gitweb/gitweb.perl

   1 #!/usr/bin/perl
   2
   3 # gitweb - simple web interface to track changes in git repositories
   4 #
   5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
   6 # (C) 2005, Christian Gierke
   7 #
   8 # This program is licensed under the GPLv2
   9
  10 use 5.008001;
  11 use strict;
  12 use warnings;
  13 # handle ACL in file access tests
  14 use filetest 'access';
  15 use CGI qw(:standard :escapeHTML -nosticky);
  16 use CGI::Util qw(unescape);
  17 use CGI::Carp qw(fatalsToBrowser set_message);
  18 use Encode;
  19 use Fcntl ':mode';
  20 use File::Find qw();
  21 use File::Basename qw(basename);
  22 use Time::HiRes qw(gettimeofday tv_interval);
  23 use Digest::MD5 qw(md5_hex);
  24
  25 binmode STDOUT, ':utf8';
  26
  27 if (!defined($CGI::VERSION) || $CGI::VERSION < 4.08) {
  28         eval 'sub CGI::multi_param { CGI::param(@_) }'
  29 }
  30
  31 our $t0 = [ gettimeofday() ];
  32 our $number_of_git_cmds = 0;
  33
  34 BEGIN {
  35         CGI->compile() if $ENV{'MOD_PERL'};
  36 }
  37
  38 our $version = "++GIT_VERSION++";
  39
  40 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
  41 sub evaluate_uri {
  42         our $cgi;
  43
  44         our $my_url = $cgi->url();
  45         our $my_uri = $cgi->url(-absolute => 1);
  46
  47         # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
  48         # needed and used only for URLs with nonempty PATH_INFO
  49         our $base_url = $my_url;
  50
  51         # When the script is used as DirectoryIndex, the URL does not contain the name
  52         # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
  53         # have to do it ourselves. We make $path_info global because it's also used
  54         # later on.
  55         #
  56         # Another issue with the script being the DirectoryIndex is that the resulting
  57         # $my_url data is not the full script URL: this is good, because we want
  58         # generated links to keep implying the script name if it wasn't explicitly
  59         # indicated in the URL we're handling, but it means that $my_url cannot be used
  60         # as base URL.
  61         # Therefore, if we needed to strip PATH_INFO, then we know that we have
  62         # to build the base URL ourselves:
  63         our $path_info = decode_utf8($ENV{"PATH_INFO"});
  64         if ($path_info) {
  65                 # $path_info has already been URL-decoded by the web server, but
  66                 # $my_url and $my_uri have not. URL-decode them so we can properly
  67                 # strip $path_info.
  68                 $my_url = unescape($my_url);
  69                 $my_uri = unescape($my_uri);
  70                 if ($my_url =~ s,\Q$path_info\E$,, &&
  71                     $my_uri =~ s,\Q$path_info\E$,, &&
  72                     defined $ENV{'SCRIPT_NAME'}) {
  73                         $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
  74                 }
  75         }
  76
  77         # target of the home link on top of all pages
  78         our $home_link = $my_uri || "/";
  79 }
  80
  81 # core git executable to use
  82 # this can just be "git" if your webserver has a sensible PATH
  83 our $GIT = "++GIT_BINDIR++/git";
  84
  85 # absolute fs-path which will be prepended to the project path
  86 #our $projectroot = "/pub/scm";
  87 our $projectroot = "++GITWEB_PROJECTROOT++";
  88
  89 # fs traversing limit for getting project list
  90 # the number is relative to the projectroot
  91 our $project_maxdepth = "++GITWEB_PROJECT_MAXDEPTH++";
  92
  93 # string of the home link on top of all pages
  94 our $home_link_str = "++GITWEB_HOME_LINK_STR++";
  95
  96 # extra breadcrumbs preceding the home link
  97 our @extra_breadcrumbs = ();
  98
  99 # name of your site or organization to appear in page titles
 100 # replace this with something more descriptive for clearer bookmarks
 101 our $site_name = "++GITWEB_SITENAME++"
 102                  || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
 103
 104 # html snippet to include in the <head> section of each page
 105 our $site_html_head_string = "++GITWEB_SITE_HTML_HEAD_STRING++";
 106 # filename of html text to include at top of each page
 107 our $site_header = "++GITWEB_SITE_HEADER++";
 108 # html text to include at home page
 109 our $home_text = "++GITWEB_HOMETEXT++";
 110 # filename of html text to include at bottom of each page
 111 our $site_footer = "++GITWEB_SITE_FOOTER++";
 112
 113 # URI of stylesheets
 114 our @stylesheets = ("++GITWEB_CSS++");
 115 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
 116 our $stylesheet = undef;
 117 # URI of GIT logo (72x27 size)
 118 our $logo = "++GITWEB_LOGO++";
 119 # URI of GIT favicon, assumed to be image/png type
 120 our $favicon = "++GITWEB_FAVICON++";
 121 # URI of gitweb.js (JavaScript code for gitweb)
 122 our $javascript = "++GITWEB_JS++";
 123
 124 # URI and label (title) of GIT logo link
 125 #our $logo_url = "https://www.kernel.org/pub/software/scm/git/docs/";
 126 #our $logo_label = "git documentation";
 127 our $logo_url = "https://git-scm.com/";
 128 our $logo_label = "git homepage";
 129
 130 # source of projects list
 131 our $projects_list = "++GITWEB_LIST++";
 132
 133 # the width (in characters) of the projects list "Description" column
 134 our $projects_list_description_width = 25;
 135
 136 # group projects by category on the projects list
 137 # (enabled if this variable evaluates to true)
 138 our $projects_list_group_categories = 0;
 139
 140 # default category if none specified
 141 # (leave the empty string for no category)
 142 our $project_list_default_category = "";
 143
 144 # default order of projects list
 145 # valid values are none, project, descr, owner, and age
 146 our $default_projects_order = "project";
 147
 148 # show repository only if this file exists
 149 # (only effective if this variable evaluates to true)
 150 our $export_ok = "++GITWEB_EXPORT_OK++";
 151
 152 # don't generate age column on the projects list page
 153 our $omit_age_column = 0;
 154
 155 # don't generate information about owners of repositories
 156 our $omit_owner=0;
 157
 158 # show repository only if this subroutine returns true
 159 # when given the path to the project, for example:
 160 #    sub { return -e "$_[0]/git-daemon-export-ok"; }
 161 our $export_auth_hook = undef;
 162
 163 # only allow viewing of repositories also shown on the overview page
 164 our $strict_export = "++GITWEB_STRICT_EXPORT++";
 165
 166 # list of git base URLs used for URL to where fetch project from,
 167 # i.e. full URL is "$git_base_url/$project"
 168 our @git_base_url_list = grep { $_ ne '' } ("++GITWEB_BASE_URL++");
 169
 170 # default blob_plain mimetype and default charset for text/plain blob
 171 our $default_blob_plain_mimetype = 'text/plain';
 172 our $default_text_plain_charset  = undef;
 173
 174 # file to use for guessing MIME types before trying /etc/mime.types
 175 # (relative to the current git repository)
 176 our $mimetypes_file = undef;
 177
 178 # assume this charset if line contains non-UTF-8 characters;
 179 # it should be valid encoding (see Encoding::Supported(3pm) for list),
 180 # for which encoding all byte sequences are valid, for example
 181 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
 182 # could be even 'utf-8' for the old behavior)
 183 our $fallback_encoding = 'latin1';
 184
 185 # rename detection options for git-diff and git-diff-tree
 186 # - default is '-M', with the cost proportional to
 187 #   (number of removed files) * (number of new files).
 188 # - more costly is '-C' (which implies '-M'), with the cost proportional to
 189 #   (number of changed files + number of removed files) * (number of new files)
 190 # - even more costly is '-C', '--find-copies-harder' with cost
 191 #   (number of files in the original tree) * (number of new files)
 192 # - one might want to include '-B' option, e.g. '-B', '-M'
 193 our @diff_opts = ('-M'); # taken from git_commit
 194
 195 # Disables features that would allow repository owners to inject script into
 196 # the gitweb domain.
 197 our $prevent_xss = 0;
 198
 199 # Path to the highlight executable to use (must be the one from
 200 # http://andre-simon.de/zip/download.php due to assumptions about parameters and output).
 201 # Useful if highlight is not installed on your webserver's PATH.
 202 # [Default: highlight]
 203 our $highlight_bin = "++HIGHLIGHT_BIN++";
 204
 205 # information about snapshot formats that gitweb is capable of serving
 206 our %known_snapshot_formats = (
 207         # name => {
 208         #       'display' => display name,
 209         #       'type' => mime type,
 210         #       'suffix' => filename suffix,
 211         #       'format' => --format for git-archive,
 212         #       'compressor' => [compressor command and arguments]
 213         #                       (array reference, optional)
 214         #       'disabled' => boolean (optional)}
 215         #
 216         'tgz' => {
 217                 'display' => 'tar.gz',
 218                 'type' => 'application/x-gzip',
 219                 'suffix' => '.tar.gz',
 220                 'format' => 'tar',
 221                 'compressor' => ['gzip', '-n']},
 222
 223         'tbz2' => {
 224                 'display' => 'tar.bz2',
 225                 'type' => 'application/x-bzip2',
 226                 'suffix' => '.tar.bz2',
 227                 'format' => 'tar',
 228                 'compressor' => ['bzip2']},
 229
 230         'txz' => {
 231                 'display' => 'tar.xz',
 232                 'type' => 'application/x-xz',
 233                 'suffix' => '.tar.xz',
 234                 'format' => 'tar',
 235                 'compressor' => ['xz'],
 236                 'disabled' => 1},
 237
 238         'zip' => {
 239                 'display' => 'zip',
 240                 'type' => 'application/x-zip',
 241                 'suffix' => '.zip',
 242                 'format' => 'zip'},
 243 );
 244
 245 # Aliases so we understand old gitweb.snapshot values in repository
 246 # configuration.
 247 our %known_snapshot_format_aliases = (
 248         'gzip'  => 'tgz',
 249         'bzip2' => 'tbz2',
 250         'xz'    => 'txz',
 251
 252         # backward compatibility: legacy gitweb config support
 253         'x-gzip' => undef, 'gz' => undef,
 254         'x-bzip2' => undef, 'bz2' => undef,
 255         'x-zip' => undef, '' => undef,
 256 );
 257
 258 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
 259 # are changed, it may be appropriate to change these values too via
 260 # $GITWEB_CONFIG.
 261 our %avatar_size = (
 262         'default' => 16,
 263         'double'  => 32
 264 );
 265
 266 # Used to set the maximum load that we will still respond to gitweb queries.
 267 # If server load exceed this value then return "503 server busy" error.
 268 # If gitweb cannot determined server load, it is taken to be 0.
 269 # Leave it undefined (or set to 'undef') to turn off load checking.
 270 our $maxload = 300;
 271
 272 # configuration for 'highlight' (http://andre-simon.de/doku/highlight/en/highlight.php)
 273 # match by basename
 274 our %highlight_basename = (
 275         #'Program' => 'py',
 276         #'Library' => 'py',
 277         'SConstruct' => 'py', # SCons equivalent of Makefile
 278         'Makefile' => 'make',
 279 );
 280 # match by extension
 281 our %highlight_ext = (
 282         # main extensions, defining name of syntax;
 283         # see files in /usr/share/highlight/langDefs/ directory
 284         (map { $_ => $_ } qw(py rb java css js tex bib xml awk bat ini spec tcl sql)),
 285         # alternate extensions, see /etc/highlight/filetypes.conf
 286         (map { $_ => 'c'   } qw(c h)),
 287         (map { $_ => 'sh'  } qw(sh bash zsh ksh)),
 288         (map { $_ => 'cpp' } qw(cpp cxx c++ cc)),
 289         (map { $_ => 'php' } qw(php php3 php4 php5 phps)),
 290         (map { $_ => 'pl'  } qw(pl perl pm)), # perhaps also 'cgi'
 291         (map { $_ => 'make'} qw(make mak mk)),
 292         (map { $_ => 'xml' } qw(xml xhtml html htm)),
 293 );
 294
 295 # You define site-wide feature defaults here; override them with
 296 # $GITWEB_CONFIG as necessary.
 297 our %feature = (
 298         # feature => {
 299         #       'sub' => feature-sub (subroutine),
 300         #       'override' => allow-override (boolean),
 301         #       'default' => [ default options...] (array reference)}
 302         #
 303         # if feature is overridable (it means that allow-override has true value),
 304         # then feature-sub will be called with default options as parameters;
 305         # return value of feature-sub indicates if to enable specified feature
 306         #
 307         # if there is no 'sub' key (no feature-sub), then feature cannot be
 308         # overridden
 309         #
 310         # use gitweb_get_feature(<feature>) to retrieve the <feature> value
 311         # (an array) or gitweb_check_feature(<feature>) to check if <feature>
 312         # is enabled
 313
 314         # Enable the 'blame' blob view, showing the last commit that modified
 315         # each line in the file. This can be very CPU-intensive.
 316
 317         # To enable system wide have in $GITWEB_CONFIG
 318         # $feature{'blame'}{'default'} = [1];
 319         # To have project specific config enable override in $GITWEB_CONFIG
 320         # $feature{'blame'}{'override'} = 1;
 321         # and in project config gitweb.blame = 0|1;
 322         'blame' => {
 323                 'sub' => sub { feature_bool('blame', @_) },
 324                 'override' => 0,
 325                 'default' => [0]},
 326
 327         # Enable the 'snapshot' link, providing a compressed archive of any
 328         # tree. This can potentially generate high traffic if you have large
 329         # project.
 330
 331         # Value is a list of formats defined in %known_snapshot_formats that
 332         # you wish to offer.
 333         # To disable system wide have in $GITWEB_CONFIG
 334         # $feature{'snapshot'}{'default'} = [];
 335         # To have project specific config enable override in $GITWEB_CONFIG
 336         # $feature{'snapshot'}{'override'} = 1;
 337         # and in project config, a comma-separated list of formats or "none"
 338         # to disable.  Example: gitweb.snapshot = tbz2,zip;
 339         'snapshot' => {
 340                 'sub' => \&feature_snapshot,
 341                 'override' => 0,
 342                 'default' => ['tgz']},
 343
 344         # Enable text search, which will list the commits which match author,
 345         # committer or commit text to a given string.  Enabled by default.
 346         # Project specific override is not supported.
 347         #
 348         # Note that this controls all search features, which means that if
 349         # it is disabled, then 'grep' and 'pickaxe' search would also be
 350         # disabled.
 351         'search' => {
 352                 'override' => 0,
 353                 'default' => [1]},
 354
 355         # Enable grep search, which will list the files in currently selected
 356         # tree containing the given string. Enabled by default. This can be
 357         # potentially CPU-intensive, of course.
 358         # Note that you need to have 'search' feature enabled too.
 359
 360         # To enable system wide have in $GITWEB_CONFIG
 361         # $feature{'grep'}{'default'} = [1];
 362         # To have project specific config enable override in $GITWEB_CONFIG
 363         # $feature{'grep'}{'override'} = 1;
 364         # and in project config gitweb.grep = 0|1;
 365         'grep' => {
 366                 'sub' => sub { feature_bool('grep', @_) },
 367                 'override' => 0,
 368                 'default' => [1]},
 369
 370         # Enable the pickaxe search, which will list the commits that modified
 371         # a given string in a file. This can be practical and quite faster
 372         # alternative to 'blame', but still potentially CPU-intensive.
 373         # Note that you need to have 'search' feature enabled too.
 374
 375         # To enable system wide have in $GITWEB_CONFIG
 376         # $feature{'pickaxe'}{'default'} = [1];
 377         # To have project specific config enable override in $GITWEB_CONFIG
 378         # $feature{'pickaxe'}{'override'} = 1;
 379         # and in project config gitweb.pickaxe = 0|1;
 380         'pickaxe' => {
 381                 'sub' => sub { feature_bool('pickaxe', @_) },
 382                 'override' => 0,
 383                 'default' => [1]},
 384
 385         # Enable showing size of blobs in a 'tree' view, in a separate
 386         # column, similar to what 'ls -l' does.  This cost a bit of IO.
 387
 388         # To disable system wide have in $GITWEB_CONFIG
 389         # $feature{'show-sizes'}{'default'} = [0];
 390         # To have project specific config enable override in $GITWEB_CONFIG
 391         # $feature{'show-sizes'}{'override'} = 1;
 392         # and in project config gitweb.showsizes = 0|1;
 393         'show-sizes' => {
 394                 'sub' => sub { feature_bool('showsizes', @_) },
 395                 'override' => 0,
 396                 'default' => [1]},
 397
 398         # Make gitweb use an alternative format of the URLs which can be
 399         # more readable and natural-looking: project name is embedded
 400         # directly in the path and the query string contains other
 401         # auxiliary information. All gitweb installations recognize
 402         # URL in either format; this configures in which formats gitweb
 403         # generates links.
 404
 405         # To enable system wide have in $GITWEB_CONFIG
 406         # $feature{'pathinfo'}{'default'} = [1];
 407         # Project specific override is not supported.
 408
 409         # Note that you will need to change the default location of CSS,
 410         # favicon, logo and possibly other files to an absolute URL. Also,
 411         # if gitweb.cgi serves as your indexfile, you will need to force
 412         # $my_uri to contain the script name in your $GITWEB_CONFIG.
 413         'pathinfo' => {
 414                 'override' => 0,
 415                 'default' => [0]},
 416
 417         # Make gitweb consider projects in project root subdirectories
 418         # to be forks of existing projects. Given project $projname.git,
 419         # projects matching $projname/*.git will not be shown in the main
 420         # projects list, instead a '+' mark will be added to $projname
 421         # there and a 'forks' view will be enabled for the project, listing
 422         # all the forks. If project list is taken from a file, forks have
 423         # to be listed after the main project.
 424
 425         # To enable system wide have in $GITWEB_CONFIG
 426         # $feature{'forks'}{'default'} = [1];
 427         # Project specific override is not supported.
 428         'forks' => {
 429                 'override' => 0,
 430                 'default' => [0]},
 431
 432         # Insert custom links to the action bar of all project pages.
 433         # This enables you mainly to link to third-party scripts integrating
 434         # into gitweb; e.g. git-browser for graphical history representation
 435         # or custom web-based repository administration interface.
 436
 437         # The 'default' value consists of a list of triplets in the form
 438         # (label, link, position) where position is the label after which
 439         # to insert the link and link is a format string where %n expands
 440         # to the project name, %f to the project path within the filesystem,
 441         # %h to the current hash (h gitweb parameter) and %b to the current
 442         # hash base (hb gitweb parameter); %% expands to %.
 443
 444         # To enable system wide have in $GITWEB_CONFIG e.g.
 445         # $feature{'actions'}{'default'} = [('graphiclog',
 446         #       '/git-browser/by-commit.html?r=%n', 'summary')];
 447         # Project specific override is not supported.
 448         'actions' => {
 449                 'override' => 0,
 450                 'default' => []},
 451
 452         # Allow gitweb scan project content tags of project repository,
 453         # and display the popular Web 2.0-ish "tag cloud" near the projects
 454         # list.  Note that this is something COMPLETELY different from the
 455         # normal Git tags.
 456
 457         # gitweb by itself can show existing tags, but it does not handle
 458         # tagging itself; you need to do it externally, outside gitweb.
 459         # The format is described in git_get_project_ctags() subroutine.
 460         # You may want to install the HTML::TagCloud Perl module to get
 461         # a pretty tag cloud instead of just a list of tags.
 462
 463         # To enable system wide have in $GITWEB_CONFIG
 464         # $feature{'ctags'}{'default'} = [1];
 465         # Project specific override is not supported.
 466
 467         # In the future whether ctags editing is enabled might depend
 468         # on the value, but using 1 should always mean no editing of ctags.
 469         'ctags' => {
 470                 'override' => 0,
 471                 'default' => [0]},
 472
 473         # The maximum number of patches in a patchset generated in patch
 474         # view. Set this to 0 or undef to disable patch view, or to a
 475         # negative number to remove any limit.
 476
 477         # To disable system wide have in $GITWEB_CONFIG
 478         # $feature{'patches'}{'default'} = [0];
 479         # To have project specific config enable override in $GITWEB_CONFIG
 480         # $feature{'patches'}{'override'} = 1;
 481         # and in project config gitweb.patches = 0|n;
 482         # where n is the maximum number of patches allowed in a patchset.
 483         'patches' => {
 484                 'sub' => \&feature_patches,
 485                 'override' => 0,
 486                 'default' => [16]},
 487
 488         # Avatar support. When this feature is enabled, views such as
 489         # shortlog or commit will display an avatar associated with
 490         # the email of the committer(s) and/or author(s).
 491
 492         # Currently available providers are gravatar and picon.
 493         # If an unknown provider is specified, the feature is disabled.
 494
 495         # Picon currently relies on the indiana.edu database.
 496
 497         # To enable system wide have in $GITWEB_CONFIG
 498         # $feature{'avatar'}{'default'} = ['<provider>'];
 499         # where <provider> is either gravatar or picon.
 500         # To have project specific config enable override in $GITWEB_CONFIG
 501         # $feature{'avatar'}{'override'} = 1;
 502         # and in project config gitweb.avatar = <provider>;
 503         'avatar' => {
 504                 'sub' => \&feature_avatar,
 505                 'override' => 0,
 506                 'default' => ['']},
 507
 508         # Enable displaying how much time and how many git commands
 509         # it took to generate and display page.  Disabled by default.
 510         # Project specific override is not supported.
 511         'timed' => {
 512                 'override' => 0,
 513                 'default' => [0]},
 514
 515         # Enable turning some links into links to actions which require
 516         # JavaScript to run (like 'blame_incremental').  Not enabled by
 517         # default.  Project specific override is currently not supported.
 518         'javascript-actions' => {
 519                 'override' => 0,
 520                 'default' => [0]},
 521
 522         # Enable and configure ability to change common timezone for dates
 523         # in gitweb output via JavaScript.  Enabled by default.
 524         # Project specific override is not supported.
 525         'javascript-timezone' => {
 526                 'override' => 0,
 527                 'default' => [
 528                         'local',     # default timezone: 'utc', 'local', or '(-|+)HHMM' format,
 529                                      # or undef to turn off this feature
 530                         'gitweb_tz', # name of cookie where to store selected timezone
 531                         'datetime',  # CSS class used to mark up dates for manipulation
 532                 ]},
 533
 534         # Syntax highlighting support. This is based on Daniel Svensson's
 535         # and Sham Chukoury's work in gitweb-xmms2.git.
 536         # It requires the 'highlight' program present in $PATH,
 537         # and therefore is disabled by default.
 538
 539         # To enable system wide have in $GITWEB_CONFIG
 540         # $feature{'highlight'}{'default'} = [1];
 541
 542         'highlight' => {
 543                 'sub' => sub { feature_bool('highlight', @_) },
 544                 'override' => 0,
 545                 'default' => [0]},
 546
 547         # Enable displaying of remote heads in the heads list
 548
 549         # To enable system wide have in $GITWEB_CONFIG
 550         # $feature{'remote_heads'}{'default'} = [1];
 551         # To have project specific config enable override in $GITWEB_CONFIG
 552         # $feature{'remote_heads'}{'override'} = 1;
 553         # and in project config gitweb.remoteheads = 0|1;
 554         'remote_heads' => {
 555                 'sub' => sub { feature_bool('remote_heads', @_) },
 556                 'override' => 0,
 557                 'default' => [0]},
 558
 559         # Enable showing branches under other refs in addition to heads
 560
 561         # To set system wide extra branch refs have in $GITWEB_CONFIG
 562         # $feature{'extra-branch-refs'}{'default'} = ['dirs', 'of', 'choice'];
 563         # To have project specific config enable override in $GITWEB_CONFIG
 564         # $feature{'extra-branch-refs'}{'override'} = 1;
 565         # and in project config gitweb.extrabranchrefs = dirs of choice
 566         # Every directory is separated with whitespace.
 567
 568         'extra-branch-refs' => {
 569                 'sub' => \&feature_extra_branch_refs,
 570                 'override' => 0,
 571                 'default' => []},
 572
 573         # Redact e-mail addresses.
 574
 575         # To enable system wide have in $GITWEB_CONFIG
 576         # $feature{'email-privacy'}{'default'} = [1];
 577         'email-privacy' => {
 578                 'sub' => sub { feature_bool('email-privacy', @_) },
 579                 'override' => 1,
 580                 'default' => [0]},
 581 );
 582
 583 sub gitweb_get_feature {
 584         my ($name) = @_;
 585         return unless exists $feature{$name};
 586         my ($sub, $override, @defaults) = (
 587                 $feature{$name}{'sub'},
 588                 $feature{$name}{'override'},
 589                 @{$feature{$name}{'default'}});
 590         # project specific override is possible only if we have project
 591         our $git_dir; # global variable, declared later
 592         if (!$override || !defined $git_dir) {
 593                 return @defaults;
 594         }
 595         if (!defined $sub) {
 596                 warn "feature $name is not overridable";
 597                 return @defaults;
 598         }
 599         return $sub->(@defaults);
 600 }
 601
 602 # A wrapper to check if a given feature is enabled.
 603 # With this, you can say
 604 #
 605 #   my $bool_feat = gitweb_check_feature('bool_feat');
 606 #   gitweb_check_feature('bool_feat') or somecode;
 607 #
 608 # instead of
 609 #
 610 #   my ($bool_feat) = gitweb_get_feature('bool_feat');
 611 #   (gitweb_get_feature('bool_feat'))[0] or somecode;
 612 #
 613 sub gitweb_check_feature {
 614         return (gitweb_get_feature(@_))[0];
 615 }
 616
 617
 618 sub feature_bool {
 619         my $key = shift;
 620         my ($val) = git_get_project_config($key, '--bool');
 621
 622         if (!defined $val) {
 623                 return ($_[0]);
 624         } elsif ($val eq 'true') {
 625                 return (1);
 626         } elsif ($val eq 'false') {
 627                 return (0);
 628         }
 629 }
 630
 631 sub feature_snapshot {
 632         my (@fmts) = @_;
 633
 634         my ($val) = git_get_project_config('snapshot');
 635
 636         if ($val) {
 637                 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
 638         }
 639
 640         return @fmts;
 641 }
 642
 643 sub feature_patches {
 644         my @val = (git_get_project_config('patches', '--int'));
 645
 646         if (@val) {
 647                 return @val;
 648         }
 649
 650         return ($_[0]);
 651 }
 652
 653 sub feature_avatar {
 654         my @val = (git_get_project_config('avatar'));
 655
 656         return @val ? @val : @_;
 657 }
 658
 659 sub feature_extra_branch_refs {
 660         my (@branch_refs) = @_;
 661         my $values = git_get_project_config('extrabranchrefs');
 662
 663         if ($values) {
 664                 $values = config_to_multi ($values);
 665                 @branch_refs = ();
 666                 foreach my $value (@{$values}) {
 667                         push @branch_refs, split /\s+/, $value;
 668                 }
 669         }
 670
 671         return @branch_refs;
 672 }
 673
 674 # checking HEAD file with -e is fragile if the repository was
 675 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
 676 # and then pruned.
 677 sub check_head_link {
 678         my ($dir) = @_;
 679         my $headfile = "$dir/HEAD";
 680         return ((-e $headfile) ||
 681                 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
 682 }
 683
 684 sub check_export_ok {
 685         my ($dir) = @_;
 686         return (check_head_link($dir) &&
 687                 (!$export_ok || -e "$dir/$export_ok") &&
 688                 (!$export_auth_hook || $export_auth_hook->($dir)));
 689 }
 690
 691 # process alternate names for backward compatibility
 692 # filter out unsupported (unknown) snapshot formats
 693 sub filter_snapshot_fmts {
 694         my @fmts = @_;
 695
 696         @fmts = map {
 697                 exists $known_snapshot_format_aliases{$_} ?
 698                        $known_snapshot_format_aliases{$_} : $_} @fmts;
 699         @fmts = grep {
 700                 exists $known_snapshot_formats{$_} &&
 701                 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
 702 }
 703
 704 sub filter_and_validate_refs {
 705         my @refs = @_;
 706         my %unique_refs = ();
 707
 708         foreach my $ref (@refs) {
 709                 die_error(500, "Invalid ref '$ref' in 'extra-branch-refs' feature") unless (is_valid_ref_format($ref));
 710                 # 'heads' are added implicitly in get_branch_refs().
 711                 $unique_refs{$ref} = 1 if ($ref ne 'heads');
 712         }
 713         return sort keys %unique_refs;
 714 }
 715
 716 # If it is set to code reference, it is code that it is to be run once per
 717 # request, allowing updating configurations that change with each request,
 718 # while running other code in config file only once.
 719 #
 720 # Otherwise, if it is false then gitweb would process config file only once;
 721 # if it is true then gitweb config would be run for each request.
 722 our $per_request_config = 1;
 723
 724 # read and parse gitweb config file given by its parameter.
 725 # returns true on success, false on recoverable error, allowing
 726 # to chain this subroutine, using first file that exists.
 727 # dies on errors during parsing config file, as it is unrecoverable.
 728 sub read_config_file {
 729         my $filename = shift;
 730         return unless defined $filename;
 731         if (-e $filename) {
 732                 do $filename;
 733                 # die if there is a problem accessing the file
 734                 die $! if $!;
 735                 # die if there are errors parsing config file
 736                 die $@ if $@;
 737                 return 1;
 738         }
 739         return;
 740 }
 741
 742 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM, $GITWEB_CONFIG_COMMON);
 743 sub evaluate_gitweb_config {
 744         our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
 745         our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
 746         our $GITWEB_CONFIG_COMMON = $ENV{'GITWEB_CONFIG_COMMON'} || "++GITWEB_CONFIG_COMMON++";
 747
 748         # Protect against duplications of file names, to not read config twice.
 749         # Only one of $GITWEB_CONFIG and $GITWEB_CONFIG_SYSTEM is used, so
 750         # there possibility of duplication of filename there doesn't matter.
 751         $GITWEB_CONFIG = ""        if ($GITWEB_CONFIG eq $GITWEB_CONFIG_COMMON);
 752         $GITWEB_CONFIG_SYSTEM = "" if ($GITWEB_CONFIG_SYSTEM eq $GITWEB_CONFIG_COMMON);
 753
 754         # Common system-wide settings for convenience.
 755         # Those settings can be overridden by GITWEB_CONFIG or GITWEB_CONFIG_SYSTEM.
 756         read_config_file($GITWEB_CONFIG_COMMON);
 757
 758         # Use first config file that exists.  This means use the per-instance
 759         # GITWEB_CONFIG if exists, otherwise use GITWEB_SYSTEM_CONFIG.
 760         read_config_file($GITWEB_CONFIG) and return;
 761         read_config_file($GITWEB_CONFIG_SYSTEM);
 762 }
 763
 764 # Get loadavg of system, to compare against $maxload.
 765 # Currently it requires '/proc/loadavg' present to get loadavg;
 766 # if it is not present it returns 0, which means no load checking.
 767 sub get_loadavg {
 768         if( -e '/proc/loadavg' ){
 769                 open my $fd, '<', '/proc/loadavg'
 770                         or return 0;
 771                 my @load = split(/\s+/, scalar <$fd>);
 772                 close $fd;
 773
 774                 # The first three columns measure CPU and IO utilization of the last one,
 775                 # five, and 10 minute periods.  The fourth column shows the number of
 776                 # currently running processes and the total number of processes in the m/n
 777                 # format.  The last column displays the last process ID used.
 778                 return $load[0] || 0;
 779         }
 780         # additional checks for load average should go here for things that don't export
 781         # /proc/loadavg
 782
 783         return 0;
 784 }
 785
 786 # version of the core git binary
 787 our $git_version;
 788 sub evaluate_git_version {
 789         our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
 790         $number_of_git_cmds++;
 791 }
 792
 793 sub check_loadavg {
 794         if (defined $maxload && get_loadavg() > $maxload) {
 795                 die_error(503, "The load average on the server is too high");
 796         }
 797 }
 798
 799 # ======================================================================
 800 # input validation and dispatch
 801
 802 # Various hash size-related values.
 803 my $sha1_len = 40;
 804 my $sha256_extra_len = 24;
 805 my $sha256_len = $sha1_len + $sha256_extra_len;
 806
 807 # A regex matching $len hex characters. $len may be a range (e.g. 7,64).
 808 sub oid_nlen_regex {
 809         my $len = shift;
 810         my $hchr = qr/[0-9a-fA-F]/;
 811         return qr/(?:(?:$hchr){$len})/;
 812 }
 813
 814 # A regex matching two sets of $nlen hex characters, prefixed by the literal
 815 # string $prefix and with the literal string $infix between them.
 816 sub oid_nlen_prefix_infix_regex {
 817         my $nlen = shift;
 818         my $prefix = shift;
 819         my $infix = shift;
 820
 821         my $rx = oid_nlen_regex($nlen);
 822
 823         return qr/^\Q$prefix\E$rx\Q$infix\E$rx$/;
 824 }
 825
 826 # A regex matching a valid object ID.
 827 our $oid_regex;
 828 {
 829         my $x = oid_nlen_regex($sha1_len);
 830         my $y = oid_nlen_regex($sha256_extra_len);
 831         $oid_regex = qr/(?:$x(?:$y)?)/;
 832 }
 833
 834 # input parameters can be collected from a variety of sources (presently, CGI
 835 # and PATH_INFO), so we define an %input_params hash that collects them all
 836 # together during validation: this allows subsequent uses (e.g. href()) to be
 837 # agnostic of the parameter origin
 838
 839 our %input_params = ();
 840
 841 # input parameters are stored with the long parameter name as key. This will
 842 # also be used in the href subroutine to convert parameters to their CGI
 843 # equivalent, and since the href() usage is the most frequent one, we store
 844 # the name -> CGI key mapping here, instead of the reverse.
 845 #
 846 # XXX: Warning: If you touch this, check the search form for updating,
 847 # too.
 848
 849 our @cgi_param_mapping = (
 850         project => "p",
 851         action => "a",
 852         file_name => "f",
 853         file_parent => "fp",
 854         hash => "h",
 855         hash_parent => "hp",
 856         hash_base => "hb",
 857         hash_parent_base => "hpb",
 858         page => "pg",
 859         order => "o",
 860         searchtext => "s",
 861         searchtype => "st",
 862         snapshot_format => "sf",
 863         extra_options => "opt",
 864         search_use_regexp => "sr",
 865         ctag => "by_tag",
 866         diff_style => "ds",
 867         project_filter => "pf",
 868         # this must be last entry (for manipulation from JavaScript)
 869         javascript => "js"
 870 );
 871 our %cgi_param_mapping = @cgi_param_mapping;
 872
 873 # we will also need to know the possible actions, for validation
 874 our %actions = (
 875         "blame" => \&git_blame,
 876         "blame_incremental" => \&git_blame_incremental,
 877         "blame_data" => \&git_blame_data,
 878         "blobdiff" => \&git_blobdiff,
 879         "blobdiff_plain" => \&git_blobdiff_plain,
 880         "blob" => \&git_blob,
 881         "blob_plain" => \&git_blob_plain,
 882         "commitdiff" => \&git_commitdiff,
 883         "commitdiff_plain" => \&git_commitdiff_plain,
 884         "commit" => \&git_commit,
 885         "forks" => \&git_forks,
 886         "heads" => \&git_heads,
 887         "history" => \&git_history,
 888         "log" => \&git_log,
 889         "patch" => \&git_patch,
 890         "patches" => \&git_patches,
 891         "remotes" => \&git_remotes,
 892         "rss" => \&git_rss,
 893         "atom" => \&git_atom,
 894         "search" => \&git_search,
 895         "search_help" => \&git_search_help,
 896         "shortlog" => \&git_shortlog,
 897         "summary" => \&git_summary,
 898         "tag" => \&git_tag,
 899         "tags" => \&git_tags,
 900         "tree" => \&git_tree,
 901         "snapshot" => \&git_snapshot,
 902         "object" => \&git_object,
 903         # those below don't need $project
 904         "opml" => \&git_opml,
 905         "project_list" => \&git_project_list,
 906         "project_index" => \&git_project_index,
 907 );
 908
 909 # finally, we have the hash of allowed extra_options for the commands that
 910 # allow them
 911 our %allowed_options = (
 912         "--no-merges" => [ qw(rss atom log shortlog history) ],
 913 );
 914
 915 # fill %input_params with the CGI parameters. All values except for 'opt'
 916 # should be single values, but opt can be an array. We should probably
 917 # build an array of parameters that can be multi-valued, but since for the time
 918 # being it's only this one, we just single it out
 919 sub evaluate_query_params {
 920         our $cgi;
 921
 922         while (my ($name, $symbol) = each %cgi_param_mapping) {
 923                 if ($symbol eq 'opt') {
 924                         $input_params{$name} = [ map { decode_utf8($_) } $cgi->multi_param($symbol) ];
 925                 } else {
 926                         $input_params{$name} = decode_utf8($cgi->param($symbol));
 927                 }
 928         }
 929 }
 930
 931 # now read PATH_INFO and update the parameter list for missing parameters
 932 sub evaluate_path_info {
 933         return if defined $input_params{'project'};
 934         return if !$path_info;
 935         $path_info =~ s,^/+,,;
 936         return if !$path_info;
 937
 938         # find which part of PATH_INFO is project
 939         my $project = $path_info;
 940         $project =~ s,/+$,,;
 941         while ($project && !check_head_link("$projectroot/$project")) {
 942                 $project =~ s,/*[^/]*$,,;
 943         }
 944         return unless $project;
 945         $input_params{'project'} = $project;
 946
 947         # do not change any parameters if an action is given using the query string
 948         return if $input_params{'action'};
 949         $path_info =~ s,^\Q$project\E/*,,;
 950
 951         # next, check if we have an action
 952         my $action = $path_info;
 953         $action =~ s,/.*$,,;
 954         if (exists $actions{$action}) {
 955                 $path_info =~ s,^$action/*,,;
 956                 $input_params{'action'} = $action;
 957         }
 958
 959         # list of actions that want hash_base instead of hash, but can have no
 960         # pathname (f) parameter
 961         my @wants_base = (
 962                 'tree',
 963                 'history',
 964         );
 965
 966         # we want to catch, among others
 967         # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
 968         my ($parentrefname, $parentpathname, $refname, $pathname) =
 969                 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
 970
 971         # first, analyze the 'current' part
 972         if (defined $pathname) {
 973                 # we got "branch:filename" or "branch:dir/"
 974                 # we could use git_get_type(branch:pathname), but:
 975                 # - it needs $git_dir
 976                 # - it does a git() call
 977                 # - the convention of terminating directories with a slash
 978                 #   makes it superfluous
 979                 # - embedding the action in the PATH_INFO would make it even
 980                 #   more superfluous
 981                 $pathname =~ s,^/+,,;
 982                 if (!$pathname || substr($pathname, -1) eq "/") {
 983                         $input_params{'action'} ||= "tree";
 984                         $pathname =~ s,/$,,;
 985                 } else {
 986                         # the default action depends on whether we had parent info
 987                         # or not
 988                         if ($parentrefname) {
 989                                 $input_params{'action'} ||= "blobdiff_plain";
 990                         } else {
 991                                 $input_params{'action'} ||= "blob_plain";
 992                         }
 993                 }
 994                 $input_params{'hash_base'} ||= $refname;
 995                 $input_params{'file_name'} ||= $pathname;
 996         } elsif (defined $refname) {
 997                 # we got "branch". In this case we have to choose if we have to
 998                 # set hash or hash_base.
 999                 #
1000                 # Most of the actions without a pathname only want hash to be
1001                 # set, except for the ones specified in @wants_base that want
1002                 # hash_base instead. It should also be noted that hand-crafted
1003                 # links having 'history' as an action and no pathname or hash
1004                 # set will fail, but that happens regardless of PATH_INFO.
1005                 if (defined $parentrefname) {
1006                         # if there is parent let the default be 'shortlog' action
1007                         # (for http://git.example.com/repo.git/A..B links); if there
1008                         # is no parent, dispatch will detect type of object and set
1009                         # action appropriately if required (if action is not set)
1010                         $input_params{'action'} ||= "shortlog";
1011                 }
1012                 if ($input_params{'action'} &&
1013                     grep { $_ eq $input_params{'action'} } @wants_base) {
1014                         $input_params{'hash_base'} ||= $refname;
1015                 } else {
1016                         $input_params{'hash'} ||= $refname;
1017                 }
1018         }
1019
1020         # next, handle the 'parent' part, if present
1021         if (defined $parentrefname) {
1022                 # a missing pathspec defaults to the 'current' filename, allowing e.g.
1023                 # someproject/blobdiff/oldrev..newrev:/filename
1024                 if ($parentpathname) {
1025                         $parentpathname =~ s,^/+,,;
1026                         $parentpathname =~ s,/$,,;
1027                         $input_params{'file_parent'} ||= $parentpathname;
1028                 } else {
1029                         $input_params{'file_parent'} ||= $input_params{'file_name'};
1030                 }
1031                 # we assume that hash_parent_base is wanted if a path was specified,
1032                 # or if the action wants hash_base instead of hash
1033                 if (defined $input_params{'file_parent'} ||
1034                         grep { $_ eq $input_params{'action'} } @wants_base) {
1035                         $input_params{'hash_parent_base'} ||= $parentrefname;
1036                 } else {
1037                         $input_params{'hash_parent'} ||= $parentrefname;
1038                 }
1039         }
1040
1041         # for the snapshot action, we allow URLs in the form
1042         # $project/snapshot/$hash.ext
1043         # where .ext determines the snapshot and gets removed from the
1044         # passed $refname to provide the $hash.
1045         #
1046         # To be able to tell that $refname includes the format extension, we
1047         # require the following two conditions to be satisfied:
1048         # - the hash input parameter MUST have been set from the $refname part
1049         #   of the URL (i.e. they must be equal)
1050         # - the snapshot format MUST NOT have been defined already (e.g. from
1051         #   CGI parameter sf)
1052         # It's also useless to try any matching unless $refname has a dot,
1053         # so we check for that too
1054         if (defined $input_params{'action'} &&
1055                 $input_params{'action'} eq 'snapshot' &&
1056                 defined $refname && index($refname, '.') != -1 &&
1057                 $refname eq $input_params{'hash'} &&
1058                 !defined $input_params{'snapshot_format'}) {
1059                 # We loop over the known snapshot formats, checking for
1060                 # extensions. Allowed extensions are both the defined suffix
1061                 # (which includes the initial dot already) and the snapshot
1062                 # format key itself, with a prepended dot
1063                 while (my ($fmt, $opt) = each %known_snapshot_formats) {
1064                         my $hash = $refname;
1065                         unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
1066                                 next;
1067                         }
1068                         my $sfx = $1;
1069                         # a valid suffix was found, so set the snapshot format
1070                         # and reset the hash parameter
1071                         $input_params{'snapshot_format'} = $fmt;
1072                         $input_params{'hash'} = $hash;
1073                         # we also set the format suffix to the one requested
1074                         # in the URL: this way a request for e.g. .tgz returns
1075                         # a .tgz instead of a .tar.gz
1076                         $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
1077                         last;
1078                 }
1079         }
1080 }
1081
1082 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
1083      $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
1084      $searchtext, $search_regexp, $project_filter);
1085 sub evaluate_and_validate_params {
1086         our $action = $input_params{'action'};
1087         if (defined $action) {
1088                 if (!is_valid_action($action)) {
1089                         die_error(400, "Invalid action parameter");
1090                 }
1091         }
1092
1093         # parameters which are pathnames
1094         our $project = $input_params{'project'};
1095         if (defined $project) {
1096                 if (!is_valid_project($project)) {
1097                         undef $project;
1098                         die_error(404, "No such project");
1099                 }
1100         }
1101
1102         our $project_filter = $input_params{'project_filter'};
1103         if (defined $project_filter) {
1104                 if (!is_valid_pathname($project_filter)) {
1105                         die_error(404, "Invalid project_filter parameter");
1106                 }
1107         }
1108
1109         our $file_name = $input_params{'file_name'};
1110         if (defined $file_name) {
1111                 if (!is_valid_pathname($file_name)) {
1112                         die_error(400, "Invalid file parameter");
1113                 }
1114         }
1115
1116         our $file_parent = $input_params{'file_parent'};
1117         if (defined $file_parent) {
1118                 if (!is_valid_pathname($file_parent)) {
1119                         die_error(400, "Invalid file parent parameter");
1120                 }
1121         }
1122
1123         # parameters which are refnames
1124         our $hash = $input_params{'hash'};
1125         if (defined $hash) {
1126                 if (!is_valid_refname($hash)) {
1127                         die_error(400, "Invalid hash parameter");
1128                 }
1129         }
1130
1131         our $hash_parent = $input_params{'hash_parent'};
1132         if (defined $hash_parent) {
1133                 if (!is_valid_refname($hash_parent)) {
1134                         die_error(400, "Invalid hash parent parameter");
1135                 }
1136         }
1137
1138         our $hash_base = $input_params{'hash_base'};
1139         if (defined $hash_base) {
1140                 if (!is_valid_refname($hash_base)) {
1141                         die_error(400, "Invalid hash base parameter");
1142                 }
1143         }
1144
1145         our @extra_options = @{$input_params{'extra_options'}};
1146         # @extra_options is always defined, since it can only be (currently) set from
1147         # CGI, and $cgi->param() returns the empty array in array context if the param
1148         # is not set
1149         foreach my $opt (@extra_options) {
1150                 if (not exists $allowed_options{$opt}) {
1151                         die_error(400, "Invalid option parameter");
1152                 }
1153                 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1154                         die_error(400, "Invalid option parameter for this action");
1155                 }
1156         }
1157
1158         our $hash_parent_base = $input_params{'hash_parent_base'};
1159         if (defined $hash_parent_base) {
1160                 if (!is_valid_refname($hash_parent_base)) {
1161                         die_error(400, "Invalid hash parent base parameter");
1162                 }
1163         }
1164
1165         # other parameters
1166         our $page = $input_params{'page'};
1167         if (defined $page) {
1168                 if ($page =~ m/[^0-9]/) {
1169                         die_error(400, "Invalid page parameter");
1170                 }
1171         }
1172
1173         our $searchtype = $input_params{'searchtype'};
1174         if (defined $searchtype) {
1175                 if ($searchtype =~ m/[^a-z]/) {
1176                         die_error(400, "Invalid searchtype parameter");
1177                 }
1178         }
1179
1180         our $search_use_regexp = $input_params{'search_use_regexp'};
1181
1182         our $searchtext = $input_params{'searchtext'};
1183         our $search_regexp = undef;
1184         if (defined $searchtext) {
1185                 if (length($searchtext) < 2) {
1186                         die_error(403, "At least two characters are required for search parameter");
1187                 }
1188                 if ($search_use_regexp) {
1189                         $search_regexp = $searchtext;
1190                         if (!eval { qr/$search_regexp/; 1; }) {
1191                                 (my $error = $@) =~ s/ at \S+ line \d+.*\n?//;
1192                                 die_error(400, "Invalid search regexp '$search_regexp'",
1193                                           esc_html($error));
1194                         }
1195                 } else {
1196                         $search_regexp = quotemeta $searchtext;
1197                 }
1198         }
1199 }
1200
1201 # path to the current git repository
1202 our $git_dir;
1203 sub evaluate_git_dir {
1204         our $git_dir = "$projectroot/$project" if $project;
1205 }
1206
1207 our (@snapshot_fmts, $git_avatar, @extra_branch_refs);
1208 sub configure_gitweb_features {
1209         # list of supported snapshot formats
1210         our @snapshot_fmts = gitweb_get_feature('snapshot');
1211         @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1212
1213         our ($git_avatar) = gitweb_get_feature('avatar');
1214         $git_avatar = '' unless $git_avatar =~ /^(?:gravatar|picon)$/s;
1215
1216         our @extra_branch_refs = gitweb_get_feature('extra-branch-refs');
1217         @extra_branch_refs = filter_and_validate_refs (@extra_branch_refs);
1218 }
1219
1220 sub get_branch_refs {
1221         return ('heads', @extra_branch_refs);
1222 }
1223
1224 # custom error handler: 'die <message>' is Internal Server Error
1225 sub handle_errors_html {
1226         my $msg = shift; # it is already HTML escaped
1227
1228         # to avoid infinite loop where error occurs in die_error,
1229         # change handler to default handler, disabling handle_errors_html
1230         set_message("Error occurred when inside die_error:\n$msg");
1231
1232         # you cannot jump out of die_error when called as error handler;
1233         # the subroutine set via CGI::Carp::set_message is called _after_
1234         # HTTP headers are already written, so it cannot write them itself
1235         die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1236 }
1237 set_message(\&handle_errors_html);
1238
1239 # dispatch
1240 sub dispatch {
1241         if (!defined $action) {
1242                 if (defined $hash) {
1243                         $action = git_get_type($hash);
1244                         $action or die_error(404, "Object does not exist");
1245                 } elsif (defined $hash_base && defined $file_name) {
1246                         $action = git_get_type("$hash_base:$file_name");
1247                         $action or die_error(404, "File or directory does not exist");
1248                 } elsif (defined $project) {
1249                         $action = 'summary';
1250                 } else {
1251                         $action = 'project_list';
1252                 }
1253         }
1254         if (!defined($actions{$action})) {
1255                 die_error(400, "Unknown action");
1256         }
1257         if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1258             !$project) {
1259                 die_error(400, "Project needed");
1260         }
1261         $actions{$action}->();
1262 }
1263
1264 sub reset_timer {
1265         our $t0 = [ gettimeofday() ]
1266                 if defined $t0;
1267         our $number_of_git_cmds = 0;
1268 }
1269
1270 our $first_request = 1;
1271 sub run_request {
1272         reset_timer();
1273
1274         evaluate_uri();
1275         if ($first_request) {
1276                 evaluate_gitweb_config();
1277                 evaluate_git_version();
1278         }
1279         if ($per_request_config) {
1280                 if (ref($per_request_config) eq 'CODE') {
1281                         $per_request_config->();
1282                 } elsif (!$first_request) {
1283                         evaluate_gitweb_config();
1284                 }
1285         }
1286         check_loadavg();
1287
1288         # $projectroot and $projects_list might be set in gitweb config file
1289         $projects_list ||= $projectroot;
1290
1291         evaluate_query_params();
1292         evaluate_path_info();
1293         evaluate_and_validate_params();
1294         evaluate_git_dir();
1295
1296         configure_gitweb_features();
1297
1298         dispatch();
1299 }
1300
1301 our $is_last_request = sub { 1 };
1302 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1303 our $CGI = 'CGI';
1304 our $cgi;
1305 our $FCGI_Stream_PRINT_raw = \&FCGI::Stream::PRINT;
1306 sub configure_as_fcgi {
1307         require CGI::Fast;
1308         our $CGI = 'CGI::Fast';
1309         # FCGI is not Unicode aware hence the UTF-8 encoding must be done manually.
1310         # However no encoding must be done within git_blob_plain() and git_snapshot()
1311         # which must still output in raw binary mode.
1312         no warnings 'redefine';
1313         my $enc = Encode::find_encoding('UTF-8');
1314         *FCGI::Stream::PRINT = sub {
1315                 my @OUTPUT = @_;
1316                 for (my $i = 1; $i < @_; $i++) {
1317                         $OUTPUT[$i] = $enc->encode($_[$i], Encode::FB_CROAK|Encode::LEAVE_SRC);
1318                 }
1319                 @_ = @OUTPUT;
1320                 goto $FCGI_Stream_PRINT_raw;
1321         };
1322
1323         my $request_number = 0;
1324         # let each child service 100 requests
1325         our $is_last_request = sub { ++$request_number > 100 };
1326 }
1327 sub evaluate_argv {
1328         my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1329         configure_as_fcgi()
1330                 if $script_name =~ /\.fcgi$/;
1331
1332         return unless (@ARGV);
1333
1334         require Getopt::Long;
1335         Getopt::Long::GetOptions(
1336                 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1337                 'nproc|n=i' => sub {
1338                         my ($arg, $val) = @_;
1339                         return unless eval { require FCGI::ProcManager; 1; };
1340                         my $proc_manager = FCGI::ProcManager->new({
1341                                 n_processes => $val,
1342                         });
1343                         our $pre_listen_hook    = sub { $proc_manager->pm_manage()        };
1344                         our $pre_dispatch_hook  = sub { $proc_manager->pm_pre_dispatch()  };
1345                         our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1346                 },
1347         );
1348 }
1349
1350 sub run {
1351         evaluate_argv();
1352
1353         $first_request = 1;
1354         $pre_listen_hook->()
1355                 if $pre_listen_hook;
1356
1357  REQUEST:
1358         while ($cgi = $CGI->new()) {
1359                 $pre_dispatch_hook->()
1360                         if $pre_dispatch_hook;
1361
1362                 run_request();
1363
1364                 $post_dispatch_hook->()
1365                         if $post_dispatch_hook;
1366                 $first_request = 0;
1367
1368                 last REQUEST if ($is_last_request->());
1369         }
1370
1371  DONE_GITWEB:
1372         1;
1373 }
1374
1375 run();
1376
1377 if (defined caller) {
1378         # wrapped in a subroutine processing requests,
1379         # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1380         return;
1381 } else {
1382         # pure CGI script, serving single request
1383         exit;
1384 }
1385
1386 ## ======================================================================
1387 ## action links
1388
1389 # possible values of extra options
1390 # -full => 0|1      - use absolute/full URL ($my_uri/$my_url as base)
1391 # -replay => 1      - start from a current view (replay with modifications)
1392 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1393 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1394 sub href {
1395         my %params = @_;
1396         # default is to use -absolute url() i.e. $my_uri
1397         my $href = $params{-full} ? $my_url : $my_uri;
1398
1399         # implicit -replay, must be first of implicit params
1400         $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1401
1402         $params{'project'} = $project unless exists $params{'project'};
1403
1404         if ($params{-replay}) {
1405                 while (my ($name, $symbol) = each %cgi_param_mapping) {
1406                         if (!exists $params{$name}) {
1407                                 $params{$name} = $input_params{$name};
1408                         }
1409                 }
1410         }
1411
1412         my $use_pathinfo = gitweb_check_feature('pathinfo');
1413         if (defined $params{'project'} &&
1414             (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1415                 # try to put as many parameters as possible in PATH_INFO:
1416                 #   - project name
1417                 #   - action
1418                 #   - hash_parent or hash_parent_base:/file_parent
1419                 #   - hash or hash_base:/filename
1420                 #   - the snapshot_format as an appropriate suffix
1421
1422                 # When the script is the root DirectoryIndex for the domain,
1423                 # $href here would be something like http://gitweb.example.com/
1424                 # Thus, we strip any trailing / from $href, to spare us double
1425                 # slashes in the final URL
1426                 $href =~ s,/$,,;
1427
1428                 # Then add the project name, if present
1429                 $href .= "/".esc_path_info($params{'project'});
1430                 delete $params{'project'};
1431
1432                 # since we destructively absorb parameters, we keep this
1433                 # boolean that remembers if we're handling a snapshot
1434                 my $is_snapshot = $params{'action'} eq 'snapshot';
1435
1436                 # Summary just uses the project path URL, any other action is
1437                 # added to the URL
1438                 if (defined $params{'action'}) {
1439                         $href .= "/".esc_path_info($params{'action'})
1440                                 unless $params{'action'} eq 'summary';
1441                         delete $params{'action'};
1442                 }
1443
1444                 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1445                 # stripping nonexistent or useless pieces
1446                 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1447                         || $params{'hash_parent'} || $params{'hash'});
1448                 if (defined $params{'hash_base'}) {
1449                         if (defined $params{'hash_parent_base'}) {
1450                                 $href .= esc_path_info($params{'hash_parent_base'});
1451                                 # skip the file_parent if it's the same as the file_name
1452                                 if (defined $params{'file_parent'}) {
1453                                         if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1454                                                 delete $params{'file_parent'};
1455                                         } elsif ($params{'file_parent'} !~ /\.\./) {
1456                                                 $href .= ":/".esc_path_info($params{'file_parent'});
1457                                                 delete $params{'file_parent'};
1458                                         }
1459                                 }
1460                                 $href .= "..";
1461                                 delete $params{'hash_parent'};
1462                                 delete $params{'hash_parent_base'};
1463                         } elsif (defined $params{'hash_parent'}) {
1464                                 $href .= esc_path_info($params{'hash_parent'}). "..";
1465                                 delete $params{'hash_parent'};
1466                         }
1467
1468                         $href .= esc_path_info($params{'hash_base'});
1469                         if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1470                                 $href .= ":/".esc_path_info($params{'file_name'});
1471                                 delete $params{'file_name'};
1472                         }
1473                         delete $params{'hash'};
1474                         delete $params{'hash_base'};
1475                 } elsif (defined $params{'hash'}) {
1476                         $href .= esc_path_info($params{'hash'});
1477                         delete $params{'hash'};
1478                 }
1479
1480                 # If the action was a snapshot, we can absorb the
1481                 # snapshot_format parameter too
1482                 if ($is_snapshot) {
1483                         my $fmt = $params{'snapshot_format'};
1484                         # snapshot_format should always be defined when href()
1485                         # is called, but just in case some code forgets, we
1486                         # fall back to the default
1487                         $fmt ||= $snapshot_fmts[0];
1488                         $href .= $known_snapshot_formats{$fmt}{'suffix'};
1489                         delete $params{'snapshot_format'};
1490                 }
1491         }
1492
1493         # now encode the parameters explicitly
1494         my @result = ();
1495         for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1496                 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1497                 if (defined $params{$name}) {
1498                         if (ref($params{$name}) eq "ARRAY") {
1499                                 foreach my $par (@{$params{$name}}) {
1500                                         push @result, $symbol . "=" . esc_param($par);
1501                                 }
1502                         } else {
1503                                 push @result, $symbol . "=" . esc_param($params{$name});
1504                         }
1505                 }
1506         }
1507         $href .= "?" . join(';', @result) if scalar @result;
1508
1509         # final transformation: trailing spaces must be escaped (URI-encoded)
1510         $href =~ s/(\s+)$/CGI::escape($1)/e;
1511
1512         if ($params{-anchor}) {
1513                 $href .= "#".esc_param($params{-anchor});
1514         }
1515
1516         return $href;
1517 }
1518
1519
1520 ## ======================================================================
1521 ## validation, quoting/unquoting and escaping
1522
1523 sub is_valid_action {
1524         my $input = shift;
1525         return undef unless exists $actions{$input};
1526         return 1;
1527 }
1528
1529 sub is_valid_project {
1530         my $input = shift;
1531
1532         return unless defined $input;
1533         if (!is_valid_pathname($input) ||
1534                 !(-d "$projectroot/$input") ||
1535                 !check_export_ok("$projectroot/$input") ||
1536                 ($strict_export && !project_in_list($input))) {
1537                 return undef;
1538         } else {
1539                 return 1;
1540         }
1541 }
1542
1543 sub is_valid_pathname {
1544         my $input = shift;
1545
1546         return undef unless defined $input;
1547         # no '.' or '..' as elements of path, i.e. no '.' or '..'
1548         # at the beginning, at the end, and between slashes.
1549         # also this catches doubled slashes
1550         if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1551                 return undef;
1552         }
1553         # no null characters
1554         if ($input =~ m!\0!) {
1555                 return undef;
1556         }
1557         return 1;
1558 }
1559
1560 sub is_valid_ref_format {
1561         my $input = shift;
1562
1563         return undef unless defined $input;
1564         # restrictions on ref name according to git-check-ref-format
1565         if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1566                 return undef;
1567         }
1568         return 1;
1569 }
1570
1571 sub is_valid_refname {
1572         my $input = shift;
1573
1574         return undef unless defined $input;
1575         # textual hashes are O.K.
1576         if ($input =~ m/^$oid_regex$/) {
1577                 return 1;
1578         }
1579         # it must be correct pathname
1580         is_valid_pathname($input) or return undef;
1581         # check git-check-ref-format restrictions
1582         is_valid_ref_format($input) or return undef;
1583         return 1;
1584 }
1585
1586 # decode sequences of octets in utf8 into Perl's internal form,
1587 # which is utf-8 with utf8 flag set if needed.  gitweb writes out
1588 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1589 sub to_utf8 {
1590         my $str = shift;
1591         return undef unless defined $str;
1592
1593         if (utf8::is_utf8($str) || utf8::decode($str)) {
1594                 return $str;
1595         } else {
1596                 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1597         }
1598 }
1599
1600 # quote unsafe chars, but keep the slash, even when it's not
1601 # correct, but quoted slashes look too horrible in bookmarks
1602 sub esc_param {
1603         my $str = shift;
1604         return undef unless defined $str;
1605         $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1606         $str =~ s/ /\+/g;
1607         return $str;
1608 }
1609
1610 # the quoting rules for path_info fragment are slightly different
1611 sub esc_path_info {
1612         my $str = shift;
1613         return undef unless defined $str;
1614
1615         # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1616         $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1617
1618         return $str;
1619 }
1620
1621 # quote unsafe chars in whole URL, so some characters cannot be quoted
1622 sub esc_url {
1623         my $str = shift;
1624         return undef unless defined $str;
1625         $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1626         $str =~ s/ /\+/g;
1627         return $str;
1628 }
1629
1630 # quote unsafe characters in HTML attributes
1631 sub esc_attr {
1632
1633         # for XHTML conformance escaping '"' to '&quot;' is not enough
1634         return esc_html(@_);
1635 }
1636
1637 # replace invalid utf8 character with SUBSTITUTION sequence
1638 sub esc_html {
1639         my $str = shift;
1640         my %opts = @_;
1641
1642         return undef unless defined $str;
1643
1644         $str = to_utf8($str);
1645         $str = $cgi->escapeHTML($str);
1646         if ($opts{'-nbsp'}) {
1647                 $str =~ s/ /&nbsp;/g;
1648         }
1649         $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1650         return $str;
1651 }
1652
1653 # quote control characters and escape filename to HTML
1654 sub esc_path {
1655         my $str = shift;
1656         my %opts = @_;
1657
1658         return undef unless defined $str;
1659
1660         $str = to_utf8($str);
1661         $str = $cgi->escapeHTML($str);
1662         if ($opts{'-nbsp'}) {
1663                 $str =~ s/ /&nbsp;/g;
1664         }
1665         $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1666         return $str;
1667 }
1668
1669 # Sanitize for use in XHTML + application/xml+xhtml (valid XML 1.0)
1670 sub sanitize {
1671         my $str = shift;
1672
1673         return undef unless defined $str;
1674
1675         $str = to_utf8($str);
1676         $str =~ s|([[:cntrl:]])|(index("\t\n\r", $1) != -1 ? $1 : quot_cec($1))|eg;
1677         return $str;
1678 }
1679
1680 # Make control characters "printable", using character escape codes (CEC)
1681 sub quot_cec {
1682         my $cntrl = shift;
1683         my %opts = @_;
1684         my %es = ( # character escape codes, aka escape sequences
1685                 "\t" => '\t',   # tab             (HT)
1686                 "\n" => '\n',   # line feed       (LF)
1687                 "\r" => '\r',   # carriage return (CR)
1688                 "\f" => '\f',   # form feed       (FF)
1689                 "\b" => '\b',   # backspace       (BS)
1690                 "\a" => '\a',   # alarm (bell)    (BEL)
1691                 "\e" => '\e',   # escape          (ESC)
1692                 "\013" => '\v', # vertical tab    (VT)
1693                 "\000" => '\0', # nul character   (NUL)
1694         );
1695         my $chr = ( (exists $es{$cntrl})
1696                     ? $es{$cntrl}
1697                     : sprintf('\%2x', ord($cntrl)) );
1698         if ($opts{-nohtml}) {
1699                 return $chr;
1700         } else {
1701                 return "<span class=\"cntrl\">$chr</span>";
1702         }
1703 }
1704
1705 # Alternatively use unicode control pictures codepoints,
1706 # Unicode "printable representation" (PR)
1707 sub quot_upr {
1708         my $cntrl = shift;
1709         my %opts = @_;
1710
1711         my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1712         if ($opts{-nohtml}) {
1713                 return $chr;
1714         } else {
1715                 return "<span class=\"cntrl\">$chr</span>";
1716         }
1717 }
1718
1719 # git may return quoted and escaped filenames
1720 sub unquote {
1721         my $str = shift;
1722
1723         sub unq {
1724                 my $seq = shift;
1725                 my %es = ( # character escape codes, aka escape sequences
1726                         't' => "\t",   # tab            (HT, TAB)
1727                         'n' => "\n",   # newline        (NL)
1728                         'r' => "\r",   # return         (CR)
1729                         'f' => "\f",   # form feed      (FF)
1730                         'b' => "\b",   # backspace      (BS)
1731                         'a' => "\a",   # alarm (bell)   (BEL)
1732                         'e' => "\e",   # escape         (ESC)
1733                         'v' => "\013", # vertical tab   (VT)
1734                 );
1735
1736                 if ($seq =~ m/^[0-7]{1,3}$/) {
1737                         # octal char sequence
1738                         return chr(oct($seq));
1739                 } elsif (exists $es{$seq}) {
1740                         # C escape sequence, aka character escape code
1741                         return $es{$seq};
1742                 }
1743                 # quoted ordinary character
1744                 return $seq;
1745         }
1746
1747         if ($str =~ m/^"(.*)"$/) {
1748                 # needs unquoting
1749                 $str = $1;
1750                 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1751         }
1752         return $str;
1753 }
1754
1755 # escape tabs (convert tabs to spaces)
1756 sub untabify {
1757         my $line = shift;
1758
1759         while ((my $pos = index($line, "\t")) != -1) {
1760                 if (my $count = (8 - ($pos % 8))) {
1761                         my $spaces = ' ' x $count;
1762                         $line =~ s/\t/$spaces/;
1763                 }
1764         }
1765
1766         return $line;
1767 }
1768
1769 sub project_in_list {
1770         my $project = shift;
1771         my @list = git_get_projects_list();
1772         return @list && scalar(grep { $_->{'path'} eq $project } @list);
1773 }
1774
1775 ## ----------------------------------------------------------------------
1776 ## HTML aware string manipulation
1777
1778 # Try to chop given string on a word boundary between position
1779 # $len and $len+$add_len. If there is no word boundary there,
1780 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1781 # (marking chopped part) would be longer than given string.
1782 sub chop_str {
1783         my $str = shift;
1784         my $len = shift;
1785         my $add_len = shift || 10;
1786         my $where = shift || 'right'; # 'left' | 'center' | 'right'
1787
1788         # Make sure perl knows it is utf8 encoded so we don't
1789         # cut in the middle of a utf8 multibyte char.
1790         $str = to_utf8($str);
1791
1792         # allow only $len chars, but don't cut a word if it would fit in $add_len
1793         # if it doesn't fit, cut it if it's still longer than the dots we would add
1794         # remove chopped character entities entirely
1795
1796         # when chopping in the middle, distribute $len into left and right part
1797         # return early if chopping wouldn't make string shorter
1798         if ($where eq 'center') {
1799                 return $str if ($len + 5 >= length($str)); # filler is length 5
1800                 $len = int($len/2);
1801         } else {
1802                 return $str if ($len + 4 >= length($str)); # filler is length 4
1803         }
1804
1805         # regexps: ending and beginning with word part up to $add_len
1806         my $endre = qr/.{$len}\w{0,$add_len}/;
1807         my $begre = qr/\w{0,$add_len}.{$len}/;
1808
1809         if ($where eq 'left') {
1810                 $str =~ m/^(.*?)($begre)$/;
1811                 my ($lead, $body) = ($1, $2);
1812                 if (length($lead) > 4) {
1813                         $lead = " ...";
1814                 }
1815                 return "$lead$body";
1816
1817         } elsif ($where eq 'center') {
1818                 $str =~ m/^($endre)(.*)$/;
1819                 my ($left, $str)  = ($1, $2);
1820                 $str =~ m/^(.*?)($begre)$/;
1821                 my ($mid, $right) = ($1, $2);
1822                 if (length($mid) > 5) {
1823                         $mid = " ... ";
1824                 }
1825                 return "$left$mid$right";
1826
1827         } else {
1828                 $str =~ m/^($endre)(.*)$/;
1829                 my $body = $1;
1830                 my $tail = $2;
1831                 if (length($tail) > 4) {
1832                         $tail = "... ";
1833                 }
1834                 return "$body$tail";
1835         }
1836 }
1837
1838 # takes the same arguments as chop_str, but also wraps a <span> around the
1839 # result with a title attribute if it does get chopped. Additionally, the
1840 # string is HTML-escaped.
1841 sub chop_and_escape_str {
1842         my ($str) = @_;
1843
1844         my $chopped = chop_str(@_);
1845         $str = to_utf8($str);
1846         if ($chopped eq $str) {
1847                 return esc_html($chopped);
1848         } else {
1849                 $str =~ s/[[:cntrl:]]/?/g;
1850                 return $cgi->span({-title=>$str}, esc_html($chopped));
1851         }
1852 }
1853
1854 # Highlight selected fragments of string, using given CSS class,
1855 # and escape HTML.  It is assumed that fragments do not overlap.
1856 # Regions are passed as list of pairs (array references).
1857 #
1858 # Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
1859 # '<span class="mark">foo</span>bar'
1860 sub esc_html_hl_regions {
1861         my ($str, $css_class, @sel) = @_;
1862         my %opts = grep { ref($_) ne 'ARRAY' } @sel;
1863         @sel     = grep { ref($_) eq 'ARRAY' } @sel;
1864         return esc_html($str, %opts) unless @sel;
1865
1866         my $out = '';
1867         my $pos = 0;
1868
1869         for my $s (@sel) {
1870                 my ($begin, $end) = @$s;
1871
1872                 # Don't create empty <span> elements.
1873                 next if $end <= $begin;
1874
1875                 my $escaped = esc_html(substr($str, $begin, $end - $begin),
1876                                        %opts);
1877
1878                 $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
1879                         if ($begin - $pos > 0);
1880                 $out .= $cgi->span({-class => $css_class}, $escaped);
1881
1882                 $pos = $end;
1883         }
1884         $out .= esc_html(substr($str, $pos), %opts)
1885                 if ($pos < length($str));
1886
1887         return $out;
1888 }
1889
1890 # return positions of beginning and end of each match
1891 sub matchpos_list {
1892         my ($str, $regexp) = @_;
1893         return unless (defined $str && defined $regexp);
1894
1895         my @matches;
1896         while ($str =~ /$regexp/g) {
1897                 push @matches, [$-[0], $+[0]];
1898         }
1899         return @matches;
1900 }
1901
1902 # highlight match (if any), and escape HTML
1903 sub esc_html_match_hl {
1904         my ($str, $regexp) = @_;
1905         return esc_html($str) unless defined $regexp;
1906
1907         my @matches = matchpos_list($str, $regexp);
1908         return esc_html($str) unless @matches;
1909
1910         return esc_html_hl_regions($str, 'match', @matches);
1911 }
1912
1913
1914 # highlight match (if any) of shortened string, and escape HTML
1915 sub esc_html_match_hl_chopped {
1916         my ($str, $chopped, $regexp) = @_;
1917         return esc_html_match_hl($str, $regexp) unless defined $chopped;
1918
1919         my @matches = matchpos_list($str, $regexp);
1920         return esc_html($chopped) unless @matches;
1921
1922         # filter matches so that we mark chopped string
1923         my $tail = "... "; # see chop_str
1924         unless ($chopped =~ s/\Q$tail\E$//) {
1925                 $tail = '';
1926         }
1927         my $chop_len = length($chopped);
1928         my $tail_len = length($tail);
1929         my @filtered;
1930
1931         for my $m (@matches) {
1932                 if ($m->[0] > $chop_len) {
1933                         push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
1934                         last;
1935                 } elsif ($m->[1] > $chop_len) {
1936                         push @filtered, [ $m->[0], $chop_len + $tail_len ];
1937                         last;
1938                 }
1939                 push @filtered, $m;
1940         }
1941
1942         return esc_html_hl_regions($chopped . $tail, 'match', @filtered);
1943 }
1944
1945 ## ----------------------------------------------------------------------
1946 ## functions returning short strings
1947
1948 # CSS class for given age value (in seconds)
1949 sub age_class {
1950         my $age = shift;
1951
1952         if (!defined $age) {
1953                 return "noage";
1954         } elsif ($age < 60*60*2) {
1955                 return "age0";
1956         } elsif ($age < 60*60*24*2) {
1957                 return "age1";
1958         } else {
1959                 return "age2";
1960         }
1961 }
1962
1963 # convert age in seconds to "nn units ago" string
1964 sub age_string {
1965         my $age = shift;
1966         my $age_str;
1967
1968         if ($age > 60*60*24*365*2) {
1969                 $age_str = (int $age/60/60/24/365);
1970                 $age_str .= " years ago";
1971         } elsif ($age > 60*60*24*(365/12)*2) {
1972                 $age_str = int $age/60/60/24/(365/12);
1973                 $age_str .= " months ago";
1974         } elsif ($age > 60*60*24*7*2) {
1975                 $age_str = int $age/60/60/24/7;
1976                 $age_str .= " weeks ago";
1977         } elsif ($age > 60*60*24*2) {
1978                 $age_str = int $age/60/60/24;
1979                 $age_str .= " days ago";
1980         } elsif ($age > 60*60*2) {
1981                 $age_str = int $age/60/60;
1982                 $age_str .= " hours ago";
1983         } elsif ($age > 60*2) {
1984                 $age_str = int $age/60;
1985                 $age_str .= " min ago";
1986         } elsif ($age > 2) {
1987                 $age_str = int $age;
1988                 $age_str .= " sec ago";
1989         } else {
1990                 $age_str .= " right now";
1991         }
1992         return $age_str;
1993 }
1994
1995 use constant {
1996         S_IFINVALID => 0030000,
1997         S_IFGITLINK => 0160000,
1998 };
1999
2000 # submodule/subproject, a commit object reference
2001 sub S_ISGITLINK {
2002         my $mode = shift;
2003
2004         return (($mode & S_IFMT) == S_IFGITLINK)
2005 }
2006
2007 # convert file mode in octal to symbolic file mode string
2008 sub mode_str {
2009         my $mode = oct shift;
2010
2011         if (S_ISGITLINK($mode)) {
2012                 return 'm---------';
2013         } elsif (S_ISDIR($mode & S_IFMT)) {
2014                 return 'drwxr-xr-x';
2015         } elsif (S_ISLNK($mode)) {
2016                 return 'lrwxrwxrwx';
2017         } elsif (S_ISREG($mode)) {
2018                 # git cares only about the executable bit
2019                 if ($mode & S_IXUSR) {
2020                         return '-rwxr-xr-x';
2021                 } else {
2022                         return '-rw-r--r--';
2023                 };
2024         } else {
2025                 return '----------';
2026         }
2027 }
2028
2029 # convert file mode in octal to file type string
2030 sub file_type {
2031         my $mode = shift;
2032
2033         if ($mode !~ m/^[0-7]+$/) {
2034                 return $mode;
2035         } else {
2036                 $mode = oct $mode;
2037         }
2038
2039         if (S_ISGITLINK($mode)) {
2040                 return "submodule";
2041         } elsif (S_ISDIR($mode & S_IFMT)) {
2042                 return "directory";
2043         } elsif (S_ISLNK($mode)) {
2044                 return "symlink";
2045         } elsif (S_ISREG($mode)) {
2046                 return "file";
2047         } else {
2048                 return "unknown";
2049         }
2050 }
2051
2052 # convert file mode in octal to file type description string
2053 sub file_type_long {
2054         my $mode = shift;
2055
2056         if ($mode !~ m/^[0-7]+$/) {
2057                 return $mode;
2058         } else {
2059                 $mode = oct $mode;
2060         }
2061
2062         if (S_ISGITLINK($mode)) {
2063                 return "submodule";
2064         } elsif (S_ISDIR($mode & S_IFMT)) {
2065                 return "directory";
2066         } elsif (S_ISLNK($mode)) {
2067                 return "symlink";
2068         } elsif (S_ISREG($mode)) {
2069                 if ($mode & S_IXUSR) {
2070                         return "executable";
2071                 } else {
2072                         return "file";
2073                 };
2074         } else {
2075                 return "unknown";
2076         }
2077 }
2078
2079
2080 ## ----------------------------------------------------------------------
2081 ## functions returning short HTML fragments, or transforming HTML fragments
2082 ## which don't belong to other sections
2083
2084 # format line of commit message.
2085 sub format_log_line_html {
2086         my $line = shift;
2087
2088         # Potentially abbreviated OID.
2089         my $regex = oid_nlen_regex("7,64");
2090
2091         $line = esc_html($line, -nbsp=>1);
2092         $line =~ s{
2093         \b
2094         (
2095             # The output of "git describe", e.g. v2.10.0-297-gf6727b0
2096             # or hadoop-20160921-113441-20-g094fb7d
2097             (?<!-) # see strbuf_check_tag_ref(). Tags can't start with -
2098             [A-Za-z0-9.-]+
2099             (?!\.) # refs can't end with ".", see check_refname_format()
2100             -g$regex
2101             |
2102             # Just a normal looking Git SHA1
2103             $regex
2104         )
2105         \b
2106     }{
2107                 $cgi->a({-href => href(action=>"object", hash=>$1),
2108                                         -class => "text"}, $1);
2109         }egx;
2110
2111         return $line;
2112 }
2113
2114 # format marker of refs pointing to given object
2115
2116 # the destination action is chosen based on object type and current context:
2117 # - for annotated tags, we choose the tag view unless it's the current view
2118 #   already, in which case we go to shortlog view
2119 # - for other refs, we keep the current view if we're in history, shortlog or
2120 #   log view, and select shortlog otherwise
2121 sub format_ref_marker {
2122         my ($refs, $id) = @_;
2123         my $markers = '';
2124
2125         if (defined $refs->{$id}) {
2126                 foreach my $ref (@{$refs->{$id}}) {
2127                         # this code exploits the fact that non-lightweight tags are the
2128                         # only indirect objects, and that they are the only objects for which
2129                         # we want to use tag instead of shortlog as action
2130                         my ($type, $name) = qw();
2131                         my $indirect = ($ref =~ s/\^\{\}$//);
2132                         # e.g. tags/v2.6.11 or heads/next
2133                         if ($ref =~ m!^(.*?)s?/(.*)$!) {
2134                                 $type = $1;
2135                                 $name = $2;
2136                         } else {
2137                                 $type = "ref";
2138                                 $name = $ref;
2139                         }
2140
2141                         my $class = $type;
2142                         $class .= " indirect" if $indirect;
2143
2144                         my $dest_action = "shortlog";
2145
2146                         if ($indirect) {
2147                                 $dest_action = "tag" unless $action eq "tag";
2148                         } elsif ($action =~ /^(history|(short)?log)$/) {
2149                                 $dest_action = $action;
2150                         }
2151
2152                         my $dest = "";
2153                         $dest .= "refs/" unless $ref =~ m!^refs/!;
2154                         $dest .= $ref;
2155
2156                         my $link = $cgi->a({
2157                                 -href => href(
2158                                         action=>$dest_action,
2159                                         hash=>$dest
2160                                 )}, esc_html($name));
2161
2162                         $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
2163                                 $link . "</span>";
2164                 }
2165         }
2166
2167         if ($markers) {
2168                 return ' <span class="refs">'. $markers . '</span>';
2169         } else {
2170                 return "";
2171         }
2172 }
2173
2174 # format, perhaps shortened and with markers, title line
2175 sub format_subject_html {
2176         my ($long, $short, $href, $extra) = @_;
2177         $extra = '' unless defined($extra);
2178
2179         if (length($short) < length($long)) {
2180                 $long =~ s/[[:cntrl:]]/?/g;
2181                 return $cgi->a({-href => $href, -class => "list subject",
2182                                 -title => to_utf8($long)},
2183                        esc_html($short)) . $extra;
2184         } else {
2185                 return $cgi->a({-href => $href, -class => "list subject"},
2186                        esc_html($long)) . $extra;
2187         }
2188 }
2189
2190 # Rather than recomputing the url for an email multiple times, we cache it
2191 # after the first hit. This gives a visible benefit in views where the avatar
2192 # for the same email is used repeatedly (e.g. shortlog).
2193 # The cache is shared by all avatar engines (currently gravatar only), which
2194 # are free to use it as preferred. Since only one avatar engine is used for any
2195 # given page, there's no risk for cache conflicts.
2196 our %avatar_cache = ();
2197
2198 # Compute the picon url for a given email, by using the picon search service over at
2199 # http://www.cs.indiana.edu/picons/search.html
2200 sub picon_url {
2201         my $email = lc shift;
2202         if (!$avatar_cache{$email}) {
2203                 my ($user, $domain) = split('@', $email);
2204                 $avatar_cache{$email} =
2205                         "//www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
2206                         "$domain/$user/" .
2207                         "users+domains+unknown/up/single";
2208         }
2209         return $avatar_cache{$email};
2210 }
2211
2212 # Compute the gravatar url for a given email, if it's not in the cache already.
2213 # Gravatar stores only the part of the URL before the size, since that's the
2214 # one computationally more expensive. This also allows reuse of the cache for
2215 # different sizes (for this particular engine).
2216 sub gravatar_url {
2217         my $email = lc shift;
2218         my $size = shift;
2219         $avatar_cache{$email} ||=
2220                 "//www.gravatar.com/avatar/" .
2221                         md5_hex($email) . "?s=";
2222         return $avatar_cache{$email} . $size;
2223 }
2224
2225 # Insert an avatar for the given $email at the given $size if the feature
2226 # is enabled.
2227 sub git_get_avatar {
2228         my ($email, %opts) = @_;
2229         my $pre_white  = ($opts{-pad_before} ? "&nbsp;" : "");
2230         my $post_white = ($opts{-pad_after}  ? "&nbsp;" : "");
2231         $opts{-size} ||= 'default';
2232         my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
2233         my $url = "";
2234         if ($git_avatar eq 'gravatar') {
2235                 $url = gravatar_url($email, $size);
2236         } elsif ($git_avatar eq 'picon') {
2237                 $url = picon_url($email);
2238         }
2239         # Other providers can be added by extending the if chain, defining $url
2240         # as needed. If no variant puts something in $url, we assume avatars
2241         # are completely disabled/unavailable.
2242         if ($url) {
2243                 return $pre_white .
2244                        "<img width=\"$size\" " .
2245                             "class=\"avatar\" " .
2246                             "src=\"".esc_url($url)."\" " .
2247                             "alt=\"\" " .
2248                        "/>" . $post_white;
2249         } else {
2250                 return "";
2251         }
2252 }
2253
2254 sub format_search_author {
2255         my ($author, $searchtype, $displaytext) = @_;
2256         my $have_search = gitweb_check_feature('search');
2257
2258         if ($have_search) {
2259                 my $performed = "";
2260                 if ($searchtype eq 'author') {
2261                         $performed = "authored";
2262                 } elsif ($searchtype eq 'committer') {
2263                         $performed = "committed";
2264                 }
2265
2266                 return $cgi->a({-href => href(action=>"search", hash=>$hash,
2267                                 searchtext=>$author,
2268                                 searchtype=>$searchtype), class=>"list",
2269                                 title=>"Search for commits $performed by $author"},
2270                                 $displaytext);
2271
2272         } else {
2273                 return $displaytext;
2274         }
2275 }
2276
2277 # format the author name of the given commit with the given tag
2278 # the author name is chopped and escaped according to the other
2279 # optional parameters (see chop_str).
2280 sub format_author_html {
2281         my $tag = shift;
2282         my $co = shift;
2283         my $author = chop_and_escape_str($co->{'author_name'}, @_);
2284         return "<$tag class=\"author\">" .
2285                format_search_author($co->{'author_name'}, "author",
2286                        git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2287                        $author) .
2288                "</$tag>";
2289 }
2290
2291 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2292 sub format_git_diff_header_line {
2293         my $line = shift;
2294         my $diffinfo = shift;
2295         my ($from, $to) = @_;
2296
2297         if ($diffinfo->{'nparents'}) {
2298                 # combined diff
2299                 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2300                 if ($to->{'href'}) {
2301                         $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2302                                          esc_path($to->{'file'}));
2303                 } else { # file was deleted (no href)
2304                         $line .= esc_path($to->{'file'});
2305                 }
2306         } else {
2307                 # "ordinary" diff
2308                 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2309                 if ($from->{'href'}) {
2310                         $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2311                                          'a/' . esc_path($from->{'file'}));
2312                 } else { # file was added (no href)
2313                         $line .= 'a/' . esc_path($from->{'file'});
2314                 }
2315                 $line .= ' ';
2316                 if ($to->{'href'}) {
2317                         $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2318                                          'b/' . esc_path($to->{'file'}));
2319                 } else { # file was deleted
2320                         $line .= 'b/' . esc_path($to->{'file'});
2321                 }
2322         }
2323
2324         return "<div class=\"diff header\">$line</div>\n";
2325 }
2326
2327 # format extended diff header line, before patch itself
2328 sub format_extended_diff_header_line {
2329         my $line = shift;
2330         my $diffinfo = shift;
2331         my ($from, $to) = @_;
2332
2333         # match <path>
2334         if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2335                 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2336                                        esc_path($from->{'file'}));
2337         }
2338         if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2339                 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2340                                  esc_path($to->{'file'}));
2341         }
2342         # match single <mode>
2343         if ($line =~ m/\s(\d{6})$/) {
2344                 $line .= '<span class="info"> (' .
2345                          file_type_long($1) .
2346                          ')</span>';
2347         }
2348         # match <hash>
2349         if ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", ",") |
2350             $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", ",")) {
2351                 # can match only for combined diff
2352                 $line = 'index ';
2353                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2354                         if ($from->{'href'}[$i]) {
2355                                 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2356                                                   -class=>"hash"},
2357                                                  substr($diffinfo->{'from_id'}[$i],0,7));
2358                         } else {
2359                                 $line .= '0' x 7;
2360                         }
2361                         # separator
2362                         $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2363                 }
2364                 $line .= '..';
2365                 if ($to->{'href'}) {
2366                         $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2367                                          substr($diffinfo->{'to_id'},0,7));
2368                 } else {
2369                         $line .= '0' x 7;
2370                 }
2371
2372         } elsif ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", "..") |
2373                  $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", "..")) {
2374                 # can match only for ordinary diff
2375                 my ($from_link, $to_link);
2376                 if ($from->{'href'}) {
2377                         $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2378                                              substr($diffinfo->{'from_id'},0,7));
2379                 } else {
2380                         $from_link = '0' x 7;
2381                 }
2382                 if ($to->{'href'}) {
2383                         $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2384                                            substr($diffinfo->{'to_id'},0,7));
2385                 } else {
2386                         $to_link = '0' x 7;
2387                 }
2388                 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2389                 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2390         }
2391
2392         return $line . "<br/>\n";
2393 }
2394
2395 # format from-file/to-file diff header
2396 sub format_diff_from_to_header {
2397         my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2398         my $line;
2399         my $result = '';
2400
2401         $line = $from_line;
2402         #assert($line =~ m/^---/) if DEBUG;
2403         # no extra formatting for "^--- /dev/null"
2404         if (! $diffinfo->{'nparents'}) {
2405                 # ordinary (single parent) diff
2406                 if ($line =~ m!^--- "?a/!) {
2407                         if ($from->{'href'}) {
2408                                 $line = '--- a/' .
2409                                         $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2410                                                 esc_path($from->{'file'}));
2411                         } else {
2412                                 $line = '--- a/' .
2413                                         esc_path($from->{'file'});
2414                         }
2415                 }
2416                 $result .= qq!<div class="diff from_file">$line</div>\n!;
2417
2418         } else {
2419                 # combined diff (merge commit)
2420                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2421                         if ($from->{'href'}[$i]) {
2422                                 $line = '--- ' .
2423                                         $cgi->a({-href=>href(action=>"blobdiff",
2424                                                              hash_parent=>$diffinfo->{'from_id'}[$i],
2425                                                              hash_parent_base=>$parents[$i],
2426                                                              file_parent=>$from->{'file'}[$i],
2427                                                              hash=>$diffinfo->{'to_id'},
2428                                                              hash_base=>$hash,
2429                                                              file_name=>$to->{'file'}),
2430                                                  -class=>"path",
2431                                                  -title=>"diff" . ($i+1)},
2432                                                 $i+1) .
2433                                         '/' .
2434                                         $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2435                                                 esc_path($from->{'file'}[$i]));
2436                         } else {
2437                                 $line = '--- /dev/null';
2438                         }
2439                         $result .= qq!<div class="diff from_file">$line</div>\n!;
2440                 }
2441         }
2442
2443         $line = $to_line;
2444         #assert($line =~ m/^\+\+\+/) if DEBUG;
2445         # no extra formatting for "^+++ /dev/null"
2446         if ($line =~ m!^\+\+\+ "?b/!) {
2447                 if ($to->{'href'}) {
2448                         $line = '+++ b/' .
2449                                 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2450                                         esc_path($to->{'file'}));
2451                 } else {
2452                         $line = '+++ b/' .
2453                                 esc_path($to->{'file'});
2454                 }
2455         }
2456         $result .= qq!<div class="diff to_file">$line</div>\n!;
2457
2458         return $result;
2459 }
2460
2461 # create note for patch simplified by combined diff
2462 sub format_diff_cc_simplified {
2463         my ($diffinfo, @parents) = @_;
2464         my $result = '';
2465
2466         $result .= "<div class=\"diff header\">" .
2467                    "diff --cc ";
2468         if (!is_deleted($diffinfo)) {
2469                 $result .= $cgi->a({-href => href(action=>"blob",
2470                                                   hash_base=>$hash,
2471                                                   hash=>$diffinfo->{'to_id'},
2472                                                   file_name=>$diffinfo->{'to_file'}),
2473                                     -class => "path"},
2474                                    esc_path($diffinfo->{'to_file'}));
2475         } else {
2476                 $result .= esc_path($diffinfo->{'to_file'});
2477         }
2478         $result .= "</div>\n" . # class="diff header"
2479                    "<div class=\"diff nodifferences\">" .
2480                    "Simple merge" .
2481                    "</div>\n"; # class="diff nodifferences"
2482
2483         return $result;
2484 }
2485
2486 sub diff_line_class {
2487         my ($line, $from, $to) = @_;
2488
2489         # ordinary diff
2490         my $num_sign = 1;
2491         # combined diff
2492         if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2493                 $num_sign = scalar @{$from->{'href'}};
2494         }
2495
2496         my @diff_line_classifier = (
2497                 { regexp => qr/^\@\@{$num_sign} /, class => "chunk_header"},
2498                 { regexp => qr/^\\/,               class => "incomplete"  },
2499                 { regexp => qr/^ {$num_sign}/,     class => "ctx" },
2500                 # classifier for context must come before classifier add/rem,
2501                 # or we would have to use more complicated regexp, for example
2502                 # qr/(?= {0,$m}\+)[+ ]{$num_sign}/, where $m = $num_sign - 1;
2503                 { regexp => qr/^[+ ]{$num_sign}/,   class => "add" },
2504                 { regexp => qr/^[- ]{$num_sign}/,   class => "rem" },
2505         );
2506         for my $clsfy (@diff_line_classifier) {
2507                 return $clsfy->{'class'}
2508                         if ($line =~ $clsfy->{'regexp'});
2509         }
2510
2511         # fallback
2512         return "";
2513 }
2514
2515 # assumes that $from and $to are defined and correctly filled,
2516 # and that $line holds a line of chunk header for unified diff
2517 sub format_unidiff_chunk_header {
2518         my ($line, $from, $to) = @_;
2519
2520         my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2521                 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2522
2523         $from_lines = 0 unless defined $from_lines;
2524         $to_lines   = 0 unless defined $to_lines;
2525
2526         if ($from->{'href'}) {
2527                 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2528                                      -class=>"list"}, $from_text);
2529         }
2530         if ($to->{'href'}) {
2531                 $to_text   = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2532                                      -class=>"list"}, $to_text);
2533         }
2534         $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2535                 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2536         return $line;
2537 }
2538
2539 # assumes that $from and $to are defined and correctly filled,
2540 # and that $line holds a line of chunk header for combined diff
2541 sub format_cc_diff_chunk_header {
2542         my ($line, $from, $to) = @_;
2543
2544         my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2545         my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2546
2547         @from_text = split(' ', $ranges);
2548         for (my $i = 0; $i < @from_text; ++$i) {
2549                 ($from_start[$i], $from_nlines[$i]) =
2550                         (split(',', substr($from_text[$i], 1)), 0);
2551         }
2552
2553         $to_text   = pop @from_text;
2554         $to_start  = pop @from_start;
2555         $to_nlines = pop @from_nlines;
2556
2557         $line = "<span class=\"chunk_info\">$prefix ";
2558         for (my $i = 0; $i < @from_text; ++$i) {
2559                 if ($from->{'href'}[$i]) {
2560                         $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2561                                           -class=>"list"}, $from_text[$i]);
2562                 } else {
2563                         $line .= $from_text[$i];
2564                 }
2565                 $line .= " ";
2566         }
2567         if ($to->{'href'}) {
2568                 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2569                                   -class=>"list"}, $to_text);
2570         } else {
2571                 $line .= $to_text;
2572         }
2573         $line .= " $prefix</span>" .
2574                  "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2575         return $line;
2576 }
2577
2578 # process patch (diff) line (not to be used for diff headers),
2579 # returning HTML-formatted (but not wrapped) line.
2580 # If the line is passed as a reference, it is treated as HTML and not
2581 # esc_html()'ed.
2582 sub format_diff_line {
2583         my ($line, $diff_class, $from, $to) = @_;
2584
2585         if (ref($line)) {
2586                 $line = $$line;
2587         } else {
2588                 chomp $line;
2589                 $line = untabify($line);
2590
2591                 if ($from && $to && $line =~ m/^\@{2} /) {
2592                         $line = format_unidiff_chunk_header($line, $from, $to);
2593                 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2594                         $line = format_cc_diff_chunk_header($line, $from, $to);
2595                 } else {
2596                         $line = esc_html($line, -nbsp=>1);
2597                 }
2598         }
2599
2600         my $diff_classes = "diff";
2601         $diff_classes .= " $diff_class" if ($diff_class);
2602         $line = "<div class=\"$diff_classes\">$line</div>\n";
2603
2604         return $line;
2605 }
2606
2607 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2608 # linked.  Pass the hash of the tree/commit to snapshot.
2609 sub format_snapshot_links {
2610         my ($hash) = @_;
2611         my $num_fmts = @snapshot_fmts;
2612         if ($num_fmts > 1) {
2613                 # A parenthesized list of links bearing format names.
2614                 # e.g. "snapshot (_tar.gz_ _zip_)"
2615                 return "snapshot (" . join(' ', map
2616                         $cgi->a({
2617                                 -href => href(
2618                                         action=>"snapshot",
2619                                         hash=>$hash,
2620                                         snapshot_format=>$_
2621                                 )
2622                         }, $known_snapshot_formats{$_}{'display'})
2623                 , @snapshot_fmts) . ")";
2624         } elsif ($num_fmts == 1) {
2625                 # A single "snapshot" link whose tooltip bears the format name.
2626                 # i.e. "_snapshot_"
2627                 my ($fmt) = @snapshot_fmts;
2628                 return
2629                         $cgi->a({
2630                                 -href => href(
2631                                         action=>"snapshot",
2632                                         hash=>$hash,
2633                                         snapshot_format=>$fmt
2634                                 ),
2635                                 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2636                         }, "snapshot");
2637         } else { # $num_fmts == 0
2638                 return undef;
2639         }
2640 }
2641
2642 ## ......................................................................
2643 ## functions returning values to be passed, perhaps after some
2644 ## transformation, to other functions; e.g. returning arguments to href()
2645
2646 # returns hash to be passed to href to generate gitweb URL
2647 # in -title key it returns description of link
2648 sub get_feed_info {
2649         my $format = shift || 'Atom';
2650         my %res = (action => lc($format));
2651         my $matched_ref = 0;
2652
2653         # feed links are possible only for project views
2654         return unless (defined $project);
2655         # some views should link to OPML, or to generic project feed,
2656         # or don't have specific feed yet (so they should use generic)
2657         return if (!$action || $action =~ /^(?:tags|heads|forks|tag|search)$/x);
2658
2659         my $branch = undef;
2660         # branches refs uses 'refs/' + $get_branch_refs()[x] + '/' prefix
2661         # (fullname) to differentiate from tag links; this also makes
2662         # possible to detect branch links
2663         for my $ref (get_branch_refs()) {
2664                 if ((defined $hash_base && $hash_base =~ m!^refs/\Q$ref\E/(.*)$!) ||
2665                     (defined $hash      && $hash      =~ m!^refs/\Q$ref\E/(.*)$!)) {
2666                         $branch = $1;
2667                         $matched_ref = $ref;
2668                         last;
2669                 }
2670         }
2671         # find log type for feed description (title)
2672         my $type = 'log';
2673         if (defined $file_name) {
2674                 $type  = "history of $file_name";
2675                 $type .= "/" if ($action eq 'tree');
2676                 $type .= " on '$branch'" if (defined $branch);
2677         } else {
2678                 $type = "log of $branch" if (defined $branch);
2679         }
2680
2681         $res{-title} = $type;
2682         $res{'hash'} = (defined $branch ? "refs/$matched_ref/$branch" : undef);
2683         $res{'file_name'} = $file_name;
2684
2685         return %res;
2686 }
2687
2688 ## ----------------------------------------------------------------------
2689 ## git utility subroutines, invoking git commands
2690
2691 # returns path to the core git executable and the --git-dir parameter as list
2692 sub git_cmd {
2693         $number_of_git_cmds++;
2694         return $GIT, '--git-dir='.$git_dir;
2695 }
2696
2697 # quote the given arguments for passing them to the shell
2698 # quote_command("command", "arg 1", "arg with ' and ! characters")
2699 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2700 # Try to avoid using this function wherever possible.
2701 sub quote_command {
2702         return join(' ',
2703                 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2704 }
2705
2706 # get HEAD ref of given project as hash
2707 sub git_get_head_hash {
2708         return git_get_full_hash(shift, 'HEAD');
2709 }
2710
2711 sub git_get_full_hash {
2712         return git_get_hash(@_);
2713 }
2714
2715 sub git_get_short_hash {
2716         return git_get_hash(@_, '--short=7');
2717 }
2718
2719 sub git_get_hash {
2720         my ($project, $hash, @options) = @_;
2721         my $o_git_dir = $git_dir;
2722         my $retval = undef;
2723         $git_dir = "$projectroot/$project";
2724         if (open my $fd, '-|', git_cmd(), 'rev-parse',
2725             '--verify', '-q', @options, $hash) {
2726                 $retval = <$fd>;
2727                 chomp $retval if defined $retval;
2728                 close $fd;
2729         }
2730         if (defined $o_git_dir) {
2731                 $git_dir = $o_git_dir;
2732         }
2733         return $retval;
2734 }
2735
2736 # get type of given object
2737 sub git_get_type {
2738         my $hash = shift;
2739
2740         open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2741         my $type = <$fd>;
2742         close $fd or return;
2743         chomp $type;
2744         return $type;
2745 }
2746
2747 # repository configuration
2748 our $config_file = '';
2749 our %config;
2750
2751 # store multiple values for single key as anonymous array reference
2752 # single values stored directly in the hash, not as [ <value> ]
2753 sub hash_set_multi {
2754         my ($hash, $key, $value) = @_;
2755
2756         if (!exists $hash->{$key}) {
2757                 $hash->{$key} = $value;
2758         } elsif (!ref $hash->{$key}) {
2759                 $hash->{$key} = [ $hash->{$key}, $value ];
2760         } else {
2761                 push @{$hash->{$key}}, $value;
2762         }
2763 }
2764
2765 # return hash of git project configuration
2766 # optionally limited to some section, e.g. 'gitweb'
2767 sub git_parse_project_config {
2768         my $section_regexp = shift;
2769         my %config;
2770
2771         local $/ = "\0";
2772
2773         open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2774                 or return;
2775
2776         while (my $keyval = <$fh>) {
2777                 chomp $keyval;
2778                 my ($key, $value) = split(/\n/, $keyval, 2);
2779
2780                 hash_set_multi(\%config, $key, $value)
2781                         if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2782         }
2783         close $fh;
2784
2785         return %config;
2786 }
2787
2788 # convert config value to boolean: 'true' or 'false'
2789 # no value, number > 0, 'true' and 'yes' values are true
2790 # rest of values are treated as false (never as error)
2791 sub config_to_bool {
2792         my $val = shift;
2793
2794         return 1 if !defined $val;             # section.key
2795
2796         # strip leading and trailing whitespace
2797         $val =~ s/^\s+//;
2798         $val =~ s/\s+$//;
2799
2800         return (($val =~ /^\d+$/ && $val) ||   # section.key = 1
2801                 ($val =~ /^(?:true|yes)$/i));  # section.key = true
2802 }
2803
2804 # convert config value to simple decimal number
2805 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2806 # to be multiplied by 1024, 1048576, or 1073741824
2807 sub config_to_int {
2808         my $val = shift;
2809
2810         # strip leading and trailing whitespace
2811         $val =~ s/^\s+//;
2812         $val =~ s/\s+$//;
2813
2814         if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2815                 $unit = lc($unit);
2816                 # unknown unit is treated as 1
2817                 return $num * ($unit eq 'g' ? 1073741824 :
2818                                $unit eq 'm' ?    1048576 :
2819                                $unit eq 'k' ?       1024 : 1);
2820         }
2821         return $val;
2822 }
2823
2824 # convert config value to array reference, if needed
2825 sub config_to_multi {
2826         my $val = shift;
2827
2828         return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2829 }
2830
2831 sub git_get_project_config {
2832         my ($key, $type) = @_;
2833
2834         return unless defined $git_dir;
2835
2836         # key sanity check
2837         return unless ($key);
2838         # only subsection, if exists, is case sensitive,
2839         # and not lowercased by 'git config -z -l'
2840         if (my ($hi, $mi, $lo) = ($key =~ /^([^.]*)\.(.*)\.([^.]*)$/)) {
2841                 $lo =~ s/_//g;
2842                 $key = join(".", lc($hi), $mi, lc($lo));
2843                 return if ($lo =~ /\W/ || $hi =~ /\W/);
2844         } else {
2845                 $key = lc($key);
2846                 $key =~ s/_//g;
2847                 return if ($key =~ /\W/);
2848         }
2849         $key =~ s/^gitweb\.//;
2850
2851         # type sanity check
2852         if (defined $type) {
2853                 $type =~ s/^--//;
2854                 $type = undef
2855                         unless ($type eq 'bool' || $type eq 'int');
2856         }
2857
2858         # get config
2859         if (!defined $config_file ||
2860             $config_file ne "$git_dir/config") {
2861                 %config = git_parse_project_config('gitweb');
2862                 $config_file = "$git_dir/config";
2863         }
2864
2865         # check if config variable (key) exists
2866         return unless exists $config{"gitweb.$key"};
2867
2868         # ensure given type
2869         if (!defined $type) {
2870                 return $config{"gitweb.$key"};
2871         } elsif ($type eq 'bool') {
2872                 # backward compatibility: 'git config --bool' returns true/false
2873                 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2874         } elsif ($type eq 'int') {
2875                 return config_to_int($config{"gitweb.$key"});
2876         }
2877         return $config{"gitweb.$key"};
2878 }
2879
2880 # get hash of given path at given ref
2881 sub git_get_hash_by_path {
2882         my $base = shift;
2883         my $path = shift || return undef;
2884         my $type = shift;
2885
2886         $path =~ s,/+$,,;
2887
2888         open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2889                 or die_error(500, "Open git-ls-tree failed");
2890         my $line = <$fd>;
2891         close $fd or return undef;
2892
2893         if (!defined $line) {
2894                 # there is no tree or hash given by $path at $base
2895                 return undef;
2896         }
2897
2898         #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
2899         $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/;
2900         if (defined $type && $type ne $2) {
2901                 # type doesn't match
2902                 return undef;
2903         }
2904         return $3;
2905 }
2906
2907 # get path of entry with given hash at given tree-ish (ref)
2908 # used to get 'from' filename for combined diff (merge commit) for renames
2909 sub git_get_path_by_hash {
2910         my $base = shift || return;
2911         my $hash = shift || return;
2912
2913         local $/ = "\0";
2914
2915         open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2916                 or return undef;
2917         while (my $line = <$fd>) {
2918                 chomp $line;
2919
2920                 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423  gitweb'
2921                 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f  gitweb/README'
2922                 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2923                         close $fd;
2924                         return $1;
2925                 }
2926         }
2927         close $fd;
2928         return undef;
2929 }
2930
2931 ## ......................................................................
2932 ## git utility functions, directly accessing git repository
2933
2934 # get the value of config variable either from file named as the variable
2935 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2936 # configuration variable in the repository config file.
2937 sub git_get_file_or_project_config {
2938         my ($path, $name) = @_;
2939
2940         $git_dir = "$projectroot/$path";
2941         open my $fd, '<', "$git_dir/$name"
2942                 or return git_get_project_config($name);
2943         my $conf = <$fd>;
2944         close $fd;
2945         if (defined $conf) {
2946                 chomp $conf;
2947         }
2948         return $conf;
2949 }
2950
2951 sub git_get_project_description {
2952         my $path = shift;
2953         return git_get_file_or_project_config($path, 'description');
2954 }
2955
2956 sub git_get_project_category {
2957         my $path = shift;
2958         return git_get_file_or_project_config($path, 'category');
2959 }
2960
2961
2962 # supported formats:
2963 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2964 #   - if its contents is a number, use it as tag weight,
2965 #   - otherwise add a tag with weight 1
2966 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2967 #   the same value multiple times increases tag weight
2968 # * `gitweb.ctag' multi-valued repo config variable
2969 sub git_get_project_ctags {
2970         my $project = shift;
2971         my $ctags = {};
2972
2973         $git_dir = "$projectroot/$project";
2974         if (opendir my $dh, "$git_dir/ctags") {
2975                 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2976                 foreach my $tagfile (@files) {
2977                         open my $ct, '<', $tagfile
2978                                 or next;
2979                         my $val = <$ct>;
2980                         chomp $val if $val;
2981                         close $ct;
2982
2983                         (my $ctag = $tagfile) =~ s#.*/##;
2984                         if ($val =~ /^\d+$/) {
2985                                 $ctags->{$ctag} = $val;
2986                         } else {
2987                                 $ctags->{$ctag} = 1;
2988                         }
2989                 }
2990                 closedir $dh;
2991
2992         } elsif (open my $fh, '<', "$git_dir/ctags") {
2993                 while (my $line = <$fh>) {
2994                         chomp $line;
2995                         $ctags->{$line}++ if $line;
2996                 }
2997                 close $fh;
2998
2999         } else {
3000                 my $taglist = config_to_multi(git_get_project_config('ctag'));
3001                 foreach my $tag (@$taglist) {
3002                         $ctags->{$tag}++;
3003                 }
3004         }
3005
3006         return $ctags;
3007 }
3008
3009 # return hash, where keys are content tags ('ctags'),
3010 # and values are sum of weights of given tag in every project
3011 sub git_gather_all_ctags {
3012         my $projects = shift;
3013         my $ctags = {};
3014
3015         foreach my $p (@$projects) {
3016                 foreach my $ct (keys %{$p->{'ctags'}}) {
3017                         $ctags->{$ct} += $p->{'ctags'}->{$ct};
3018                 }
3019         }
3020
3021         return $ctags;
3022 }
3023
3024 sub git_populate_project_tagcloud {
3025         my $ctags = shift;
3026
3027         # First, merge different-cased tags; tags vote on casing
3028         my %ctags_lc;
3029         foreach (keys %$ctags) {
3030                 $ctags_lc{lc $_}->{count} += $ctags->{$_};
3031                 if (not $ctags_lc{lc $_}->{topcount}
3032                     or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
3033                         $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
3034                         $ctags_lc{lc $_}->{topname} = $_;
3035                 }
3036         }
3037
3038         my $cloud;
3039         my $matched = $input_params{'ctag'};
3040         if (eval { require HTML::TagCloud; 1; }) {
3041                 $cloud = HTML::TagCloud->new;
3042                 foreach my $ctag (sort keys %ctags_lc) {
3043                         # Pad the title with spaces so that the cloud looks
3044                         # less crammed.
3045                         my $title = esc_html($ctags_lc{$ctag}->{topname});
3046                         $title =~ s/ /&nbsp;/g;
3047                         $title =~ s/^/&nbsp;/g;
3048                         $title =~ s/$/&nbsp;/g;
3049                         if (defined $matched && $matched eq $ctag) {
3050                                 $title = qq(<span class="match">$title</span>);
3051                         }
3052                         $cloud->add($title, href(project=>undef, ctag=>$ctag),
3053                                     $ctags_lc{$ctag}->{count});
3054                 }
3055         } else {
3056                 $cloud = {};
3057                 foreach my $ctag (keys %ctags_lc) {
3058                         my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
3059                         if (defined $matched && $matched eq $ctag) {
3060                                 $title = qq(<span class="match">$title</span>);
3061                         }
3062                         $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
3063                         $cloud->{$ctag}{ctag} =
3064                                 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
3065                 }
3066         }
3067         return $cloud;
3068 }
3069
3070 sub git_show_project_tagcloud {
3071         my ($cloud, $count) = @_;
3072         if (ref $cloud eq 'HTML::TagCloud') {
3073                 return $cloud->html_and_css($count);
3074         } else {
3075                 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
3076                 return
3077                         '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
3078                         join (', ', map {
3079                                 $cloud->{$_}->{'ctag'}
3080                         } splice(@tags, 0, $count)) .
3081                         '</div>';
3082         }
3083 }
3084
3085 sub git_get_project_url_list {
3086         my $path = shift;
3087
3088         $git_dir = "$projectroot/$path";
3089         open my $fd, '<', "$git_dir/cloneurl"
3090                 or return wantarray ?
3091                 @{ config_to_multi(git_get_project_config('url')) } :
3092                    config_to_multi(git_get_project_config('url'));
3093         my @git_project_url_list = map { chomp; $_ } <$fd>;
3094         close $fd;
3095
3096         return wantarray ? @git_project_url_list : \@git_project_url_list;
3097 }
3098
3099 sub git_get_projects_list {
3100         my $filter = shift || '';
3101         my $paranoid = shift;
3102         my @list;
3103
3104         if (-d $projects_list) {
3105                 # search in directory
3106                 my $dir = $projects_list;
3107                 # remove the trailing "/"
3108                 $dir =~ s!/+$!!;
3109                 my $pfxlen = length("$dir");
3110                 my $pfxdepth = ($dir =~ tr!/!!);
3111                 # when filtering, search only given subdirectory
3112                 if ($filter && !$paranoid) {
3113                         $dir .= "/$filter";
3114                         $dir =~ s!/+$!!;
3115                 }
3116
3117                 File::Find::find({
3118                         follow_fast => 1, # follow symbolic links
3119                         follow_skip => 2, # ignore duplicates
3120                         dangling_symlinks => 0, # ignore dangling symlinks, silently
3121                         wanted => sub {
3122                                 # global variables
3123                                 our $project_maxdepth;
3124                                 our $projectroot;
3125                                 # skip project-list toplevel, if we get it.
3126                                 return if (m!^[/.]$!);
3127                                 # only directories can be git repositories
3128                                 return unless (-d $_);
3129                                 # need search permission
3130                                 return unless (-x $_);
3131                                 # don't traverse too deep (Find is super slow on os x)
3132                                 # $project_maxdepth excludes depth of $projectroot
3133                                 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
3134                                         $File::Find::prune = 1;
3135                                         return;
3136                                 }
3137
3138                                 my $path = substr($File::Find::name, $pfxlen + 1);
3139                                 # paranoidly only filter here
3140                                 if ($paranoid && $filter && $path !~ m!^\Q$filter\E/!) {
3141                                         next;
3142                                 }
3143                                 # we check related file in $projectroot
3144                                 if (check_export_ok("$projectroot/$path")) {
3145                                         push @list, { path => $path };
3146                                         $File::Find::prune = 1;
3147                                 }
3148                         },
3149                 }, "$dir");
3150
3151         } elsif (-f $projects_list) {
3152                 # read from file(url-encoded):
3153                 # 'git%2Fgit.git Linus+Torvalds'
3154                 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3155                 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3156                 open my $fd, '<', $projects_list or return;
3157         PROJECT:
3158                 while (my $line = <$fd>) {
3159                         chomp $line;
3160                         my ($path, $owner) = split ' ', $line;
3161                         $path = unescape($path);
3162                         $owner = unescape($owner);
3163                         if (!defined $path) {
3164                                 next;
3165                         }
3166                         # if $filter is rpovided, check if $path begins with $filter
3167                         if ($filter && $path !~ m!^\Q$filter\E/!) {
3168                                 next;
3169                         }
3170                         if (check_export_ok("$projectroot/$path")) {
3171                                 my $pr = {
3172                                         path => $path
3173                                 };
3174                                 if ($owner) {
3175                                         $pr->{'owner'} = to_utf8($owner);
3176                                 }
3177                                 push @list, $pr;
3178                         }
3179                 }
3180                 close $fd;
3181         }
3182         return @list;
3183 }
3184
3185 # written with help of Tree::Trie module (Perl Artistic License, GPL compatible)
3186 # as side effects it sets 'forks' field to list of forks for forked projects
3187 sub filter_forks_from_projects_list {
3188         my $projects = shift;
3189
3190         my %trie; # prefix tree of directories (path components)
3191         # generate trie out of those directories that might contain forks
3192         foreach my $pr (@$projects) {
3193                 my $path = $pr->{'path'};
3194                 $path =~ s/\.git$//;      # forks of 'repo.git' are in 'repo/' directory
3195                 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
3196                 next unless ($path);      # skip '.git' repository: tests, git-instaweb
3197                 next unless (-d "$projectroot/$path"); # containing directory exists
3198                 $pr->{'forks'} = [];      # there can be 0 or more forks of project
3199
3200                 # add to trie
3201                 my @dirs = split('/', $path);
3202                 # walk the trie, until either runs out of components or out of trie
3203                 my $ref = \%trie;
3204                 while (scalar @dirs &&
3205                        exists($ref->{$dirs[0]})) {
3206                         $ref = $ref->{shift @dirs};
3207                 }
3208                 # create rest of trie structure from rest of components
3209                 foreach my $dir (@dirs) {
3210                         $ref = $ref->{$dir} = {};
3211                 }
3212                 # create end marker, store $pr as a data
3213                 $ref->{''} = $pr if (!exists $ref->{''});
3214         }
3215
3216         # filter out forks, by finding shortest prefix match for paths
3217         my @filtered;
3218  PROJECT:
3219         foreach my $pr (@$projects) {
3220                 # trie lookup
3221                 my $ref = \%trie;
3222         DIR:
3223                 foreach my $dir (split('/', $pr->{'path'})) {
3224                         if (exists $ref->{''}) {
3225                                 # found [shortest] prefix, is a fork - skip it
3226                                 push @{$ref->{''}{'forks'}}, $pr;
3227                                 next PROJECT;
3228                         }
3229                         if (!exists $ref->{$dir}) {
3230                                 # not in trie, cannot have prefix, not a fork
3231                                 push @filtered, $pr;
3232                                 next PROJECT;
3233                         }
3234                         # If the dir is there, we just walk one step down the trie.
3235                         $ref = $ref->{$dir};
3236                 }
3237                 # we ran out of trie
3238                 # (shouldn't happen: it's either no match, or end marker)
3239                 push @filtered, $pr;
3240         }
3241
3242         return @filtered;
3243 }
3244
3245 # note: fill_project_list_info must be run first,
3246 # for 'descr_long' and 'ctags' to be filled
3247 sub search_projects_list {
3248         my ($projlist, %opts) = @_;
3249         my $tagfilter  = $opts{'tagfilter'};
3250         my $search_re = $opts{'search_regexp'};
3251
3252         return @$projlist
3253                 unless ($tagfilter || $search_re);
3254
3255         # searching projects require filling to be run before it;
3256         fill_project_list_info($projlist,
3257                                $tagfilter  ? 'ctags' : (),
3258                                $search_re ? ('path', 'descr') : ());
3259         my @projects;
3260  PROJECT:
3261         foreach my $pr (@$projlist) {
3262
3263                 if ($tagfilter) {
3264                         next unless ref($pr->{'ctags'}) eq 'HASH';
3265                         next unless
3266                                 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
3267                 }
3268
3269                 if ($search_re) {
3270                         next unless
3271                                 $pr->{'path'} =~ /$search_re/ ||
3272                                 $pr->{'descr_long'} =~ /$search_re/;
3273                 }
3274
3275                 push @projects, $pr;
3276         }
3277
3278         return @projects;
3279 }
3280
3281 our $gitweb_project_owner = undef;
3282 sub git_get_project_list_from_file {
3283
3284         return if (defined $gitweb_project_owner);
3285
3286         $gitweb_project_owner = {};
3287         # read from file (url-encoded):
3288         # 'git%2Fgit.git Linus+Torvalds'
3289         # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3290         # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3291         if (-f $projects_list) {
3292                 open(my $fd, '<', $projects_list);
3293                 while (my $line = <$fd>) {
3294                         chomp $line;
3295                         my ($pr, $ow) = split ' ', $line;
3296                         $pr = unescape($pr);
3297                         $ow = unescape($ow);
3298                         $gitweb_project_owner->{$pr} = to_utf8($ow);
3299                 }
3300                 close $fd;
3301         }
3302 }
3303
3304 sub git_get_project_owner {
3305         my $project = shift;
3306         my $owner;
3307
3308         return undef unless $project;
3309         $git_dir = "$projectroot/$project";
3310
3311         if (!defined $gitweb_project_owner) {
3312                 git_get_project_list_from_file();
3313         }
3314
3315         if (exists $gitweb_project_owner->{$project}) {
3316                 $owner = $gitweb_project_owner->{$project};
3317         }
3318         if (!defined $owner){
3319                 $owner = git_get_project_config('owner');
3320         }
3321         if (!defined $owner) {
3322                 $owner = get_file_owner("$git_dir");
3323         }
3324
3325         return $owner;
3326 }
3327
3328 sub git_get_last_activity {
3329         my ($path) = @_;
3330         my $fd;
3331
3332         $git_dir = "$projectroot/$path";
3333         open($fd, "-|", git_cmd(), 'for-each-ref',
3334              '--format=%(committer)',
3335              '--sort=-committerdate',
3336              '--count=1',
3337              map { "refs/$_" } get_branch_refs ()) or return;
3338         my $most_recent = <$fd>;
3339         close $fd or return;
3340         if (defined $most_recent &&
3341             $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
3342                 my $timestamp = $1;
3343                 my $age = time - $timestamp;
3344                 return ($age, age_string($age));
3345         }
3346         return (undef, undef);
3347 }
3348
3349 # Implementation note: when a single remote is wanted, we cannot use 'git
3350 # remote show -n' because that command always work (assuming it's a remote URL
3351 # if it's not defined), and we cannot use 'git remote show' because that would
3352 # try to make a network roundtrip. So the only way to find if that particular
3353 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3354 # and when we find what we want.
3355 sub git_get_remotes_list {
3356         my $wanted = shift;
3357         my %remotes = ();
3358
3359         open my $fd, '-|' , git_cmd(), 'remote', '-v';
3360         return unless $fd;
3361         while (my $remote = <$fd>) {
3362                 chomp $remote;
3363                 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3364                 next if $wanted and not $remote eq $wanted;
3365                 my ($url, $key) = ($1, $2);
3366
3367                 $remotes{$remote} ||= { 'heads' => () };
3368                 $remotes{$remote}{$key} = $url;
3369         }
3370         close $fd or return;
3371         return wantarray ? %remotes : \%remotes;
3372 }
3373
3374 # Takes a hash of remotes as first parameter and fills it by adding the
3375 # available remote heads for each of the indicated remotes.
3376 sub fill_remote_heads {
3377         my $remotes = shift;
3378         my @heads = map { "remotes/$_" } keys %$remotes;
3379         my @remoteheads = git_get_heads_list(undef, @heads);
3380         foreach my $remote (keys %$remotes) {
3381                 $remotes->{$remote}{'heads'} = [ grep {
3382                         $_->{'name'} =~ s!^$remote/!!
3383                         } @remoteheads ];
3384         }
3385 }
3386
3387 sub git_get_references {
3388         my $type = shift || "";
3389         my %refs;
3390         # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3391         # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3392         open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3393                 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3394                 or return;
3395
3396         while (my $line = <$fd>) {
3397                 chomp $line;
3398                 if ($line =~ m!^($oid_regex)\srefs/($type.*)$!) {
3399                         if (defined $refs{$1}) {
3400                                 push @{$refs{$1}}, $2;
3401                         } else {
3402                                 $refs{$1} = [ $2 ];
3403                         }
3404                 }
3405         }
3406         close $fd or return;
3407         return \%refs;
3408 }
3409
3410 sub git_get_rev_name_tags {
3411         my $hash = shift || return undef;
3412
3413         open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3414                 or return;
3415         my $name_rev = <$fd>;
3416         close $fd;
3417
3418         if ($name_rev =~ m|^$hash tags/(.*)$|) {
3419                 return $1;
3420         } else {
3421                 # catches also '$hash undefined' output
3422                 return undef;
3423         }
3424 }
3425
3426 ## ----------------------------------------------------------------------
3427 ## parse to hash functions
3428
3429 sub parse_date {
3430         my $epoch = shift;
3431         my $tz = shift || "-0000";
3432
3433         my %date;
3434         my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3435         my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3436         my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3437         $date{'hour'} = $hour;
3438         $date{'minute'} = $min;
3439         $date{'mday'} = $mday;
3440         $date{'day'} = $days[$wday];
3441         $date{'month'} = $months[$mon];
3442         $date{'rfc2822'}   = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3443                              $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3444         $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3445                              $mday, $months[$mon], $hour ,$min;
3446         $date{'iso-8601'}  = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3447                              1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3448
3449         my ($tz_sign, $tz_hour, $tz_min) =
3450                 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3451         $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3452         my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3453         ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3454         $date{'hour_local'} = $hour;
3455         $date{'minute_local'} = $min;
3456         $date{'tz_local'} = $tz;
3457         $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3458                                   1900+$year, $mon+1, $mday,
3459                                   $hour, $min, $sec, $tz);
3460         return %date;
3461 }
3462
3463 sub hide_mailaddrs_if_private {
3464         my $line = shift;
3465         return $line unless gitweb_check_feature('email-privacy');
3466         $line =~ s/<[^@>]+@[^>]+>/<redacted>/g;
3467         return $line;
3468 }
3469
3470 sub parse_tag {
3471         my $tag_id = shift;
3472         my %tag;
3473         my @comment;
3474
3475         open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3476         $tag{'id'} = $tag_id;
3477         while (my $line = <$fd>) {
3478                 chomp $line;
3479                 if ($line =~ m/^object ($oid_regex)$/) {
3480                         $tag{'object'} = $1;
3481                 } elsif ($line =~ m/^type (.+)$/) {
3482                         $tag{'type'} = $1;
3483                 } elsif ($line =~ m/^tag (.+)$/) {
3484                         $tag{'name'} = $1;
3485                 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3486                         $tag{'author'} = hide_mailaddrs_if_private($1);
3487                         $tag{'author_epoch'} = $2;
3488                         $tag{'author_tz'} = $3;
3489                         if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3490                                 $tag{'author_name'}  = $1;
3491                                 $tag{'author_email'} = $2;
3492                         } else {
3493                                 $tag{'author_name'} = $tag{'author'};
3494                         }
3495                 } elsif ($line =~ m/--BEGIN/) {
3496                         push @comment, $line;
3497                         last;
3498                 } elsif ($line eq "") {
3499                         last;
3500                 }
3501         }
3502         push @comment, <$fd>;
3503         $tag{'comment'} = \@comment;
3504         close $fd or return;
3505         if (!defined $tag{'name'}) {
3506                 return
3507         };
3508         return %tag
3509 }
3510
3511 sub parse_commit_text {
3512         my ($commit_text, $withparents) = @_;
3513         my @commit_lines = split '\n', $commit_text;
3514         my %co;
3515
3516         pop @commit_lines; # Remove '\0'
3517
3518         if (! @commit_lines) {
3519                 return;
3520         }
3521
3522         my $header = shift @commit_lines;
3523         if ($header !~ m/^$oid_regex/) {
3524                 return;
3525         }
3526         ($co{'id'}, my @parents) = split ' ', $header;
3527         while (my $line = shift @commit_lines) {
3528                 last if $line eq "\n";
3529                 if ($line =~ m/^tree ($oid_regex)$/) {
3530                         $co{'tree'} = $1;
3531                 } elsif ((!defined $withparents) && ($line =~ m/^parent ($oid_regex)$/)) {
3532                         push @parents, $1;
3533                 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3534                         $co{'author'} = hide_mailaddrs_if_private(to_utf8($1));
3535                         $co{'author_epoch'} = $2;
3536                         $co{'author_tz'} = $3;
3537                         if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3538                                 $co{'author_name'}  = $1;
3539                                 $co{'author_email'} = $2;
3540                         } else {
3541                                 $co{'author_name'} = $co{'author'};
3542                         }
3543                 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3544                         $co{'committer'} = hide_mailaddrs_if_private(to_utf8($1));
3545                         $co{'committer_epoch'} = $2;
3546                         $co{'committer_tz'} = $3;
3547                         if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3548                                 $co{'committer_name'}  = $1;
3549                                 $co{'committer_email'} = $2;
3550                         } else {
3551                                 $co{'committer_name'} = $co{'committer'};
3552                         }
3553                 }
3554         }
3555         if (!defined $co{'tree'}) {
3556                 return;
3557         };
3558         $co{'parents'} = \@parents;
3559         $co{'parent'} = $parents[0];
3560
3561         foreach my $title (@commit_lines) {
3562                 $title =~ s/^    //;
3563                 if ($title ne "") {
3564                         $co{'title'} = chop_str($title, 80, 5);
3565                         $co{'title_short'} = chop_str($title, 50, 5);
3566                         last;
3567                 }
3568         }
3569         if (! defined $co{'title'} || $co{'title'} eq "") {
3570                 $co{'title'} = $co{'title_short'} = '(no commit message)';
3571         }
3572         # remove added spaces, redact e-mail addresses if applicable.
3573         foreach my $line (@commit_lines) {
3574                 $line =~ s/^    //;
3575                 $line = hide_mailaddrs_if_private($line);
3576         }
3577         $co{'comment'} = \@commit_lines;
3578
3579         my $age = time - $co{'committer_epoch'};
3580         $co{'age'} = $age;
3581         $co{'age_string'} = age_string($age);
3582         my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3583         if ($age > 60*60*24*7*2) {
3584                 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3585                 $co{'age_string_age'} = $co{'age_string'};
3586         } else {
3587                 $co{'age_string_date'} = $co{'age_string'};
3588                 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3589         }
3590         return %co;
3591 }
3592
3593 sub parse_commit {
3594         my ($commit_id) = @_;
3595         my %co;
3596
3597         local $/ = "\0";
3598
3599         open my $fd, "-|", git_cmd(), "rev-list",
3600                 "--parents",
3601                 "--header",
3602                 "--max-count=1",
3603                 $commit_id,
3604                 "--",
3605                 or die_error(500, "Open git-rev-list failed");
3606         %co = parse_commit_text(<$fd>, 1);
3607         close $fd;
3608
3609         return %co;
3610 }
3611
3612 sub parse_commits {
3613         my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3614         my @cos;
3615
3616         $maxcount ||= 1;
3617         $skip ||= 0;
3618
3619         local $/ = "\0";
3620
3621         open my $fd, "-|", git_cmd(), "rev-list",
3622                 "--header",
3623                 @args,
3624                 ("--max-count=" . $maxcount),
3625                 ("--skip=" . $skip),
3626                 @extra_options,
3627                 $commit_id,
3628                 "--",
3629                 ($filename ? ($filename) : ())
3630                 or die_error(500, "Open git-rev-list failed");
3631         while (my $line = <$fd>) {
3632                 my %co = parse_commit_text($line);
3633                 push @cos, \%co;
3634         }
3635         close $fd;
3636
3637         return wantarray ? @cos : \@cos;
3638 }
3639
3640 # parse line of git-diff-tree "raw" output
3641 sub parse_difftree_raw_line {
3642         my $line = shift;
3643         my %res;
3644
3645         # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M   ls-files.c'
3646         # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M   rev-tree.c'
3647         if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ($oid_regex) ($oid_regex) (.)([0-9]{0,3})\t(.*)$/) {
3648                 $res{'from_mode'} = $1;
3649                 $res{'to_mode'} = $2;
3650                 $res{'from_id'} = $3;
3651                 $res{'to_id'} = $4;
3652                 $res{'status'} = $5;
3653                 $res{'similarity'} = $6;
3654                 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3655                         ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3656                 } else {
3657                         $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3658                 }
3659         }
3660         # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3661         # combined diff (for merge commit)
3662         elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:$oid_regex )+)([a-zA-Z]+)\t(.*)$//) {
3663                 $res{'nparents'}  = length($1);
3664                 $res{'from_mode'} = [ split(' ', $2) ];
3665                 $res{'to_mode'} = pop @{$res{'from_mode'}};
3666                 $res{'from_id'} = [ split(' ', $3) ];
3667                 $res{'to_id'} = pop @{$res{'from_id'}};
3668                 $res{'status'} = [ split('', $4) ];
3669                 $res{'to_file'} = unquote($5);
3670         }
3671         # 'c512b523472485aef4fff9e57b229d9d243c967f'
3672         elsif ($line =~ m/^($oid_regex)$/) {
3673                 $res{'commit'} = $1;
3674         }
3675
3676         return wantarray ? %res : \%res;
3677 }
3678
3679 # wrapper: return parsed line of git-diff-tree "raw" output
3680 # (the argument might be raw line, or parsed info)
3681 sub parsed_difftree_line {
3682         my $line_or_ref = shift;
3683
3684         if (ref($line_or_ref) eq "HASH") {
3685                 # pre-parsed (or generated by hand)
3686                 return $line_or_ref;
3687         } else {
3688                 return parse_difftree_raw_line($line_or_ref);
3689         }
3690 }
3691
3692 # parse line of git-ls-tree output
3693 sub parse_ls_tree_line {
3694         my $line = shift;
3695         my %opts = @_;
3696         my %res;
3697
3698         if ($opts{'-l'}) {
3699                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa   16717  panic.c'
3700                 $line =~ m/^([0-9]+) (.+) ($oid_regex) +(-|[0-9]+)\t(.+)$/s;
3701
3702                 $res{'mode'} = $1;
3703                 $res{'type'} = $2;
3704                 $res{'hash'} = $3;
3705                 $res{'size'} = $4;
3706                 if ($opts{'-z'}) {
3707                         $res{'name'} = $5;
3708                 } else {
3709                         $res{'name'} = unquote($5);
3710                 }
3711         } else {
3712                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
3713                 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t(.+)$/s;
3714
3715                 $res{'mode'} = $1;
3716                 $res{'type'} = $2;
3717                 $res{'hash'} = $3;
3718                 if ($opts{'-z'}) {
3719                         $res{'name'} = $4;
3720                 } else {
3721                         $res{'name'} = unquote($4);
3722                 }
3723         }
3724
3725         return wantarray ? %res : \%res;
3726 }
3727
3728 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3729 sub parse_from_to_diffinfo {
3730         my ($diffinfo, $from, $to, @parents) = @_;
3731
3732         if ($diffinfo->{'nparents'}) {
3733                 # combined diff
3734                 $from->{'file'} = [];
3735                 $from->{'href'} = [];
3736                 fill_from_file_info($diffinfo, @parents)
3737                         unless exists $diffinfo->{'from_file'};
3738                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3739                         $from->{'file'}[$i] =
3740                                 defined $diffinfo->{'from_file'}[$i] ?
3741                                         $diffinfo->{'from_file'}[$i] :
3742                                         $diffinfo->{'to_file'};
3743                         if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3744                                 $from->{'href'}[$i] = href(action=>"blob",
3745                                                            hash_base=>$parents[$i],
3746                                                            hash=>$diffinfo->{'from_id'}[$i],
3747                                                            file_name=>$from->{'file'}[$i]);
3748                         } else {
3749                                 $from->{'href'}[$i] = undef;
3750                         }
3751                 }
3752         } else {
3753                 # ordinary (not combined) diff
3754                 $from->{'file'} = $diffinfo->{'from_file'};
3755                 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3756                         $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3757                                                hash=>$diffinfo->{'from_id'},
3758                                                file_name=>$from->{'file'});
3759                 } else {
3760                         delete $from->{'href'};
3761                 }
3762         }
3763
3764         $to->{'file'} = $diffinfo->{'to_file'};
3765         if (!is_deleted($diffinfo)) { # file exists in result
3766                 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3767                                      hash=>$diffinfo->{'to_id'},
3768                                      file_name=>$to->{'file'});
3769         } else {
3770                 delete $to->{'href'};
3771         }
3772 }
3773
3774 ## ......................................................................
3775 ## parse to array of hashes functions
3776
3777 sub git_get_heads_list {
3778         my ($limit, @classes) = @_;
3779         @classes = get_branch_refs() unless @classes;
3780         my @patterns = map { "refs/$_" } @classes;
3781         my @headslist;
3782
3783         open my $fd, '-|', git_cmd(), 'for-each-ref',
3784                 ($limit ? '--count='.($limit+1) : ()),
3785                 '--sort=-HEAD', '--sort=-committerdate',
3786                 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3787                 @patterns
3788                 or return;
3789         while (my $line = <$fd>) {
3790                 my %ref_item;
3791
3792                 chomp $line;
3793                 my ($refinfo, $committerinfo) = split(/\0/, $line);
3794                 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3795                 my ($committer, $epoch, $tz) =
3796                         ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3797                 $ref_item{'fullname'}  = $name;
3798                 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
3799                 $name =~ s!^refs/($strip_refs|remotes)/!!;
3800                 $ref_item{'name'} = $name;
3801                 # for refs neither in 'heads' nor 'remotes' we want to
3802                 # show their ref dir
3803                 my $ref_dir = (defined $1) ? $1 : '';
3804                 if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
3805                     $ref_item{'name'} .= ' (' . $ref_dir . ')';
3806                 }
3807
3808                 $ref_item{'id'}    = $hash;
3809                 $ref_item{'title'} = $title || '(no commit message)';
3810                 $ref_item{'epoch'} = $epoch;
3811                 if ($epoch) {
3812                         $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3813                 } else {
3814                         $ref_item{'age'} = "unknown";
3815                 }
3816
3817                 push @headslist, \%ref_item;
3818         }
3819         close $fd;
3820
3821         return wantarray ? @headslist : \@headslist;
3822 }
3823
3824 sub git_get_tags_list {
3825         my $limit = shift;
3826         my @tagslist;
3827
3828         open my $fd, '-|', git_cmd(), 'for-each-ref',
3829                 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3830                 '--format=%(objectname) %(objecttype) %(refname) '.
3831                 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3832                 'refs/tags'
3833                 or return;
3834         while (my $line = <$fd>) {
3835                 my %ref_item;
3836
3837                 chomp $line;
3838                 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3839                 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3840                 my ($creator, $epoch, $tz) =
3841                         ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3842                 $ref_item{'fullname'} = $name;
3843                 $name =~ s!^refs/tags/!!;
3844
3845                 $ref_item{'type'} = $type;
3846                 $ref_item{'id'} = $id;
3847                 $ref_item{'name'} = $name;
3848                 if ($type eq "tag") {
3849                         $ref_item{'subject'} = $title;
3850                         $ref_item{'reftype'} = $reftype;
3851                         $ref_item{'refid'}   = $refid;
3852                 } else {
3853                         $ref_item{'reftype'} = $type;
3854                         $ref_item{'refid'}   = $id;
3855                 }
3856
3857                 if ($type eq "tag" || $type eq "commit") {
3858                         $ref_item{'epoch'} = $epoch;
3859                         if ($epoch) {
3860                                 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3861                         } else {
3862                                 $ref_item{'age'} = "unknown";
3863                         }
3864                 }
3865
3866                 push @tagslist, \%ref_item;
3867         }
3868         close $fd;
3869
3870         return wantarray ? @tagslist : \@tagslist;
3871 }
3872
3873 ## ----------------------------------------------------------------------
3874 ## filesystem-related functions
3875
3876 sub get_file_owner {
3877         my $path = shift;
3878
3879         my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3880         my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3881         if (!defined $gcos) {
3882                 return undef;
3883         }
3884         my $owner = $gcos;
3885         $owner =~ s/[,;].*$//;
3886         return to_utf8($owner);
3887 }
3888
3889 # assume that file exists
3890 sub insert_file {
3891         my $filename = shift;
3892
3893         open my $fd, '<', $filename;
3894         print map { to_utf8($_) } <$fd>;
3895         close $fd;
3896 }
3897
3898 ## ......................................................................
3899 ## mimetype related functions
3900
3901 sub mimetype_guess_file {
3902         my $filename = shift;
3903         my $mimemap = shift;
3904         -r $mimemap or return undef;
3905
3906         my %mimemap;
3907         open(my $mh, '<', $mimemap) or return undef;
3908         while (<$mh>) {
3909                 next if m/^#/; # skip comments
3910                 my ($mimetype, @exts) = split(/\s+/);
3911                 foreach my $ext (@exts) {
3912                         $mimemap{$ext} = $mimetype;
3913                 }
3914         }
3915         close($mh);
3916
3917         $filename =~ /\.([^.]*)$/;
3918         return $mimemap{$1};
3919 }
3920
3921 sub mimetype_guess {
3922         my $filename = shift;
3923         my $mime;
3924         $filename =~ /\./ or return undef;
3925
3926         if ($mimetypes_file) {
3927                 my $file = $mimetypes_file;
3928                 if ($file !~ m!^/!) { # if it is relative path
3929                         # it is relative to project
3930                         $file = "$projectroot/$project/$file";
3931                 }
3932                 $mime = mimetype_guess_file($filename, $file);
3933         }
3934         $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3935         return $mime;
3936 }
3937
3938 sub blob_mimetype {
3939         my $fd = shift;
3940         my $filename = shift;
3941
3942         if ($filename) {
3943                 my $mime = mimetype_guess($filename);
3944                 $mime and return $mime;
3945         }
3946
3947         # just in case
3948         return $default_blob_plain_mimetype unless $fd;
3949
3950         if (-T $fd) {
3951                 return 'text/plain';
3952         } elsif (! $filename) {
3953                 return 'application/octet-stream';
3954         } elsif ($filename =~ m/\.png$/i) {
3955                 return 'image/png';
3956         } elsif ($filename =~ m/\.gif$/i) {
3957                 return 'image/gif';
3958         } elsif ($filename =~ m/\.jpe?g$/i) {
3959                 return 'image/jpeg';
3960         } else {
3961                 return 'application/octet-stream';
3962         }
3963 }
3964
3965 sub blob_contenttype {
3966         my ($fd, $file_name, $type) = @_;
3967
3968         $type ||= blob_mimetype($fd, $file_name);
3969         if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3970                 $type .= "; charset=$default_text_plain_charset";
3971         }
3972
3973         return $type;
3974 }
3975
3976 # guess file syntax for syntax highlighting; return undef if no highlighting
3977 # the name of syntax can (in the future) depend on syntax highlighter used
3978 sub guess_file_syntax {
3979         my ($highlight, $file_name) = @_;
3980         return undef unless ($highlight && defined $file_name);
3981         my $basename = basename($file_name, '.in');
3982         return $highlight_basename{$basename}
3983                 if exists $highlight_basename{$basename};
3984
3985         $basename =~ /\.([^.]*)$/;
3986         my $ext = $1 or return undef;
3987         return $highlight_ext{$ext}
3988                 if exists $highlight_ext{$ext};
3989
3990         return undef;
3991 }
3992
3993 # run highlighter and return FD of its output,
3994 # or return original FD if no highlighting
3995 sub run_highlighter {
3996         my ($fd, $highlight, $syntax) = @_;
3997         return $fd unless ($highlight);
3998
3999         close $fd;
4000         my $syntax_arg = (defined $syntax) ? "--syntax $syntax" : "--force";
4001         open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
4002                   quote_command($^X, '-CO', '-MEncode=decode,FB_DEFAULT', '-pse',
4003                     '$_ = decode($fe, $_, FB_DEFAULT) if !utf8::decode($_);',
4004                     '--', "-fe=$fallback_encoding")." | ".
4005                   quote_command($highlight_bin).
4006                   " --replace-tabs=8 --fragment $syntax_arg |"
4007                 or die_error(500, "Couldn't open file or run syntax highlighter");
4008         return $fd;
4009 }
4010
4011 ## ======================================================================
4012 ## functions printing HTML: header, footer, error page
4013
4014 sub get_page_title {
4015         my $title = to_utf8($site_name);
4016
4017         unless (defined $project) {
4018                 if (defined $project_filter) {
4019                         $title .= " - projects in '" . esc_path($project_filter) . "'";
4020                 }
4021                 return $title;
4022         }
4023         $title .= " - " . to_utf8($project);
4024
4025         return $title unless (defined $action);
4026         $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
4027
4028         return $title unless (defined $file_name);
4029         $title .= " - " . esc_path($file_name);
4030         if ($action eq "tree" && $file_name !~ m|/$|) {
4031                 $title .= "/";
4032         }
4033
4034         return $title;
4035 }
4036
4037 sub get_content_type_html {
4038         # require explicit support from the UA if we are to send the page as
4039         # 'application/xhtml+xml', otherwise send it as plain old 'text/html'.
4040         # we have to do this because MSIE sometimes globs '*/*', pretending to
4041         # support xhtml+xml but choking when it gets what it asked for.
4042         if (defined $cgi->http('HTTP_ACCEPT') &&
4043             $cgi->http('HTTP_ACCEPT') =~ m/(,|;|\s|^)application\/xhtml\+xml(,|;|\s|$)/ &&
4044             $cgi->Accept('application/xhtml+xml') != 0) {
4045                 return 'application/xhtml+xml';
4046         } else {
4047                 return 'text/html';
4048         }
4049 }
4050
4051 sub print_feed_meta {
4052         if (defined $project) {
4053                 my %href_params = get_feed_info();
4054                 if (!exists $href_params{'-title'}) {
4055                         $href_params{'-title'} = 'log';
4056                 }
4057
4058                 foreach my $format (qw(RSS Atom)) {
4059                         my $type = lc($format);
4060                         my %link_attr = (
4061                                 '-rel' => 'alternate',
4062                                 '-title' => esc_attr("$project - $href_params{'-title'} - $format feed"),
4063                                 '-type' => "application/$type+xml"
4064                         );
4065
4066                         $href_params{'extra_options'} = undef;
4067                         $href_params{'action'} = $type;
4068                         $link_attr{'-href'} = esc_attr(href(%href_params));
4069                         print "<link ".
4070                               "rel=\"$link_attr{'-rel'}\" ".
4071                               "title=\"$link_attr{'-title'}\" ".
4072                               "href=\"$link_attr{'-href'}\" ".
4073                               "type=\"$link_attr{'-type'}\" ".
4074                               "/>\n";
4075
4076                         $href_params{'extra_options'} = '--no-merges';
4077                         $link_attr{'-href'} = esc_attr(href(%href_params));
4078                         $link_attr{'-title'} .= ' (no merges)';
4079                         print "<link ".
4080                               "rel=\"$link_attr{'-rel'}\" ".
4081                               "title=\"$link_attr{'-title'}\" ".
4082                               "href=\"$link_attr{'-href'}\" ".
4083                               "type=\"$link_attr{'-type'}\" ".
4084                               "/>\n";
4085                 }
4086
4087         } else {
4088                 printf('<link rel="alternate" title="%s projects list" '.
4089                        'href="%s" type="text/plain; charset=utf-8" />'."\n",
4090                        esc_attr($site_name),
4091                        esc_attr(href(project=>undef, action=>"project_index")));
4092                 printf('<link rel="alternate" title="%s projects feeds" '.
4093                        'href="%s" type="text/x-opml" />'."\n",
4094                        esc_attr($site_name),
4095                        esc_attr(href(project=>undef, action=>"opml")));
4096         }
4097 }
4098
4099 sub print_header_links {
4100         my $status = shift;
4101
4102         # print out each stylesheet that exist, providing backwards capability
4103         # for those people who defined $stylesheet in a config file
4104         if (defined $stylesheet) {
4105                 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4106         } else {
4107                 foreach my $stylesheet (@stylesheets) {
4108                         next unless $stylesheet;
4109                         print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4110                 }
4111         }
4112         print_feed_meta()
4113                 if ($status eq '200 OK');
4114         if (defined $favicon) {
4115                 print qq(<link rel="shortcut icon" href=").esc_url($favicon).qq(" type="image/png" />\n);
4116         }
4117 }
4118
4119 sub print_nav_breadcrumbs_path {
4120         my $dirprefix = undef;
4121         while (my $part = shift) {
4122                 $dirprefix .= "/" if defined $dirprefix;
4123                 $dirprefix .= $part;
4124                 print $cgi->a({-href => href(project => undef,
4125                                              project_filter => $dirprefix,
4126                                              action => "project_list")},
4127                               esc_html($part)) . " / ";
4128         }
4129 }
4130
4131 sub print_nav_breadcrumbs {
4132         my %opts = @_;
4133
4134         for my $crumb (@extra_breadcrumbs, [ $home_link_str => $home_link ]) {
4135                 print $cgi->a({-href => esc_url($crumb->[1])}, $crumb->[0]) . " / ";
4136         }
4137         if (defined $project) {
4138                 my @dirname = split '/', $project;
4139                 my $projectbasename = pop @dirname;
4140                 print_nav_breadcrumbs_path(@dirname);
4141                 print $cgi->a({-href => href(action=>"summary")}, esc_html($projectbasename));
4142                 if (defined $action) {
4143                         my $action_print = $action ;
4144                         if (defined $opts{-action_extra}) {
4145                                 $action_print = $cgi->a({-href => href(action=>$action)},
4146                                         $action);
4147                         }
4148                         print " / $action_print";
4149                 }
4150                 if (defined $opts{-action_extra}) {
4151                         print " / $opts{-action_extra}";
4152                 }
4153                 print "\n";
4154         } elsif (defined $project_filter) {
4155                 print_nav_breadcrumbs_path(split '/', $project_filter);
4156         }
4157 }
4158
4159 sub print_search_form {
4160         if (!defined $searchtext) {
4161                 $searchtext = "";
4162         }
4163         my $search_hash;
4164         if (defined $hash_base) {
4165                 $search_hash = $hash_base;
4166         } elsif (defined $hash) {
4167                 $search_hash = $hash;
4168         } else {
4169                 $search_hash = "HEAD";
4170         }
4171         my $action = $my_uri;
4172         my $use_pathinfo = gitweb_check_feature('pathinfo');
4173         if ($use_pathinfo) {
4174                 $action .= "/".esc_url($project);
4175         }
4176         print $cgi->start_form(-method => "get", -action => $action) .
4177               "<div class=\"search\">\n" .
4178               (!$use_pathinfo &&
4179               $cgi->input({-name=>"p", -value=>$project, -type=>"hidden"}) . "\n") .
4180               $cgi->input({-name=>"a", -value=>"search", -type=>"hidden"}) . "\n" .
4181               $cgi->input({-name=>"h", -value=>$search_hash, -type=>"hidden"}) . "\n" .
4182               $cgi->popup_menu(-name => 'st', -default => 'commit',
4183                                -values => ['commit', 'grep', 'author', 'committer', 'pickaxe']) .
4184               " " . $cgi->a({-href => href(action=>"search_help"),
4185                              -title => "search help" }, "?") . " search:\n",
4186               $cgi->textfield(-name => "s", -value => $searchtext, -override => 1) . "\n" .
4187               "<span title=\"Extended regular expression\">" .
4188               $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
4189                              -checked => $search_use_regexp) .
4190               "</span>" .
4191               "</div>" .
4192               $cgi->end_form() . "\n";
4193 }
4194
4195 sub git_header_html {
4196         my $status = shift || "200 OK";
4197         my $expires = shift;
4198         my %opts = @_;
4199
4200         my $title = get_page_title();
4201         print $cgi->header(-type=>get_content_type_html(), -charset => 'utf-8',
4202                            -status=> $status, -expires => $expires)
4203                 unless ($opts{'-no_http_header'});
4204         my $mod_perl_version = $ENV{'MOD_PERL'} ? " $ENV{'MOD_PERL'}" : '';
4205         print <<EOF;
4206 <?xml version="1.0" encoding="utf-8"?>
4207 <!DOCTYPE html [
4208         <!ENTITY nbsp "&#xA0;">
4209         <!ENTITY sdot "&#x22C5;">
4210 ]>
4211 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en-US" lang="en-US">
4212 <!-- git web interface version $version, (C) 2005-2006, Kay Sievers <kay.sievers\@vrfy.org>, Christian Gierke -->
4213 <!-- git core binaries version $git_version -->
4214 <head>
4215 <meta name="generator" content="gitweb/$version git/$git_version$mod_perl_version"/>
4216 <meta name="robots" content="index, nofollow"/>
4217 <title>$title</title>
4218 EOF
4219         # the stylesheet, favicon etc urls won't work correctly with path_info
4220         # unless we set the appropriate base URL
4221         if ($ENV{'PATH_INFO'}) {
4222                 print "<base href=\"".esc_url($base_url)."\" />\n";
4223         }
4224         print_header_links($status);
4225
4226         if (defined $site_html_head_string) {
4227                 print to_utf8($site_html_head_string);
4228         }
4229
4230         print "</head>\n" .
4231               "<body>\n";
4232
4233         if (defined $site_header && -f $site_header) {
4234                 insert_file($site_header);
4235         }
4236
4237         print "<div class=\"page_header\">\n";
4238         if (defined $logo) {
4239                 print $cgi->a({-href => esc_url($logo_url),
4240                                -title => $logo_label},
4241                               $cgi->img({-src => esc_url($logo),
4242                                          -width => 72, -height => 27,
4243                                          -alt => "git",
4244                                          -class => "logo"}));
4245         }
4246         print_nav_breadcrumbs(%opts);
4247         print "</div>\n";
4248
4249         my $have_search = gitweb_check_feature('search');
4250         if (defined $project && $have_search) {
4251                 print_search_form();
4252         }
4253 }
4254
4255 sub git_footer_html {
4256         my $feed_class = 'rss_logo';
4257
4258         print "<div class=\"page_footer\">\n";
4259         if (defined $project) {
4260                 my $descr = git_get_project_description($project);
4261                 if (defined $descr) {
4262                         print "<div class=\"page_footer_text\">" . esc_html($descr) . "</div>\n";
4263                 }
4264
4265                 my %href_params = get_feed_info();
4266                 if (!%href_params) {
4267                         $feed_class .= ' generic';
4268                 }
4269                 $href_params{'-title'} ||= 'log';
4270
4271                 foreach my $format (qw(RSS Atom)) {
4272                         $href_params{'action'} = lc($format);
4273                         print $cgi->a({-href => href(%href_params),
4274                                       -title => "$href_params{'-title'} $format feed",
4275                                       -class => $feed_class}, $format)."\n";
4276                 }
4277
4278         } else {
4279                 print $cgi->a({-href => href(project=>undef, action=>"opml",
4280                                              project_filter => $project_filter),
4281                               -class => $feed_class}, "OPML") . " ";
4282                 print $cgi->a({-href => href(project=>undef, action=>"project_index",
4283                                              project_filter => $project_filter),
4284                               -class => $feed_class}, "TXT") . "\n";
4285         }
4286         print "</div>\n"; # class="page_footer"
4287
4288         if (defined $t0 && gitweb_check_feature('timed')) {
4289                 print "<div id=\"generating_info\">\n";
4290                 print 'This page took '.
4291                       '<span id="generating_time" class="time_span">'.
4292                       tv_interval($t0, [ gettimeofday() ]).
4293                       ' seconds </span>'.
4294                       ' and '.
4295                       '<span id="generating_cmd">'.
4296                       $number_of_git_cmds.
4297                       '</span> git commands '.
4298                       " to generate.\n";
4299                 print "</div>\n"; # class="page_footer"
4300         }
4301
4302         if (defined $site_footer && -f $site_footer) {
4303                 insert_file($site_footer);
4304         }
4305
4306         print qq!<script type="text/javascript" src="!.esc_url($javascript).qq!"></script>\n!;
4307         if (defined $action &&
4308             $action eq 'blame_incremental') {
4309                 print qq!<script type="text/javascript">\n!.
4310                       qq!startBlame("!. esc_attr(href(action=>"blame_data", -replay=>1)) .qq!",\n!.
4311                       qq!           "!. esc_attr(href()) .qq!");\n!.
4312                       qq!</script>\n!;
4313         } else {
4314                 my ($jstimezone, $tz_cookie, $datetime_class) =
4315                         gitweb_get_feature('javascript-timezone');
4316
4317                 print qq!<script type="text/javascript">\n!.
4318                       qq!window.onload = function () {\n!;
4319                 if (gitweb_check_feature('javascript-actions')) {
4320                         print qq!       fixLinks();\n!;
4321                 }
4322                 if ($jstimezone && $tz_cookie && $datetime_class) {
4323                         print qq!       var tz_cookie = { name: '$tz_cookie', expires: 14, path: '/' };\n!. # in days
4324                               qq!       onloadTZSetup('$jstimezone', tz_cookie, '$datetime_class');\n!;
4325                 }
4326                 print qq!};\n!.
4327                       qq!</script>\n!;
4328         }
4329
4330         print "</body>\n" .
4331               "</html>";
4332 }
4333
4334 # die_error(<http_status_code>, <error_message>[, <detailed_html_description>])
4335 # Example: die_error(404, 'Hash not found')
4336 # By convention, use the following status codes (as defined in RFC 2616):
4337 # 400: Invalid or missing CGI parameters, or
4338 #      requested object exists but has wrong type.
4339 # 403: Requested feature (like "pickaxe" or "snapshot") not enabled on
4340 #      this server or project.
4341 # 404: Requested object/revision/project doesn't exist.
4342 # 500: The server isn't configured properly, or
4343 #      an internal error occurred (e.g. failed assertions caused by bugs), or
4344 #      an unknown error occurred (e.g. the git binary died unexpectedly).
4345 # 503: The server is currently unavailable (because it is overloaded,
4346 #      or down for maintenance).  Generally, this is a temporary state.
4347 sub die_error {
4348         my $status = shift || 500;
4349         my $error = esc_html(shift) || "Internal Server Error";
4350         my $extra = shift;
4351         my %opts = @_;
4352
4353         my %http_responses = (
4354                 400 => '400 Bad Request',
4355                 403 => '403 Forbidden',
4356                 404 => '404 Not Found',
4357                 500 => '500 Internal Server Error',
4358                 503 => '503 Service Unavailable',
4359         );
4360         git_header_html($http_responses{$status}, undef, %opts);
4361         print <<EOF;
4362 <div class="page_body">
4363 <br /><br />
4364 $status - $error
4365 <br />
4366 EOF
4367         if (defined $extra) {
4368                 print "<hr />\n" .
4369                       "$extra\n";
4370         }
4371         print "</div>\n";
4372
4373         git_footer_html();
4374         goto DONE_GITWEB
4375                 unless ($opts{'-error_handler'});
4376 }
4377
4378 ## ----------------------------------------------------------------------
4379 ## functions printing or outputting HTML: navigation
4380
4381 sub git_print_page_nav {
4382         my ($current, $suppress, $head, $treehead, $treebase, $extra) = @_;
4383         $extra = '' if !defined $extra; # pager or formats
4384
4385         my @navs = qw(summary shortlog log commit commitdiff tree);
4386         if ($suppress) {
4387                 @navs = grep { $_ ne $suppress } @navs;
4388         }
4389
4390         my %arg = map { $_ => {action=>$_} } @navs;
4391         if (defined $head) {
4392                 for (qw(commit commitdiff)) {
4393                         $arg{$_}{'hash'} = $head;
4394                 }
4395                 if ($current =~ m/^(tree | log | shortlog | commit | commitdiff | search)$/x) {
4396                         for (qw(shortlog log)) {
4397                                 $arg{$_}{'hash'} = $head;
4398                         }
4399                 }
4400         }
4401
4402         $arg{'tree'}{'hash'} = $treehead if defined $treehead;
4403         $arg{'tree'}{'hash_base'} = $treebase if defined $treebase;
4404
4405         my @actions = gitweb_get_feature('actions');
4406         my %repl = (
4407                 '%' => '%',
4408                 'n' => $project,         # project name
4409                 'f' => $git_dir,         # project path within filesystem
4410                 'h' => $treehead || '',  # current hash ('h' parameter)
4411                 'b' => $treebase || '',  # hash base ('hb' parameter)
4412         );
4413         while (@actions) {
4414                 my ($label, $link, $pos) = splice(@actions,0,3);
4415                 # insert
4416                 @navs = map { $_ eq $pos ? ($_, $label) : $_ } @navs;
4417                 # munch munch
4418                 $link =~ s/%([%nfhb])/$repl{$1}/g;
4419                 $arg{$label}{'_href'} = $link;
4420         }
4421
4422         print "<div class=\"page_nav\">\n" .
4423                 (join " | ",
4424                  map { $_ eq $current ?
4425                        $_ : $cgi->a({-href => ($arg{$_}{_href} ? $arg{$_}{_href} : href(%{$arg{$_}}))}, "$_")
4426                  } @navs);
4427         print "<br/>\n$extra<br/>\n" .
4428               "</div>\n";
4429 }
4430
4431 # returns a submenu for the navigation of the refs views (tags, heads,
4432 # remotes) with the current view disabled and the remotes view only
4433 # available if the feature is enabled
4434 sub format_ref_views {
4435         my ($current) = @_;
4436         my @ref_views = qw{tags heads};
4437         push @ref_views, 'remotes' if gitweb_check_feature('remote_heads');
4438         return join " | ", map {
4439                 $_ eq $current ? $_ :
4440                 $cgi->a({-href => href(action=>$_)}, $_)
4441         } @ref_views
4442 }
4443
4444 sub format_paging_nav {
4445         my ($action, $page, $has_next_link) = @_;
4446         my $paging_nav;
4447
4448
4449         if ($page > 0) {
4450                 $paging_nav .=
4451                         $cgi->a({-href => href(-replay=>1, page=>undef)}, "first") .
4452                         " &sdot; " .
4453                         $cgi->a({-href => href(-replay=>1, page=>$page-1),
4454                                  -accesskey => "p", -title => "Alt-p"}, "prev");
4455         } else {
4456                 $paging_nav .= "first &sdot; prev";
4457         }
4458
4459         if ($has_next_link) {
4460                 $paging_nav .= " &sdot; " .
4461                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
4462                                  -accesskey => "n", -title => "Alt-n"}, "next");
4463         } else {
4464                 $paging_nav .= " &sdot; next";
4465         }
4466
4467         return $paging_nav;
4468 }
4469
4470 ## ......................................................................
4471 ## functions printing or outputting HTML: div
4472
4473 sub git_print_header_div {
4474         my ($action, $title, $hash, $hash_base) = @_;
4475         my %args = ();
4476
4477         $args{'action'} = $action;
4478         $args{'hash'} = $hash if $hash;
4479         $args{'hash_base'} = $hash_base if $hash_base;
4480
4481         print "<div class=\"header\">\n" .
4482               $cgi->a({-href => href(%args), -class => "title"},
4483               $title ? $title : $action) .
4484               "\n</div>\n";
4485 }
4486
4487 sub format_repo_url {
4488         my ($name, $url) = @_;
4489         return "<tr class=\"metadata_url\"><td>$name</td><td>$url</td></tr>\n";
4490 }
4491
4492 # Group output by placing it in a DIV element and adding a header.
4493 # Options for start_div() can be provided by passing a hash reference as the
4494 # first parameter to the function.
4495 # Options to git_print_header_div() can be provided by passing an array
4496 # reference. This must follow the options to start_div if they are present.
4497 # The content can be a scalar, which is output as-is, a scalar reference, which
4498 # is output after html escaping, an IO handle passed either as *handle or
4499 # *handle{IO}, or a function reference. In the latter case all following
4500 # parameters will be taken as argument to the content function call.
4501 sub git_print_section {
4502         my ($div_args, $header_args, $content);
4503         my $arg = shift;
4504         if (ref($arg) eq 'HASH') {
4505                 $div_args = $arg;
4506                 $arg = shift;
4507         }
4508         if (ref($arg) eq 'ARRAY') {
4509                 $header_args = $arg;
4510                 $arg = shift;
4511         }
4512         $content = $arg;
4513
4514         print $cgi->start_div($div_args);
4515         git_print_header_div(@$header_args);
4516
4517         if (ref($content) eq 'CODE') {
4518                 $content->(@_);
4519         } elsif (ref($content) eq 'SCALAR') {
4520                 print esc_html($$content);
4521         } elsif (ref($content) eq 'GLOB' or ref($content) eq 'IO::Handle') {
4522                 print <$content>;
4523         } elsif (!ref($content) && defined($content)) {
4524                 print $content;
4525         }
4526
4527         print $cgi->end_div;
4528 }
4529
4530 sub format_timestamp_html {
4531         my $date = shift;
4532         my $strtime = $date->{'rfc2822'};
4533
4534         my (undef, undef, $datetime_class) =
4535                 gitweb_get_feature('javascript-timezone');
4536         if ($datetime_class) {
4537                 $strtime = qq!<span class="$datetime_class">$strtime</span>!;
4538         }
4539
4540         my $localtime_format = '(%02d:%02d %s)';
4541         if ($date->{'hour_local'} < 6) {
4542                 $localtime_format = '(<span class="atnight">%02d:%02d</span> %s)';
4543         }
4544         $strtime .= ' ' .
4545                     sprintf($localtime_format,
4546                             $date->{'hour_local'}, $date->{'minute_local'}, $date->{'tz_local'});
4547
4548         return $strtime;
4549 }
4550
4551 # Outputs the author name and date in long form
4552 sub git_print_authorship {
4553         my $co = shift;
4554         my %opts = @_;
4555         my $tag = $opts{-tag} || 'div';
4556         my $author = $co->{'author_name'};
4557
4558         my %ad = parse_date($co->{'author_epoch'}, $co->{'author_tz'});
4559         print "<$tag class=\"author_date\">" .
4560               format_search_author($author, "author", esc_html($author)) .
4561               " [".format_timestamp_html(\%ad)."]".
4562               git_get_avatar($co->{'author_email'}, -pad_before => 1) .
4563               "</$tag>\n";
4564 }
4565
4566 # Outputs table rows containing the full author or committer information,
4567 # in the format expected for 'commit' view (& similar).
4568 # Parameters are a commit hash reference, followed by the list of people
4569 # to output information for. If the list is empty it defaults to both
4570 # author and committer.
4571 sub git_print_authorship_rows {
4572         my $co = shift;
4573         # too bad we can't use @people = @_ || ('author', 'committer')
4574         my @people = @_;
4575         @people = ('author', 'committer') unless @people;
4576         foreach my $who (@people) {
4577                 my %wd = parse_date($co->{"${who}_epoch"}, $co->{"${who}_tz"});
4578                 print "<tr><td>$who</td><td>" .
4579                       format_search_author($co->{"${who}_name"}, $who,
4580                                            esc_html($co->{"${who}_name"})) . " " .
4581                       format_search_author($co->{"${who}_email"}, $who,
4582                                            esc_html("<" . $co->{"${who}_email"} . ">")) .
4583                       "</td><td rowspan=\"2\">" .
4584                       git_get_avatar($co->{"${who}_email"}, -size => 'double') .
4585                       "</td></tr>\n" .
4586                       "<tr>" .
4587                       "<td></td><td>" .
4588                       format_timestamp_html(\%wd) .
4589                       "</td>" .
4590                       "</tr>\n";
4591         }
4592 }
4593
4594 sub git_print_page_path {
4595         my $name = shift;
4596         my $type = shift;
4597         my $hb = shift;
4598
4599
4600         print "<div class=\"page_path\">";
4601         print $cgi->a({-href => href(action=>"tree", hash_base=>$hb),
4602                       -title => 'tree root'}, to_utf8("[$project]"));
4603         print " / ";
4604         if (defined $name) {
4605                 my @dirname = split '/', $name;
4606                 my $basename = pop @dirname;
4607                 my $fullname = '';
4608
4609                 foreach my $dir (@dirname) {
4610                         $fullname .= ($fullname ? '/' : '') . $dir;
4611                         print $cgi->a({-href => href(action=>"tree", file_name=>$fullname,
4612                                                      hash_base=>$hb),
4613                                       -title => $fullname}, esc_path($dir));
4614                         print " / ";
4615                 }
4616                 if (defined $type && $type eq 'blob') {
4617                         print $cgi->a({-href => href(action=>"blob_plain", file_name=>$file_name,
4618                                                      hash_base=>$hb),
4619                                       -title => $name}, esc_path($basename));
4620                 } elsif (defined $type && $type eq 'tree') {
4621                         print $cgi->a({-href => href(action=>"tree", file_name=>$file_name,
4622                                                      hash_base=>$hb),
4623                                       -title => $name}, esc_path($basename));
4624                         print " / ";
4625                 } else {
4626                         print esc_path($basename);
4627                 }
4628         }
4629         print "<br/></div>\n";
4630 }
4631
4632 sub git_print_log {
4633         my $log = shift;
4634         my %opts = @_;
4635
4636         if ($opts{'-remove_title'}) {
4637                 # remove title, i.e. first line of log
4638                 shift @$log;
4639         }
4640         # remove leading empty lines
4641         while (defined $log->[0] && $log->[0] eq "") {
4642                 shift @$log;
4643         }
4644
4645         # print log
4646         my $skip_blank_line = 0;
4647         foreach my $line (@$log) {
4648                 if ($line =~ m/^\s*([A-Z][-A-Za-z]*-([Bb]y|[Tt]o)|C[Cc]|(Clos|Fix)es): /) {
4649                         if (! $opts{'-remove_signoff'}) {
4650                                 print "<span class=\"signoff\">" . esc_html($line) . "</span><br/>\n";
4651                                 $skip_blank_line = 1;
4652                         }
4653                         next;
4654                 }
4655
4656                 if ($line =~ m,\s*([a-z]*link): (https?://\S+),i) {
4657                         if (! $opts{'-remove_signoff'}) {
4658                                 print "<span class=\"signoff\">" . esc_html($1) . ": " .
4659                                         "<a href=\"" . esc_html($2) . "\">" . esc_html($2) . "</a>" .
4660                                         "</span><br/>\n";
4661                                 $skip_blank_line = 1;
4662                         }
4663                         next;
4664                 }
4665
4666                 # print only one empty line
4667                 # do not print empty line after signoff
4668                 if ($line eq "") {
4669                         next if ($skip_blank_line);
4670                         $skip_blank_line = 1;
4671                 } else {
4672                         $skip_blank_line = 0;
4673                 }
4674
4675                 print format_log_line_html($line) . "<br/>\n";
4676         }
4677
4678         if ($opts{'-final_empty_line'}) {
4679                 # end with single empty line
4680                 print "<br/>\n" unless $skip_blank_line;
4681         }
4682 }
4683
4684 # return link target (what link points to)
4685 sub git_get_link_target {
4686         my $hash = shift;
4687         my $link_target;
4688
4689         # read link
4690         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
4691                 or return;
4692         {
4693                 local $/ = undef;
4694                 $link_target = <$fd>;
4695         }
4696         close $fd
4697                 or return;
4698
4699         return $link_target;
4700 }
4701
4702 # given link target, and the directory (basedir) the link is in,
4703 # return target of link relative to top directory (top tree);
4704 # return undef if it is not possible (including absolute links).
4705 sub normalize_link_target {
4706         my ($link_target, $basedir) = @_;
4707
4708         # absolute symlinks (beginning with '/') cannot be normalized
4709         return if (substr($link_target, 0, 1) eq '/');
4710
4711         # normalize link target to path from top (root) tree (dir)
4712         my $path;
4713         if ($basedir) {
4714                 $path = $basedir . '/' . $link_target;
4715         } else {
4716                 # we are in top (root) tree (dir)
4717                 $path = $link_target;
4718         }
4719
4720         # remove //, /./, and /../
4721         my @path_parts;
4722         foreach my $part (split('/', $path)) {
4723                 # discard '.' and ''
4724                 next if (!$part || $part eq '.');
4725                 # handle '..'
4726                 if ($part eq '..') {
4727                         if (@path_parts) {
4728                                 pop @path_parts;
4729                         } else {
4730                                 # link leads outside repository (outside top dir)
4731                                 return;
4732                         }
4733                 } else {
4734                         push @path_parts, $part;
4735                 }
4736         }
4737         $path = join('/', @path_parts);
4738
4739         return $path;
4740 }
4741
4742 # print tree entry (row of git_tree), but without encompassing <tr> element
4743 sub git_print_tree_entry {
4744         my ($t, $basedir, $hash_base, $have_blame) = @_;
4745
4746         my %base_key = ();
4747         $base_key{'hash_base'} = $hash_base if defined $hash_base;
4748
4749         # The format of a table row is: mode list link.  Where mode is
4750         # the mode of the entry, list is the name of the entry, an href,
4751         # and link is the action links of the entry.
4752
4753         print "<td class=\"mode\">" . mode_str($t->{'mode'}) . "</td>\n";
4754         if (exists $t->{'size'}) {
4755                 print "<td class=\"size\">$t->{'size'}</td>\n";
4756         }
4757         if ($t->{'type'} eq "blob") {
4758                 print "<td class=\"list\">" .
4759                         $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4760                                                file_name=>"$basedir$t->{'name'}", %base_key),
4761                                 -class => "list"}, esc_path($t->{'name'}));
4762                 if (S_ISLNK(oct $t->{'mode'})) {
4763                         my $link_target = git_get_link_target($t->{'hash'});
4764                         if ($link_target) {
4765                                 my $norm_target = normalize_link_target($link_target, $basedir);
4766                                 if (defined $norm_target) {
4767                                         print " -> " .
4768                                               $cgi->a({-href => href(action=>"object", hash_base=>$hash_base,
4769                                                                      file_name=>$norm_target),
4770                                                        -title => $norm_target}, esc_path($link_target));
4771                                 } else {
4772                                         print " -> " . esc_path($link_target);
4773                                 }
4774                         }
4775                 }
4776                 print "</td>\n";
4777                 print "<td class=\"link\">";
4778                 print $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4779                                              file_name=>"$basedir$t->{'name'}", %base_key)},
4780                               "blob");
4781                 if ($have_blame) {
4782                         print " | " .
4783                               $cgi->a({-href => href(action=>"blame", hash=>$t->{'hash'},
4784                                                      file_name=>"$basedir$t->{'name'}", %base_key)},
4785                                       "blame");
4786                 }
4787                 if (defined $hash_base) {
4788                         print " | " .
4789                               $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4790                                                      hash=>$t->{'hash'}, file_name=>"$basedir$t->{'name'}")},
4791                                       "history");
4792                 }
4793                 print " | " .
4794                         $cgi->a({-href => href(action=>"blob_plain", hash_base=>$hash_base,
4795                                                file_name=>"$basedir$t->{'name'}")},
4796                                 "raw");
4797                 print "</td>\n";
4798
4799         } elsif ($t->{'type'} eq "tree") {
4800                 print "<td class=\"list\">";
4801                 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4802                                              file_name=>"$basedir$t->{'name'}",
4803                                              %base_key)},
4804                               esc_path($t->{'name'}));
4805                 print "</td>\n";
4806                 print "<td class=\"link\">";
4807                 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4808                                              file_name=>"$basedir$t->{'name'}",
4809                                              %base_key)},
4810                               "tree");
4811                 if (defined $hash_base) {
4812                         print " | " .
4813                               $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4814                                                      file_name=>"$basedir$t->{'name'}")},
4815                                       "history");
4816                 }
4817                 print "</td>\n";
4818         } else {
4819                 # unknown object: we can only present history for it
4820                 # (this includes 'commit' object, i.e. submodule support)
4821                 print "<td class=\"list\">" .
4822                       esc_path($t->{'name'}) .
4823                       "</td>\n";
4824                 print "<td class=\"link\">";
4825                 if (defined $hash_base) {
4826                         print $cgi->a({-href => href(action=>"history",
4827                                                      hash_base=>$hash_base,
4828                                                      file_name=>"$basedir$t->{'name'}")},
4829                                       "history");
4830                 }
4831                 print "</td>\n";
4832         }
4833 }
4834
4835 ## ......................................................................
4836 ## functions printing large fragments of HTML
4837
4838 # get pre-image filenames for merge (combined) diff
4839 sub fill_from_file_info {
4840         my ($diff, @parents) = @_;
4841
4842         $diff->{'from_file'} = [ ];
4843         $diff->{'from_file'}[$diff->{'nparents'} - 1] = undef;
4844         for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4845                 if ($diff->{'status'}[$i] eq 'R' ||
4846                     $diff->{'status'}[$i] eq 'C') {
4847                         $diff->{'from_file'}[$i] =
4848                                 git_get_path_by_hash($parents[$i], $diff->{'from_id'}[$i]);
4849                 }
4850         }
4851
4852         return $diff;
4853 }
4854
4855 # is current raw difftree line of file deletion
4856 sub is_deleted {
4857         my $diffinfo = shift;
4858
4859         return $diffinfo->{'to_id'} eq ('0' x 40) || $diffinfo->{'to_id'} eq ('0' x 64);
4860 }
4861
4862 # does patch correspond to [previous] difftree raw line
4863 # $diffinfo  - hashref of parsed raw diff format
4864 # $patchinfo - hashref of parsed patch diff format
4865 #              (the same keys as in $diffinfo)
4866 sub is_patch_split {
4867         my ($diffinfo, $patchinfo) = @_;
4868
4869         return defined $diffinfo && defined $patchinfo
4870                 && $diffinfo->{'to_file'} eq $patchinfo->{'to_file'};
4871 }
4872
4873
4874 sub git_difftree_body {
4875         my ($difftree, $hash, @parents) = @_;
4876         my ($parent) = $parents[0];
4877         my $have_blame = gitweb_check_feature('blame');
4878         print "<div class=\"list_head\">\n";
4879         if ($#{$difftree} > 10) {
4880                 print(($#{$difftree} + 1) . " files changed:\n");
4881         }
4882         print "</div>\n";
4883
4884         print "<table class=\"" .
4885               (@parents > 1 ? "combined " : "") .
4886               "diff_tree\">\n";
4887
4888         # header only for combined diff in 'commitdiff' view
4889         my $has_header = @$difftree && @parents > 1 && $action eq 'commitdiff';
4890         if ($has_header) {
4891                 # table header
4892                 print "<thead><tr>\n" .
4893                        "<th></th><th></th>\n"; # filename, patchN link
4894                 for (my $i = 0; $i < @parents; $i++) {
4895                         my $par = $parents[$i];
4896                         print "<th>" .
4897                               $cgi->a({-href => href(action=>"commitdiff",
4898                                                      hash=>$hash, hash_parent=>$par),
4899                                        -title => 'commitdiff to parent number ' .
4900                                                   ($i+1) . ': ' . substr($par,0,7)},
4901                                       $i+1) .
4902                               "&nbsp;</th>\n";
4903                 }
4904                 print "</tr></thead>\n<tbody>\n";
4905         }
4906
4907         my $alternate = 1;
4908         my $patchno = 0;
4909         foreach my $line (@{$difftree}) {
4910                 my $diff = parsed_difftree_line($line);
4911
4912                 if ($alternate) {
4913                         print "<tr class=\"dark\">\n";
4914                 } else {
4915                         print "<tr class=\"light\">\n";
4916                 }
4917                 $alternate ^= 1;
4918
4919                 if (exists $diff->{'nparents'}) { # combined diff
4920
4921                         fill_from_file_info($diff, @parents)
4922                                 unless exists $diff->{'from_file'};
4923
4924                         if (!is_deleted($diff)) {
4925                                 # file exists in the result (child) commit
4926                                 print "<td>" .
4927                                       $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
4928                                                              file_name=>$diff->{'to_file'},
4929                                                              hash_base=>$hash),
4930                                               -class => "list"}, esc_path($diff->{'to_file'})) .
4931                                       "</td>\n";
4932                         } else {
4933                                 print "<td>" .
4934                                       esc_path($diff->{'to_file'}) .
4935                                       "</td>\n";
4936                         }
4937
4938                         if ($action eq 'commitdiff') {
4939                                 # link to patch
4940                                 $patchno++;
4941                                 print "<td class=\"link\">" .
4942                                       $cgi->a({-href => href(-anchor=>"patch$patchno")},
4943                                               "patch") .
4944                                       " | " .
4945                                       "</td>\n";
4946                         }
4947
4948                         my $has_history = 0;
4949                         my $not_deleted = 0;
4950                         for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4951                                 my $hash_parent = $parents[$i];
4952                                 my $from_hash = $diff->{'from_id'}[$i];
4953                                 my $from_path = $diff->{'from_file'}[$i];
4954                                 my $status = $diff->{'status'}[$i];
4955
4956                                 $has_history ||= ($status ne 'A');
4957                                 $not_deleted ||= ($status ne 'D');
4958
4959                                 if ($status eq 'A') {
4960                                         print "<td  class=\"link\" align=\"right\"> | </td>\n";
4961                                 } elsif ($status eq 'D') {
4962                                         print "<td class=\"link\">" .
4963                                               $cgi->a({-href => href(action=>"blob",
4964                                                                      hash_base=>$hash,
4965                                                                      hash=>$from_hash,
4966                                                                      file_name=>$from_path)},
4967                                                       "blob" . ($i+1)) .
4968                                               " | </td>\n";
4969                                 } else {
4970                                         if ($diff->{'to_id'} eq $from_hash) {
4971                                                 print "<td class=\"link nochange\">";
4972                                         } else {
4973                                                 print "<td class=\"link\">";
4974                                         }
4975                                         print $cgi->a({-href => href(action=>"blobdiff",
4976                                                                      hash=>$diff->{'to_id'},
4977                                                                      hash_parent=>$from_hash,
4978                                                                      hash_base=>$hash,
4979                                                                      hash_parent_base=>$hash_parent,
4980                                                                      file_name=>$diff->{'to_file'},
4981                                                                      file_parent=>$from_path)},
4982                                                       "diff" . ($i+1)) .
4983                                               " | </td>\n";
4984                                 }
4985                         }
4986
4987                         print "<td class=\"link\">";
4988                         if ($not_deleted) {
4989                                 print $cgi->a({-href => href(action=>"blob",
4990                                                              hash=>$diff->{'to_id'},
4991                                                              file_name=>$diff->{'to_file'},
4992                                                              hash_base=>$hash)},
4993                                               "blob");
4994                                 print " | " if ($has_history);
4995                         }
4996                         if ($has_history) {
4997                                 print $cgi->a({-href => href(action=>"history",
4998                                                              file_name=>$diff->{'to_file'},
4999                                                              hash_base=>$hash)},
5000                                               "history");
5001                         }
5002                         print "</td>\n";
5003
5004                         print "</tr>\n";
5005                         next; # instead of 'else' clause, to avoid extra indent
5006                 }
5007                 # else ordinary diff
5008
5009                 my ($to_mode_oct, $to_mode_str, $to_file_type);
5010                 my ($from_mode_oct, $from_mode_str, $from_file_type);
5011                 if ($diff->{'to_mode'} ne ('0' x 6)) {
5012                         $to_mode_oct = oct $diff->{'to_mode'};
5013                         if (S_ISREG($to_mode_oct)) { # only for regular file
5014                                 $to_mode_str = sprintf("%04o", $to_mode_oct & 0777); # permission bits
5015                         }
5016                         $to_file_type = file_type($diff->{'to_mode'});
5017                 }
5018                 if ($diff->{'from_mode'} ne ('0' x 6)) {
5019                         $from_mode_oct = oct $diff->{'from_mode'};
5020                         if (S_ISREG($from_mode_oct)) { # only for regular file
5021                                 $from_mode_str = sprintf("%04o", $from_mode_oct & 0777); # permission bits
5022                         }
5023                         $from_file_type = file_type($diff->{'from_mode'});
5024                 }
5025
5026                 if ($diff->{'status'} eq "A") { # created
5027                         my $mode_chng = "<span class=\"file_status new\">[new $to_file_type";
5028                         $mode_chng   .= " with mode: $to_mode_str" if $to_mode_str;
5029                         $mode_chng   .= "]</span>";
5030                         print "<td>";
5031                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5032                                                      hash_base=>$hash, file_name=>$diff->{'file'}),
5033                                       -class => "list"}, esc_path($diff->{'file'}));
5034                         print "</td>\n";
5035                         print "<td>$mode_chng</td>\n";
5036                         print "<td class=\"link\">";
5037                         if ($action eq 'commitdiff') {
5038                                 # link to patch
5039                                 $patchno++;
5040                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5041                                               "patch") .
5042                                       " | ";
5043                         }
5044                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5045                                                      hash_base=>$hash, file_name=>$diff->{'file'})},
5046                                       "blob");
5047                         print "</td>\n";
5048
5049                 } elsif ($diff->{'status'} eq "D") { # deleted
5050                         my $mode_chng = "<span class=\"file_status deleted\">[deleted $from_file_type]</span>";
5051                         print "<td>";
5052                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'from_id'},
5053                                                      hash_base=>$parent, file_name=>$diff->{'file'}),
5054                                        -class => "list"}, esc_path($diff->{'file'}));
5055                         print "</td>\n";
5056                         print "<td>$mode_chng</td>\n";
5057                         print "<td class=\"link\">";
5058                         if ($action eq 'commitdiff') {
5059                                 # link to patch
5060                                 $patchno++;
5061                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5062                                               "patch") .
5063                                       " | ";
5064                         }
5065                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'from_id'},
5066                                                      hash_base=>$parent, file_name=>$diff->{'file'})},
5067                                       "blob") . " | ";
5068                         if ($have_blame) {
5069                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$parent,
5070                                                              file_name=>$diff->{'file'})},
5071                                               "blame") . " | ";
5072                         }
5073                         print $cgi->a({-href => href(action=>"history", hash_base=>$parent,
5074                                                      file_name=>$diff->{'file'})},
5075                                       "history");
5076                         print "</td>\n";
5077
5078                 } elsif ($diff->{'status'} eq "M" || $diff->{'status'} eq "T") { # modified, or type changed
5079                         my $mode_chnge = "";
5080                         if ($diff->{'from_mode'} != $diff->{'to_mode'}) {
5081                                 $mode_chnge = "<span class=\"file_status mode_chnge\">[changed";
5082                                 if ($from_file_type ne $to_file_type) {
5083                                         $mode_chnge .= " from $from_file_type to $to_file_type";
5084                                 }
5085                                 if (($from_mode_oct & 0777) != ($to_mode_oct & 0777)) {
5086                                         if ($from_mode_str && $to_mode_str) {
5087                                                 $mode_chnge .= " mode: $from_mode_str->$to_mode_str";
5088                                         } elsif ($to_mode_str) {
5089                                                 $mode_chnge .= " mode: $to_mode_str";
5090                                         }
5091                                 }
5092                                 $mode_chnge .= "]</span>\n";
5093                         }
5094                         print "<td>";
5095                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5096                                                      hash_base=>$hash, file_name=>$diff->{'file'}),
5097                                       -class => "list"}, esc_path($diff->{'file'}));
5098                         print "</td>\n";
5099                         print "<td>$mode_chnge</td>\n";
5100                         print "<td class=\"link\">";
5101                         if ($action eq 'commitdiff') {
5102                                 # link to patch
5103                                 $patchno++;
5104                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5105                                               "patch") .
5106                                       " | ";
5107                         } elsif ($diff->{'to_id'} ne $diff->{'from_id'}) {
5108                                 # "commit" view and modified file (not onlu mode changed)
5109                                 print $cgi->a({-href => href(action=>"blobdiff",
5110                                                              hash=>$diff->{'to_id'}, hash_parent=>$diff->{'from_id'},
5111                                                              hash_base=>$hash, hash_parent_base=>$parent,
5112                                                              file_name=>$diff->{'file'})},
5113                                               "diff") .
5114                                       " | ";
5115                         }
5116                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5117                                                      hash_base=>$hash, file_name=>$diff->{'file'})},
5118                                        "blob") . " | ";
5119                         if ($have_blame) {
5120                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$hash,
5121                                                              file_name=>$diff->{'file'})},
5122                                               "blame") . " | ";
5123                         }
5124                         print $cgi->a({-href => href(action=>"history", hash_base=>$hash,
5125                                                      file_name=>$diff->{'file'})},
5126                                       "history");
5127                         print "</td>\n";
5128
5129                 } elsif ($diff->{'status'} eq "R" || $diff->{'status'} eq "C") { # renamed or copied
5130                         my %status_name = ('R' => 'moved', 'C' => 'copied');
5131                         my $nstatus = $status_name{$diff->{'status'}};
5132                         my $mode_chng = "";
5133                         if ($diff->{'from_mode'} != $diff->{'to_mode'}) {
5134                                 # mode also for directories, so we cannot use $to_mode_str
5135                                 $mode_chng = sprintf(", mode: %04o", $to_mode_oct & 0777);
5136                         }
5137                         print "<td>" .
5138                               $cgi->a({-href => href(action=>"blob", hash_base=>$hash,
5139                                                      hash=>$diff->{'to_id'}, file_name=>$diff->{'to_file'}),
5140                                       -class => "list"}, esc_path($diff->{'to_file'})) . "</td>\n" .
5141                               "<td><span class=\"file_status $nstatus\">[$nstatus from " .
5142                               $cgi->a({-href => href(action=>"blob", hash_base=>$parent,
5143                                                      hash=>$diff->{'from_id'}, file_name=>$diff->{'from_file'}),
5144                                       -class => "list"}, esc_path($diff->{'from_file'})) .
5145                               " with " . (int $diff->{'similarity'}) . "% similarity$mode_chng]</span></td>\n" .
5146                               "<td class=\"link\">";
5147                         if ($action eq 'commitdiff') {
5148                                 # link to patch
5149                                 $patchno++;
5150                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5151                                               "patch") .
5152                                       " | ";
5153                         } elsif ($diff->{'to_id'} ne $diff->{'from_id'}) {
5154                                 # "commit" view and modified file (not only pure rename or copy)
5155                                 print $cgi->a({-href => href(action=>"blobdiff",
5156                                                              hash=>$diff->{'to_id'}, hash_parent=>$diff->{'from_id'},
5157                                                              hash_base=>$hash, hash_parent_base=>$parent,
5158                                                              file_name=>$diff->{'to_file'}, file_parent=>$diff->{'from_file'})},
5159                                               "diff") .
5160                                       " | ";
5161                         }
5162                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5163                                                      hash_base=>$parent, file_name=>$diff->{'to_file'})},
5164                                       "blob") . " | ";
5165                         if ($have_blame) {
5166                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$hash,
5167                                                              file_name=>$diff->{'to_file'})},
5168                                               "blame") . " | ";
5169                         }
5170                         print $cgi->a({-href => href(action=>"history", hash_base=>$hash,
5171                                                     file_name=>$diff->{'to_file'})},
5172                                       "history");
5173                         print "</td>\n";
5174
5175                 } # we should not encounter Unmerged (U) or Unknown (X) status
5176                 print "</tr>\n";
5177         }
5178         print "</tbody>" if $has_header;
5179         print "</table>\n";
5180 }
5181
5182 # Print context lines and then rem/add lines in a side-by-side manner.
5183 sub print_sidebyside_diff_lines {
5184         my ($ctx, $rem, $add) = @_;
5185
5186         # print context block before add/rem block
5187         if (@$ctx) {
5188                 print join '',
5189                         '<div class="chunk_block ctx">',
5190                                 '<div class="old">',
5191                                 @$ctx,
5192                                 '</div>',
5193                                 '<div class="new">',
5194                                 @$ctx,
5195                                 '</div>',
5196                         '</div>';
5197         }
5198
5199         if (!@$add) {
5200                 # pure removal
5201                 print join '',
5202                         '<div class="chunk_block rem">',
5203                                 '<div class="old">',
5204                                 @$rem,
5205                                 '</div>',
5206                         '</div>';
5207         } elsif (!@$rem) {
5208                 # pure addition
5209                 print join '',
5210                         '<div class="chunk_block add">',
5211                                 '<div class="new">',
5212                                 @$add,
5213                                 '</div>',
5214                         '</div>';
5215         } else {
5216                 print join '',
5217                         '<div class="chunk_block chg">',
5218                                 '<div class="old">',
5219                                 @$rem,
5220                                 '</div>',
5221                                 '<div class="new">',
5222                                 @$add,
5223                                 '</div>',
5224                         '</div>';
5225         }
5226 }
5227
5228 # Print context lines and then rem/add lines in inline manner.
5229 sub print_inline_diff_lines {
5230         my ($ctx, $rem, $add) = @_;
5231
5232         print @$ctx, @$rem, @$add;
5233 }
5234
5235 # Format removed and added line, mark changed part and HTML-format them.
5236 # Implementation is based on contrib/diff-highlight
5237 sub format_rem_add_lines_pair {
5238         my ($rem, $add, $num_parents) = @_;
5239
5240         # We need to untabify lines before split()'ing them;
5241         # otherwise offsets would be invalid.
5242         chomp $rem;
5243         chomp $add;
5244         $rem = untabify($rem);
5245         $add = untabify($add);
5246
5247         my @rem = split(//, $rem);
5248         my @add = split(//, $add);
5249         my ($esc_rem, $esc_add);
5250         # Ignore leading +/- characters for each parent.
5251         my ($prefix_len, $suffix_len) = ($num_parents, 0);
5252         my ($prefix_has_nonspace, $suffix_has_nonspace);
5253
5254         my $shorter = (@rem < @add) ? @rem : @add;
5255         while ($prefix_len < $shorter) {
5256                 last if ($rem[$prefix_len] ne $add[$prefix_len]);
5257
5258                 $prefix_has_nonspace = 1 if ($rem[$prefix_len] !~ /\s/);
5259                 $prefix_len++;
5260         }
5261
5262         while ($prefix_len + $suffix_len < $shorter) {
5263                 last if ($rem[-1 - $suffix_len] ne $add[-1 - $suffix_len]);
5264
5265                 $suffix_has_nonspace = 1 if ($rem[-1 - $suffix_len] !~ /\s/);
5266                 $suffix_len++;
5267         }
5268
5269         # Mark lines that are different from each other, but have some common
5270         # part that isn't whitespace.  If lines are completely different, don't
5271         # mark them because that would make output unreadable, especially if
5272         # diff consists of multiple lines.
5273         if ($prefix_has_nonspace || $suffix_has_nonspace) {
5274                 $esc_rem = esc_html_hl_regions($rem, 'marked',
5275                         [$prefix_len, @rem - $suffix_len], -nbsp=>1);
5276                 $esc_add = esc_html_hl_regions($add, 'marked',
5277                         [$prefix_len, @add - $suffix_len], -nbsp=>1);
5278         } else {
5279                 $esc_rem = esc_html($rem, -nbsp=>1);
5280                 $esc_add = esc_html($add, -nbsp=>1);
5281         }
5282
5283         return format_diff_line(\$esc_rem, 'rem'),
5284                format_diff_line(\$esc_add, 'add');
5285 }
5286
5287 # HTML-format diff context, removed and added lines.
5288 sub format_ctx_rem_add_lines {
5289         my ($ctx, $rem, $add, $num_parents) = @_;
5290         my (@new_ctx, @new_rem, @new_add);
5291         my $can_highlight = 0;
5292         my $is_combined = ($num_parents > 1);
5293
5294         # Highlight if every removed line has a corresponding added line.
5295         if (@$add > 0 && @$add == @$rem) {
5296                 $can_highlight = 1;
5297
5298                 # Highlight lines in combined diff only if the chunk contains
5299                 # diff between the same version, e.g.
5300                 #
5301                 #    - a
5302                 #   -  b
5303                 #    + c
5304                 #   +  d
5305                 #
5306                 # Otherwise the highlighting would be confusing.
5307                 if ($is_combined) {
5308                         for (my $i = 0; $i < @$add; $i++) {
5309                                 my $prefix_rem = substr($rem->[$i], 0, $num_parents);
5310                                 my $prefix_add = substr($add->[$i], 0, $num_parents);
5311
5312                                 $prefix_rem =~ s/-/+/g;
5313
5314                                 if ($prefix_rem ne $prefix_add) {
5315                                         $can_highlight = 0;
5316                                         last;
5317                                 }
5318                         }
5319                 }
5320         }
5321
5322         if ($can_highlight) {
5323                 for (my $i = 0; $i < @$add; $i++) {
5324                         my ($line_rem, $line_add) = format_rem_add_lines_pair(
5325                                 $rem->[$i], $add->[$i], $num_parents);
5326                         push @new_rem, $line_rem;
5327                         push @new_add, $line_add;
5328                 }
5329         } else {
5330                 @new_rem = map { format_diff_line($_, 'rem') } @$rem;
5331                 @new_add = map { format_diff_line($_, 'add') } @$add;
5332         }
5333
5334         @new_ctx = map { format_diff_line($_, 'ctx') } @$ctx;
5335
5336         return (\@new_ctx, \@new_rem, \@new_add);
5337 }
5338
5339 # Print context lines and then rem/add lines.
5340 sub print_diff_lines {
5341         my ($ctx, $rem, $add, $diff_style, $num_parents) = @_;
5342         my $is_combined = $num_parents > 1;
5343
5344         ($ctx, $rem, $add) = format_ctx_rem_add_lines($ctx, $rem, $add,
5345                 $num_parents);
5346
5347         if ($diff_style eq 'sidebyside' && !$is_combined) {
5348                 print_sidebyside_diff_lines($ctx, $rem, $add);
5349         } else {
5350                 # default 'inline' style and unknown styles
5351                 print_inline_diff_lines($ctx, $rem, $add);
5352         }
5353 }
5354
5355 sub print_diff_chunk {
5356         my ($diff_style, $num_parents, $from, $to, @chunk) = @_;
5357         my (@ctx, @rem, @add);
5358
5359         # The class of the previous line.
5360         my $prev_class = '';
5361
5362         return unless @chunk;
5363
5364         # incomplete last line might be among removed or added lines,
5365         # or both, or among context lines: find which
5366         for (my $i = 1; $i < @chunk; $i++) {
5367                 if ($chunk[$i][0] eq 'incomplete') {
5368                         $chunk[$i][0] = $chunk[$i-1][0];
5369                 }
5370         }
5371
5372         # guardian
5373         push @chunk, ["", ""];
5374
5375         foreach my $line_info (@chunk) {
5376                 my ($class, $line) = @$line_info;
5377
5378                 # print chunk headers
5379                 if ($class && $class eq 'chunk_header') {
5380                         print format_diff_line($line, $class, $from, $to);
5381                         next;
5382                 }
5383
5384                 ## print from accumulator when have some add/rem lines or end
5385                 # of chunk (flush context lines), or when have add and rem
5386                 # lines and new block is reached (otherwise add/rem lines could
5387                 # be reordered)
5388                 if (!$class || ((@rem || @add) && $class eq 'ctx') ||
5389                     (@rem && @add && $class ne $prev_class)) {
5390                         print_diff_lines(\@ctx, \@rem, \@add,
5391                                          $diff_style, $num_parents);
5392                         @ctx = @rem = @add = ();
5393                 }
5394
5395                 ## adding lines to accumulator
5396                 # guardian value
5397                 last unless $line;
5398                 # rem, add or change
5399                 if ($class eq 'rem') {
5400                         push @rem, $line;
5401                 } elsif ($class eq 'add') {
5402                         push @add, $line;
5403                 }
5404                 # context line
5405                 if ($class eq 'ctx') {
5406                         push @ctx, $line;
5407                 }
5408
5409                 $prev_class = $class;
5410         }
5411 }
5412
5413 sub git_patchset_body {
5414         my ($fd, $diff_style, $difftree, $hash, @hash_parents) = @_;
5415         my ($hash_parent) = $hash_parents[0];
5416
5417         my $is_combined = (@hash_parents > 1);
5418         my $patch_idx = 0;
5419         my $patch_number = 0;
5420         my $patch_line;
5421         my $diffinfo;
5422         my $to_name;
5423         my (%from, %to);
5424         my @chunk; # for side-by-side diff
5425
5426         print "<div class=\"patchset\">\n";
5427
5428         # skip to first patch
5429         while ($patch_line = <$fd>) {
5430                 chomp $patch_line;
5431
5432                 last if ($patch_line =~ m/^diff /);
5433         }
5434
5435  PATCH:
5436         while ($patch_line) {
5437
5438                 # parse "git diff" header line
5439                 if ($patch_line =~ m/^diff --git (\"(?:[^\\\"]*(?:\\.[^\\\"]*)*)\"|[^ "]*) (.*)$/) {
5440                         # $1 is from_name, which we do not use
5441                         $to_name = unquote($2);
5442                         $to_name =~ s!^b/!!;
5443                 } elsif ($patch_line =~ m/^diff --(cc|combined) ("?.*"?)$/) {
5444                         # $1 is 'cc' or 'combined', which we do not use
5445                         $to_name = unquote($2);
5446                 } else {
5447                         $to_name = undef;
5448                 }
5449
5450                 # check if current patch belong to current raw line
5451                 # and parse raw git-diff line if needed
5452                 if (is_patch_split($diffinfo, { 'to_file' => $to_name })) {
5453                         # this is continuation of a split patch
5454                         print "<div class=\"patch cont\">\n";
5455                 } else {
5456                         # advance raw git-diff output if needed
5457                         $patch_idx++ if defined $diffinfo;
5458
5459                         # read and prepare patch information
5460                         $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5461
5462                         # compact combined diff output can have some patches skipped
5463                         # find which patch (using pathname of result) we are at now;
5464                         if ($is_combined) {
5465                                 while ($to_name ne $diffinfo->{'to_file'}) {
5466                                         print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n" .
5467                                               format_diff_cc_simplified($diffinfo, @hash_parents) .
5468                                               "</div>\n";  # class="patch"
5469
5470                                         $patch_idx++;
5471                                         $patch_number++;
5472
5473                                         last if $patch_idx > $#$difftree;
5474                                         $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5475                                 }
5476                         }
5477
5478                         # modifies %from, %to hashes
5479                         parse_from_to_diffinfo($diffinfo, \%from, \%to, @hash_parents);
5480
5481                         # this is first patch for raw difftree line with $patch_idx index
5482                         # we index @$difftree array from 0, but number patches from 1
5483                         print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n";
5484                 }
5485
5486                 # git diff header
5487                 #assert($patch_line =~ m/^diff /) if DEBUG;
5488                 #assert($patch_line !~ m!$/$!) if DEBUG; # is chomp-ed
5489                 $patch_number++;
5490                 # print "git diff" header
5491                 print format_git_diff_header_line($patch_line, $diffinfo,
5492                                                   \%from, \%to);
5493
5494                 # print extended diff header
5495                 print "<div class=\"diff extended_header\">\n";
5496         EXTENDED_HEADER:
5497                 while ($patch_line = <$fd>) {
5498                         chomp $patch_line;
5499
5500                         last EXTENDED_HEADER if ($patch_line =~ m/^--- |^diff /);
5501
5502                         print format_extended_diff_header_line($patch_line, $diffinfo,
5503                                                                \%from, \%to);
5504                 }
5505                 print "</div>\n"; # class="diff extended_header"
5506
5507                 # from-file/to-file diff header
5508                 if (! $patch_line) {
5509                         print "</div>\n"; # class="patch"
5510                         last PATCH;
5511                 }
5512                 next PATCH if ($patch_line =~ m/^diff /);
5513                 #assert($patch_line =~ m/^---/) if DEBUG;
5514
5515                 my $last_patch_line = $patch_line;
5516                 $patch_line = <$fd>;
5517                 chomp $patch_line;
5518                 #assert($patch_line =~ m/^\+\+\+/) if DEBUG;
5519
5520                 print format_diff_from_to_header($last_patch_line, $patch_line,
5521                                                  $diffinfo, \%from, \%to,
5522                                                  @hash_parents);
5523
5524                 # the patch itself
5525         LINE:
5526                 while ($patch_line = <$fd>) {
5527                         chomp $patch_line;
5528
5529                         next PATCH if ($patch_line =~ m/^diff /);
5530
5531                         my $class = diff_line_class($patch_line, \%from, \%to);
5532
5533                         if ($class eq 'chunk_header') {
5534                                 print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
5535                                 @chunk = ();
5536                         }
5537
5538                         push @chunk, [ $class, $patch_line ];
5539                 }
5540
5541         } continue {
5542                 if (@chunk) {
5543                         print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
5544                         @chunk = ();
5545                 }
5546                 print "</div>\n"; # class="patch"
5547         }
5548
5549         # for compact combined (--cc) format, with chunk and patch simplification
5550         # the patchset might be empty, but there might be unprocessed raw lines
5551         for (++$patch_idx if $patch_number > 0;
5552              $patch_idx < @$difftree;
5553              ++$patch_idx) {
5554                 # read and prepare patch information
5555                 $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5556
5557                 # generate anchor for "patch" links in difftree / whatchanged part
5558                 print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n" .
5559                       format_diff_cc_simplified($diffinfo, @hash_parents) .
5560                       "</div>\n";  # class="patch"
5561
5562                 $patch_number++;
5563         }
5564
5565         if ($patch_number == 0) {
5566                 if (@hash_parents > 1) {
5567                         print "<div class=\"diff nodifferences\">Trivial merge</div>\n";
5568                 } else {
5569                         print "<div class=\"diff nodifferences\">No differences found</div>\n";
5570                 }
5571         }
5572
5573         print "</div>\n"; # class="patchset"
5574 }
5575
5576 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5577
5578 sub git_project_search_form {
5579         my ($searchtext, $search_use_regexp) = @_;
5580
5581         my $limit = '';
5582         if ($project_filter) {
5583                 $limit = " in '$project_filter/'";
5584         }
5585
5586         print "<div class=\"projsearch\">\n";
5587         print $cgi->start_form(-method => 'get', -action => $my_uri) .
5588               $cgi->hidden(-name => 'a', -value => 'project_list')  . "\n";
5589         print $cgi->hidden(-name => 'pf', -value => $project_filter). "\n"
5590                 if (defined $project_filter);
5591         print $cgi->textfield(-name => 's', -value => $searchtext,
5592                               -title => "Search project by name and description$limit",
5593                               -size => 60) . "\n" .
5594               "<span title=\"Extended regular expression\">" .
5595               $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
5596                              -checked => $search_use_regexp) .
5597               "</span>\n" .
5598               $cgi->submit(-name => 'btnS', -value => 'Search') .
5599               $cgi->end_form() . "\n" .
5600               $cgi->a({-href => href(project => undef, searchtext => undef,
5601                                      project_filter => $project_filter)},
5602                       esc_html("List all projects$limit")) . "<br />\n";
5603         print "</div>\n";
5604 }
5605
5606 # entry for given @keys needs filling if at least one of keys in list
5607 # is not present in %$project_info
5608 sub project_info_needs_filling {
5609         my ($project_info, @keys) = @_;
5610
5611         # return List::MoreUtils::any { !exists $project_info->{$_} } @keys;
5612         foreach my $key (@keys) {
5613                 if (!exists $project_info->{$key}) {
5614                         return 1;
5615                 }
5616         }
5617         return;
5618 }
5619
5620 # fills project list info (age, description, owner, category, forks, etc.)
5621 # for each project in the list, removing invalid projects from
5622 # returned list, or fill only specified info.
5623 #
5624 # Invalid projects are removed from the returned list if and only if you
5625 # ask 'age' or 'age_string' to be filled, because they are the only fields
5626 # that run unconditionally git command that requires repository, and
5627 # therefore do always check if project repository is invalid.
5628 #
5629 # USAGE:
5630 # * fill_project_list_info(\@project_list, 'descr_long', 'ctags')
5631 #   ensures that 'descr_long' and 'ctags' fields are filled
5632 # * @project_list = fill_project_list_info(\@project_list)
5633 #   ensures that all fields are filled (and invalid projects removed)
5634 #
5635 # NOTE: modifies $projlist, but does not remove entries from it
5636 sub fill_project_list_info {
5637         my ($projlist, @wanted_keys) = @_;
5638         my @projects;
5639         my $filter_set = sub { return @_; };
5640         if (@wanted_keys) {
5641                 my %wanted_keys = map { $_ => 1 } @wanted_keys;
5642                 $filter_set = sub { return grep { $wanted_keys{$_} } @_; };
5643         }
5644
5645         my $show_ctags = gitweb_check_feature('ctags');
5646  PROJECT:
5647         foreach my $pr (@$projlist) {
5648                 if (project_info_needs_filling($pr, $filter_set->('age', 'age_string'))) {
5649                         my (@activity) = git_get_last_activity($pr->{'path'});
5650                         unless (@activity) {
5651                                 next PROJECT;
5652                         }
5653                         ($pr->{'age'}, $pr->{'age_string'}) = @activity;
5654                 }
5655                 if (project_info_needs_filling($pr, $filter_set->('descr', 'descr_long'))) {
5656                         my $descr = git_get_project_description($pr->{'path'}) || "";
5657                         $descr = to_utf8($descr);
5658                         $pr->{'descr_long'} = $descr;
5659                         $pr->{'descr'} = chop_str($descr, $projects_list_description_width, 5);
5660                 }
5661                 if (project_info_needs_filling($pr, $filter_set->('owner'))) {
5662                         $pr->{'owner'} = git_get_project_owner("$pr->{'path'}") || "";
5663                 }
5664                 if ($show_ctags &&
5665                     project_info_needs_filling($pr, $filter_set->('ctags'))) {
5666                         $pr->{'ctags'} = git_get_project_ctags($pr->{'path'});
5667                 }
5668                 if ($projects_list_group_categories &&
5669                     project_info_needs_filling($pr, $filter_set->('category'))) {
5670                         my $cat = git_get_project_category($pr->{'path'}) ||
5671                                                            $project_list_default_category;
5672                         $pr->{'category'} = to_utf8($cat);
5673                 }
5674
5675                 push @projects, $pr;
5676         }
5677
5678         return @projects;
5679 }
5680
5681 sub sort_projects_list {
5682         my ($projlist, $order) = @_;
5683
5684         sub order_str {
5685                 my $key = shift;
5686                 return sub { $a->{$key} cmp $b->{$key} };
5687         }
5688
5689         sub order_num_then_undef {
5690                 my $key = shift;
5691                 return sub {
5692                         defined $a->{$key} ?
5693                                 (defined $b->{$key} ? $a->{$key} <=> $b->{$key} : -1) :
5694                                 (defined $b->{$key} ? 1 : 0)
5695                 };
5696         }
5697
5698         my %orderings = (
5699                 project => order_str('path'),
5700                 descr => order_str('descr_long'),
5701                 owner => order_str('owner'),
5702                 age => order_num_then_undef('age'),
5703         );
5704
5705         my $ordering = $orderings{$order};
5706         return defined $ordering ? sort $ordering @$projlist : @$projlist;
5707 }
5708
5709 # returns a hash of categories, containing the list of project
5710 # belonging to each category
5711 sub build_projlist_by_category {
5712         my ($projlist, $from, $to) = @_;
5713         my %categories;
5714
5715         $from = 0 unless defined $from;
5716         $to = $#$projlist if (!defined $to || $#$projlist < $to);
5717
5718         for (my $i = $from; $i <= $to; $i++) {
5719                 my $pr = $projlist->[$i];
5720                 push @{$categories{ $pr->{'category'} }}, $pr;
5721         }
5722
5723         return wantarray ? %categories : \%categories;
5724 }
5725
5726 # print 'sort by' <th> element, generating 'sort by $name' replay link
5727 # if that order is not selected
5728 sub print_sort_th {
5729         print format_sort_th(@_);
5730 }
5731
5732 sub format_sort_th {
5733         my ($name, $order, $header) = @_;
5734         my $sort_th = "";
5735         $header ||= ucfirst($name);
5736
5737         if ($order eq $name) {
5738                 $sort_th .= "<th>$header</th>\n";
5739         } else {
5740                 $sort_th .= "<th>" .
5741                             $cgi->a({-href => href(-replay=>1, order=>$name),
5742                                      -class => "header"}, $header) .
5743                             "</th>\n";
5744         }
5745
5746         return $sort_th;
5747 }
5748
5749 sub git_project_list_rows {
5750         my ($projlist, $from, $to, $check_forks) = @_;
5751
5752         $from = 0 unless defined $from;
5753         $to = $#$projlist if (!defined $to || $#$projlist < $to);
5754
5755         my $alternate = 1;
5756         for (my $i = $from; $i <= $to; $i++) {
5757                 my $pr = $projlist->[$i];
5758
5759                 if ($alternate) {
5760                         print "<tr class=\"dark\">\n";
5761                 } else {
5762                         print "<tr class=\"light\">\n";
5763                 }
5764                 $alternate ^= 1;
5765
5766                 if ($check_forks) {
5767                         print "<td>";
5768                         if ($pr->{'forks'}) {
5769                                 my $nforks = scalar @{$pr->{'forks'}};
5770                                 if ($nforks > 0) {
5771                                         print $cgi->a({-href => href(project=>$pr->{'path'}, action=>"forks"),
5772                                                        -title => "$nforks forks"}, "+");
5773                                 } else {
5774                                         print $cgi->span({-title => "$nforks forks"}, "+");
5775                                 }
5776                         }
5777                         print "</td>\n";
5778                 }
5779                 print "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
5780                                         -class => "list"},
5781                                        esc_html_match_hl($pr->{'path'}, $search_regexp)) .
5782                       "</td>\n" .
5783                       "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
5784                                         -class => "list",
5785                                         -title => $pr->{'descr_long'}},
5786                                         $search_regexp
5787                                         ? esc_html_match_hl_chopped($pr->{'descr_long'},
5788                                                                     $pr->{'descr'}, $search_regexp)
5789                                         : esc_html($pr->{'descr'})) .
5790                       "</td>\n";
5791                 unless ($omit_owner) {
5792                         print "<td><i>" . chop_and_escape_str($pr->{'owner'}, 15) . "</i></td>\n";
5793                 }
5794                 unless ($omit_age_column) {
5795                         print "<td class=\"". age_class($pr->{'age'}) . "\">" .
5796                             (defined $pr->{'age_string'} ? $pr->{'age_string'} : "No commits") . "</td>\n";
5797                 }
5798                 print"<td class=\"link\">" .
5799                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary")}, "summary")   . " | " .
5800                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"shortlog")}, "shortlog") . " | " .
5801                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"log")}, "log") . " | " .
5802                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"tree")}, "tree") .
5803                       ($pr->{'forks'} ? " | " . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"forks")}, "forks") : '') .
5804                       "</td>\n" .
5805                       "</tr>\n";
5806         }
5807 }
5808
5809 sub git_project_list_body {
5810         # actually uses global variable $project
5811         my ($projlist, $order, $from, $to, $extra, $no_header) = @_;
5812         my @projects = @$projlist;
5813
5814         my $check_forks = gitweb_check_feature('forks');
5815         my $show_ctags  = gitweb_check_feature('ctags');
5816         my $tagfilter = $show_ctags ? $input_params{'ctag'} : undef;
5817         $check_forks = undef
5818                 if ($tagfilter || $search_regexp);
5819
5820         # filtering out forks before filling info allows to do less work
5821         @projects = filter_forks_from_projects_list(\@projects)
5822                 if ($check_forks);
5823         # search_projects_list pre-fills required info
5824         @projects = search_projects_list(\@projects,
5825                                          'search_regexp' => $search_regexp,
5826                                          'tagfilter'  => $tagfilter)
5827                 if ($tagfilter || $search_regexp);
5828         # fill the rest
5829         my @all_fields = ('descr', 'descr_long', 'ctags', 'category');
5830         push @all_fields, ('age', 'age_string') unless($omit_age_column);
5831         push @all_fields, 'owner' unless($omit_owner);
5832         @projects = fill_project_list_info(\@projects, @all_fields);
5833
5834         $order ||= $default_projects_order;
5835         $from = 0 unless defined $from;
5836         $to = $#projects if (!defined $to || $#projects < $to);
5837
5838         # short circuit
5839         if ($from > $to) {
5840                 print "<center>\n".
5841                       "<b>No such projects found</b><br />\n".
5842                       "Click ".$cgi->a({-href=>href(project=>undef)},"here")." to view all projects<br />\n".
5843                       "</center>\n<br />\n";
5844                 return;
5845         }
5846
5847         @projects = sort_projects_list(\@projects, $order);
5848
5849         if ($show_ctags) {
5850                 my $ctags = git_gather_all_ctags(\@projects);
5851                 my $cloud = git_populate_project_tagcloud($ctags);
5852                 print git_show_project_tagcloud($cloud, 64);
5853         }
5854
5855         print "<table class=\"project_list\">\n";
5856         unless ($no_header) {
5857                 print "<tr>\n";
5858                 if ($check_forks) {
5859                         print "<th></th>\n";
5860                 }
5861                 print_sort_th('project', $order, 'Project');
5862                 print_sort_th('descr', $order, 'Description');
5863                 print_sort_th('owner', $order, 'Owner') unless $omit_owner;
5864                 print_sort_th('age', $order, 'Last Change') unless $omit_age_column;
5865                 print "<th></th>\n" . # for links
5866                       "</tr>\n";
5867         }
5868
5869         if ($projects_list_group_categories) {
5870                 # only display categories with projects in the $from-$to window
5871                 @projects = sort {$a->{'category'} cmp $b->{'category'}} @projects[$from..$to];
5872                 my %categories = build_projlist_by_category(\@projects, $from, $to);
5873                 foreach my $cat (sort keys %categories) {
5874                         unless ($cat eq "") {
5875                                 print "<tr>\n";
5876                                 if ($check_forks) {
5877                                         print "<td></td>\n";
5878                                 }
5879                                 print "<td class=\"category\" colspan=\"5\">".esc_html($cat)."</td>\n";
5880                                 print "</tr>\n";
5881                         }
5882
5883                         git_project_list_rows($categories{$cat}, undef, undef, $check_forks);
5884                 }
5885         } else {
5886                 git_project_list_rows(\@projects, $from, $to, $check_forks);
5887         }
5888
5889         if (defined $extra) {
5890                 print "<tr>\n";
5891                 if ($check_forks) {
5892                         print "<td></td>\n";
5893                 }
5894                 print "<td colspan=\"5\">$extra</td>\n" .
5895                       "</tr>\n";
5896         }
5897         print "</table>\n";
5898 }
5899
5900 sub git_log_body {
5901         # uses global variable $project
5902         my ($commitlist, $from, $to, $refs, $extra) = @_;
5903
5904         $from = 0 unless defined $from;
5905         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
5906
5907         for (my $i = 0; $i <= $to; $i++) {
5908                 my %co = %{$commitlist->[$i]};
5909                 next if !%co;
5910                 my $commit = $co{'id'};
5911                 my $ref = format_ref_marker($refs, $commit);
5912                 git_print_header_div('commit',
5913                                "<span class=\"age\">$co{'age_string'}</span>" .
5914                                esc_html($co{'title'}) . $ref,
5915                                $commit);
5916                 print "<div class=\"title_text\">\n" .
5917                       "<div class=\"log_link\">\n" .
5918                       $cgi->a({-href => href(action=>"commit", hash=>$commit)}, "commit") .
5919                       " | " .
5920                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff") .
5921                       " | " .
5922                       $cgi->a({-href => href(action=>"tree", hash=>$commit, hash_base=>$commit)}, "tree") .
5923                       "<br/>\n" .
5924                       "</div>\n";
5925                       git_print_authorship(\%co, -tag => 'span');
5926                       print "<br/>\n</div>\n";
5927
5928                 print "<div class=\"log_body\">\n";
5929                 git_print_log($co{'comment'}, -final_empty_line=> 1);
5930                 print "</div>\n";
5931         }
5932         if ($extra) {
5933                 print "<div class=\"page_nav\">\n";
5934                 print "$extra\n";
5935                 print "</div>\n";
5936         }
5937 }
5938
5939 sub git_shortlog_body {
5940         # uses global variable $project
5941         my ($commitlist, $from, $to, $refs, $extra) = @_;
5942
5943         $from = 0 unless defined $from;
5944         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
5945
5946         print "<table class=\"shortlog\">\n";
5947         my $alternate = 1;
5948         for (my $i = $from; $i <= $to; $i++) {
5949                 my %co = %{$commitlist->[$i]};
5950                 my $commit = $co{'id'};
5951                 my $ref = format_ref_marker($refs, $commit);
5952                 if ($alternate) {
5953                         print "<tr class=\"dark\">\n";
5954                 } else {
5955                         print "<tr class=\"light\">\n";
5956                 }
5957                 $alternate ^= 1;
5958                 # git_summary() used print "<td><i>$co{'age_string'}</i></td>\n" .
5959                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
5960                       format_author_html('td', \%co, 10) . "<td>";
5961                 print format_subject_html($co{'title'}, $co{'title_short'},
5962                                           href(action=>"commit", hash=>$commit), $ref);
5963                 print "</td>\n" .
5964                       "<td class=\"link\">" .
5965                       $cgi->a({-href => href(action=>"commit", hash=>$commit)}, "commit") . " | " .
5966                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff") . " | " .
5967                       $cgi->a({-href => href(action=>"tree", hash=>$commit, hash_base=>$commit)}, "tree");
5968                 my $snapshot_links = format_snapshot_links($commit);
5969                 if (defined $snapshot_links) {
5970                         print " | " . $snapshot_links;
5971                 }
5972                 print "</td>\n" .
5973                       "</tr>\n";
5974         }
5975         if (defined $extra) {
5976                 print "<tr>\n" .
5977                       "<td colspan=\"4\">$extra</td>\n" .
5978                       "</tr>\n";
5979         }
5980         print "</table>\n";
5981 }
5982
5983 sub git_history_body {
5984         # Warning: assumes constant type (blob or tree) during history
5985         my ($commitlist, $from, $to, $refs, $extra,
5986             $file_name, $file_hash, $ftype) = @_;
5987
5988         $from = 0 unless defined $from;
5989         $to = $#{$commitlist} unless (defined $to && $to <= $#{$commitlist});
5990
5991         print "<table class=\"history\">\n";
5992         my $alternate = 1;
5993         for (my $i = $from; $i <= $to; $i++) {
5994                 my %co = %{$commitlist->[$i]};
5995                 if (!%co) {
5996                         next;
5997                 }
5998                 my $commit = $co{'id'};
5999
6000                 my $ref = format_ref_marker($refs, $commit);
6001
6002                 if ($alternate) {
6003                         print "<tr class=\"dark\">\n";
6004                 } else {
6005                         print "<tr class=\"light\">\n";
6006                 }
6007                 $alternate ^= 1;
6008                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6009         # shortlog:   format_author_html('td', \%co, 10)
6010                       format_author_html('td', \%co, 15, 3) . "<td>";
6011                 # originally git_history used chop_str($co{'title'}, 50)
6012                 print format_subject_html($co{'title'}, $co{'title_short'},
6013                                           href(action=>"commit", hash=>$commit), $ref);
6014                 print "</td>\n" .
6015                       "<td class=\"link\">" .
6016                       $cgi->a({-href => href(action=>$ftype, hash_base=>$commit, file_name=>$file_name)}, $ftype) . " | " .
6017                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff");
6018
6019                 if ($ftype eq 'blob') {
6020                         print " | " .
6021                               $cgi->a({-href => href(action=>"blob_plain", hash_base=>$commit, file_name=>$file_name)}, "raw");
6022
6023                         my $blob_current = $file_hash;
6024                         my $blob_parent  = git_get_hash_by_path($commit, $file_name);
6025                         if (defined $blob_current && defined $blob_parent &&
6026                                         $blob_current ne $blob_parent) {
6027                                 print " | " .
6028                                         $cgi->a({-href => href(action=>"blobdiff",
6029                                                                hash=>$blob_current, hash_parent=>$blob_parent,
6030                                                                hash_base=>$hash_base, hash_parent_base=>$commit,
6031                                                                file_name=>$file_name)},
6032                                                 "diff to current");
6033                         }
6034                 }
6035                 print "</td>\n" .
6036                       "</tr>\n";
6037         }
6038         if (defined $extra) {
6039                 print "<tr>\n" .
6040                       "<td colspan=\"4\">$extra</td>\n" .
6041                       "</tr>\n";
6042         }
6043         print "</table>\n";
6044 }
6045
6046 sub git_tags_body {
6047         # uses global variable $project
6048         my ($taglist, $from, $to, $extra) = @_;
6049         $from = 0 unless defined $from;
6050         $to = $#{$taglist} if (!defined $to || $#{$taglist} < $to);
6051
6052         print "<table class=\"tags\">\n";
6053         my $alternate = 1;
6054         for (my $i = $from; $i <= $to; $i++) {
6055                 my $entry = $taglist->[$i];
6056                 my %tag = %$entry;
6057                 my $comment = $tag{'subject'};
6058                 my $comment_short;
6059                 if (defined $comment) {
6060                         $comment_short = chop_str($comment, 30, 5);
6061                 }
6062                 if ($alternate) {
6063                         print "<tr class=\"dark\">\n";
6064                 } else {
6065                         print "<tr class=\"light\">\n";
6066                 }
6067                 $alternate ^= 1;
6068                 if (defined $tag{'age'}) {
6069                         print "<td><i>$tag{'age'}</i></td>\n";
6070                 } else {
6071                         print "<td></td>\n";
6072                 }
6073                 print "<td>" .
6074                       $cgi->a({-href => href(action=>$tag{'reftype'}, hash=>$tag{'refid'}),
6075                                -class => "list name"}, esc_html($tag{'name'})) .
6076                       "</td>\n" .
6077                       "<td>";
6078                 if (defined $comment) {
6079                         print format_subject_html($comment, $comment_short,
6080                                                   href(action=>"tag", hash=>$tag{'id'}));
6081                 }
6082                 print "</td>\n" .
6083                       "<td class=\"selflink\">";
6084                 if ($tag{'type'} eq "tag") {
6085                         print $cgi->a({-href => href(action=>"tag", hash=>$tag{'id'})}, "tag");
6086                 } else {
6087                         print "&nbsp;";
6088                 }
6089                 print "</td>\n" .
6090                       "<td class=\"link\">" . " | " .
6091                       $cgi->a({-href => href(action=>$tag{'reftype'}, hash=>$tag{'refid'})}, $tag{'reftype'});
6092                 if ($tag{'reftype'} eq "commit") {
6093                         print " | " . $cgi->a({-href => href(action=>"shortlog", hash=>$tag{'fullname'})}, "shortlog") .
6094                               " | " . $cgi->a({-href => href(action=>"log", hash=>$tag{'fullname'})}, "log");
6095                 } elsif ($tag{'reftype'} eq "blob") {
6096                         print " | " . $cgi->a({-href => href(action=>"blob_plain", hash=>$tag{'refid'})}, "raw");
6097                 }
6098                 print "</td>\n" .
6099                       "</tr>";
6100         }
6101         if (defined $extra) {
6102                 print "<tr>\n" .
6103                       "<td colspan=\"5\">$extra</td>\n" .
6104                       "</tr>\n";
6105         }
6106         print "</table>\n";
6107 }
6108
6109 sub git_heads_body {
6110         # uses global variable $project
6111         my ($headlist, $head_at, $from, $to, $extra) = @_;
6112         $from = 0 unless defined $from;
6113         $to = $#{$headlist} if (!defined $to || $#{$headlist} < $to);
6114
6115         print "<table class=\"heads\">\n";
6116         my $alternate = 1;
6117         for (my $i = $from; $i <= $to; $i++) {
6118                 my $entry = $headlist->[$i];
6119                 my %ref = %$entry;
6120                 my $curr = defined $head_at && $ref{'id'} eq $head_at;
6121                 if ($alternate) {
6122                         print "<tr class=\"dark\">\n";
6123                 } else {
6124                         print "<tr class=\"light\">\n";
6125                 }
6126                 $alternate ^= 1;
6127                 print "<td><i>$ref{'age'}</i></td>\n" .
6128                       ($curr ? "<td class=\"current_head\">" : "<td>") .
6129                       $cgi->a({-href => href(action=>"shortlog", hash=>$ref{'fullname'}),
6130                                -class => "list name"},esc_html($ref{'name'})) .
6131                       "</td>\n" .
6132                       "<td class=\"link\">" .
6133                       $cgi->a({-href => href(action=>"shortlog", hash=>$ref{'fullname'})}, "shortlog") . " | " .
6134                       $cgi->a({-href => href(action=>"log", hash=>$ref{'fullname'})}, "log") . " | " .
6135                       $cgi->a({-href => href(action=>"tree", hash=>$ref{'fullname'}, hash_base=>$ref{'fullname'})}, "tree") .
6136                       "</td>\n" .
6137                       "</tr>";
6138         }
6139         if (defined $extra) {
6140                 print "<tr>\n" .
6141                       "<td colspan=\"3\">$extra</td>\n" .
6142                       "</tr>\n";
6143         }
6144         print "</table>\n";
6145 }
6146
6147 # Display a single remote block
6148 sub git_remote_block {
6149         my ($remote, $rdata, $limit, $head) = @_;
6150
6151         my $heads = $rdata->{'heads'};
6152         my $fetch = $rdata->{'fetch'};
6153         my $push = $rdata->{'push'};
6154
6155         my $urls_table = "<table class=\"projects_list\">\n" ;
6156
6157         if (defined $fetch) {
6158                 if ($fetch eq $push) {
6159                         $urls_table .= format_repo_url("URL", $fetch);
6160                 } else {
6161                         $urls_table .= format_repo_url("Fetch URL", $fetch);
6162                         $urls_table .= format_repo_url("Push URL", $push) if defined $push;
6163                 }
6164         } elsif (defined $push) {
6165                 $urls_table .= format_repo_url("Push URL", $push);
6166         } else {
6167                 $urls_table .= format_repo_url("", "No remote URL");
6168         }
6169
6170         $urls_table .= "</table>\n";
6171
6172         my $dots;
6173         if (defined $limit && $limit < @$heads) {
6174                 $dots = $cgi->a({-href => href(action=>"remotes", hash=>$remote)}, "...");
6175         }
6176
6177         print $urls_table;
6178         git_heads_body($heads, $head, 0, $limit, $dots);
6179 }
6180
6181 # Display a list of remote names with the respective fetch and push URLs
6182 sub git_remotes_list {
6183         my ($remotedata, $limit) = @_;
6184         print "<table class=\"heads\">\n";
6185         my $alternate = 1;
6186         my @remotes = sort keys %$remotedata;
6187
6188         my $limited = $limit && $limit < @remotes;
6189
6190         $#remotes = $limit - 1 if $limited;
6191
6192         while (my $remote = shift @remotes) {
6193                 my $rdata = $remotedata->{$remote};
6194                 my $fetch = $rdata->{'fetch'};
6195                 my $push = $rdata->{'push'};
6196                 if ($alternate) {
6197                         print "<tr class=\"dark\">\n";
6198                 } else {
6199                         print "<tr class=\"light\">\n";
6200                 }
6201                 $alternate ^= 1;
6202                 print "<td>" .
6203                       $cgi->a({-href=> href(action=>'remotes', hash=>$remote),
6204                                -class=> "list name"},esc_html($remote)) .
6205                       "</td>";
6206                 print "<td class=\"link\">" .
6207                       (defined $fetch ? $cgi->a({-href=> $fetch}, "fetch") : "fetch") .
6208                       " | " .
6209                       (defined $push ? $cgi->a({-href=> $push}, "push") : "push") .
6210                       "</td>";
6211
6212                 print "</tr>\n";
6213         }
6214
6215         if ($limited) {
6216                 print "<tr>\n" .
6217                       "<td colspan=\"3\">" .
6218                       $cgi->a({-href => href(action=>"remotes")}, "...") .
6219                       "</td>\n" . "</tr>\n";
6220         }
6221
6222         print "</table>";
6223 }
6224
6225 # Display remote heads grouped by remote, unless there are too many
6226 # remotes, in which case we only display the remote names
6227 sub git_remotes_body {
6228         my ($remotedata, $limit, $head) = @_;
6229         if ($limit and $limit < keys %$remotedata) {
6230                 git_remotes_list($remotedata, $limit);
6231         } else {
6232                 fill_remote_heads($remotedata);
6233                 while (my ($remote, $rdata) = each %$remotedata) {
6234                         git_print_section({-class=>"remote", -id=>$remote},
6235                                 ["remotes", $remote, $remote], sub {
6236                                         git_remote_block($remote, $rdata, $limit, $head);
6237                                 });
6238                 }
6239         }
6240 }
6241
6242 sub git_search_message {
6243         my %co = @_;
6244
6245         my $greptype;
6246         if ($searchtype eq 'commit') {
6247                 $greptype = "--grep=";
6248         } elsif ($searchtype eq 'author') {
6249                 $greptype = "--author=";
6250         } elsif ($searchtype eq 'committer') {
6251                 $greptype = "--committer=";
6252         }
6253         $greptype .= $searchtext;
6254         my @commitlist = parse_commits($hash, 101, (100 * $page), undef,
6255                                        $greptype, '--regexp-ignore-case',
6256                                        $search_use_regexp ? '--extended-regexp' : '--fixed-strings');
6257
6258         my $paging_nav = '';
6259         if ($page > 0) {
6260                 $paging_nav .=
6261                         $cgi->a({-href => href(-replay=>1, page=>undef)},
6262                                 "first") .
6263                         " &sdot; " .
6264                         $cgi->a({-href => href(-replay=>1, page=>$page-1),
6265                                  -accesskey => "p", -title => "Alt-p"}, "prev");
6266         } else {
6267                 $paging_nav .= "first &sdot; prev";
6268         }
6269         my $next_link = '';
6270         if ($#commitlist >= 100) {
6271                 $next_link =
6272                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
6273                                  -accesskey => "n", -title => "Alt-n"}, "next");
6274                 $paging_nav .= " &sdot; $next_link";
6275         } else {
6276                 $paging_nav .= " &sdot; next";
6277         }
6278
6279         git_header_html();
6280
6281         git_print_page_nav('','', $hash,$co{'tree'},$hash, $paging_nav);
6282         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6283         if ($page == 0 && !@commitlist) {
6284                 print "<p>No match.</p>\n";
6285         } else {
6286                 git_search_grep_body(\@commitlist, 0, 99, $next_link);
6287         }
6288
6289         git_footer_html();
6290 }
6291
6292 sub git_search_changes {
6293         my %co = @_;
6294
6295         local $/ = "\n";
6296         open my $fd, '-|', git_cmd(), '--no-pager', 'log', @diff_opts,
6297                 '--pretty=format:%H', '--no-abbrev', '--raw', "-S$searchtext",
6298                 ($search_use_regexp ? '--pickaxe-regex' : ())
6299                         or die_error(500, "Open git-log failed");
6300
6301         git_header_html();
6302
6303         git_print_page_nav('','', $hash,$co{'tree'},$hash);
6304         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6305
6306         print "<table class=\"pickaxe search\">\n";
6307         my $alternate = 1;
6308         undef %co;
6309         my @files;
6310         while (my $line = <$fd>) {
6311                 chomp $line;
6312                 next unless $line;
6313
6314                 my %set = parse_difftree_raw_line($line);
6315                 if (defined $set{'commit'}) {
6316                         # finish previous commit
6317                         if (%co) {
6318                                 print "</td>\n" .
6319                                       "<td class=\"link\">" .
6320                                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})},
6321                                               "commit") .
6322                                       " | " .
6323                                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'},
6324                                                              hash_base=>$co{'id'})},
6325                                               "tree") .
6326                                       "</td>\n" .
6327                                       "</tr>\n";
6328                         }
6329
6330                         if ($alternate) {
6331                                 print "<tr class=\"dark\">\n";
6332                         } else {
6333                                 print "<tr class=\"light\">\n";
6334                         }
6335                         $alternate ^= 1;
6336                         %co = parse_commit($set{'commit'});
6337                         my $author = chop_and_escape_str($co{'author_name'}, 15, 5);
6338                         print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6339                               "<td><i>$author</i></td>\n" .
6340                               "<td>" .
6341                               $cgi->a({-href => href(action=>"commit", hash=>$co{'id'}),
6342                                       -class => "list subject"},
6343                                       chop_and_escape_str($co{'title'}, 50) . "<br/>");
6344                 } elsif (defined $set{'to_id'}) {
6345                         next if is_deleted(\%set);
6346
6347                         print $cgi->a({-href => href(action=>"blob", hash_base=>$co{'id'},
6348                                                      hash=>$set{'to_id'}, file_name=>$set{'to_file'}),
6349                                       -class => "list"},
6350                                       "<span class=\"match\">" . esc_path($set{'file'}) . "</span>") .
6351                               "<br/>\n";
6352                 }
6353         }
6354         close $fd;
6355
6356         # finish last commit (warning: repetition!)
6357         if (%co) {
6358                 print "</td>\n" .
6359                       "<td class=\"link\">" .
6360                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})},
6361                               "commit") .
6362                       " | " .
6363                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'},
6364                                              hash_base=>$co{'id'})},
6365                               "tree") .
6366                       "</td>\n" .
6367                       "</tr>\n";
6368         }
6369
6370         print "</table>\n";
6371
6372         git_footer_html();
6373 }
6374
6375 sub git_search_files {
6376         my %co = @_;
6377
6378         local $/ = "\n";
6379         open my $fd, "-|", git_cmd(), 'grep', '-n', '-z',
6380                 $search_use_regexp ? ('-E', '-i') : '-F',
6381                 $searchtext, $co{'tree'}
6382                         or die_error(500, "Open git-grep failed");
6383
6384         git_header_html();
6385
6386         git_print_page_nav('','', $hash,$co{'tree'},$hash);
6387         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6388
6389         print "<table class=\"grep_search\">\n";
6390         my $alternate = 1;
6391         my $matches = 0;
6392         my $lastfile = '';
6393         my $file_href;
6394         while (my $line = <$fd>) {
6395                 chomp $line;
6396                 my ($file, $lno, $ltext, $binary);
6397                 last if ($matches++ > 1000);
6398                 if ($line =~ /^Binary file (.+) matches$/) {
6399                         $file = $1;
6400                         $binary = 1;
6401                 } else {
6402                         ($file, $lno, $ltext) = split(/\0/, $line, 3);
6403                         $file =~ s/^$co{'tree'}://;
6404                 }
6405                 if ($file ne $lastfile) {
6406                         $lastfile and print "</td></tr>\n";
6407                         if ($alternate++) {
6408                                 print "<tr class=\"dark\">\n";
6409                         } else {
6410                                 print "<tr class=\"light\">\n";
6411                         }
6412                         $file_href = href(action=>"blob", hash_base=>$co{'id'},
6413                                           file_name=>$file);
6414                         print "<td class=\"list\">".
6415                                 $cgi->a({-href => $file_href, -class => "list"}, esc_path($file));
6416                         print "</td><td>\n";
6417                         $lastfile = $file;
6418                 }
6419                 if ($binary) {
6420                         print "<div class=\"binary\">Binary file</div>\n";
6421                 } else {
6422                         $ltext = untabify($ltext);
6423                         if ($ltext =~ m/^(.*)($search_regexp)(.*)$/i) {
6424                                 $ltext = esc_html($1, -nbsp=>1);
6425                                 $ltext .= '<span class="match">';
6426                                 $ltext .= esc_html($2, -nbsp=>1);
6427                                 $ltext .= '</span>';
6428                                 $ltext .= esc_html($3, -nbsp=>1);
6429                         } else {
6430                                 $ltext = esc_html($ltext, -nbsp=>1);
6431                         }
6432                         print "<div class=\"pre\">" .
6433                                 $cgi->a({-href => $file_href.'#l'.$lno,
6434                                         -class => "linenr"}, sprintf('%4i', $lno)) .
6435                                 ' ' .  $ltext . "</div>\n";
6436                 }
6437         }
6438         if ($lastfile) {
6439                 print "</td></tr>\n";
6440                 if ($matches > 1000) {
6441                         print "<div class=\"diff nodifferences\">Too many matches, listing trimmed</div>\n";
6442                 }
6443         } else {
6444                 print "<div class=\"diff nodifferences\">No matches found</div>\n";
6445         }
6446         close $fd;
6447
6448         print "</table>\n";
6449
6450         git_footer_html();
6451 }
6452
6453 sub git_search_grep_body {
6454         my ($commitlist, $from, $to, $extra) = @_;
6455         $from = 0 unless defined $from;
6456         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
6457
6458         print "<table class=\"commit_search\">\n";
6459         my $alternate = 1;
6460         for (my $i = $from; $i <= $to; $i++) {
6461                 my %co = %{$commitlist->[$i]};
6462                 if (!%co) {
6463                         next;
6464                 }
6465                 my $commit = $co{'id'};
6466                 if ($alternate) {
6467                         print "<tr class=\"dark\">\n";
6468                 } else {
6469                         print "<tr class=\"light\">\n";
6470                 }
6471                 $alternate ^= 1;
6472                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6473                       format_author_html('td', \%co, 15, 5) .
6474                       "<td>" .
6475                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'}),
6476                                -class => "list subject"},
6477                               chop_and_escape_str($co{'title'}, 50) . "<br/>");
6478                 my $comment = $co{'comment'};
6479                 foreach my $line (@$comment) {
6480                         if ($line =~ m/^(.*?)($search_regexp)(.*)$/i) {
6481                                 my ($lead, $match, $trail) = ($1, $2, $3);
6482                                 $match = chop_str($match, 70, 5, 'center');
6483                                 my $contextlen = int((80 - length($match))/2);
6484                                 $contextlen = 30 if ($contextlen > 30);
6485                                 $lead  = chop_str($lead,  $contextlen, 10, 'left');
6486                                 $trail = chop_str($trail, $contextlen, 10, 'right');
6487
6488                                 $lead  = esc_html($lead);
6489                                 $match = esc_html($match);
6490                                 $trail = esc_html($trail);
6491
6492                                 print "$lead<span class=\"match\">$match</span>$trail<br />";
6493                         }
6494                 }
6495                 print "</td>\n" .
6496                       "<td class=\"link\">" .
6497                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})}, "commit") .
6498                       " | " .
6499                       $cgi->a({-href => href(action=>"commitdiff", hash=>$co{'id'})}, "commitdiff") .
6500                       " | " .
6501                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$co{'id'})}, "tree");
6502                 print "</td>\n" .
6503                       "</tr>\n";
6504         }
6505         if (defined $extra) {
6506                 print "<tr>\n" .
6507                       "<td colspan=\"3\">$extra</td>\n" .
6508                       "</tr>\n";
6509         }
6510         print "</table>\n";
6511 }
6512
6513 ## ======================================================================
6514 ## ======================================================================
6515 ## actions
6516
6517 sub git_project_list {
6518         my $order = $input_params{'order'};
6519         if (defined $order && $order !~ m/none|project|descr|owner|age/) {
6520                 die_error(400, "Unknown order parameter");
6521         }
6522
6523         my @list = git_get_projects_list($project_filter, $strict_export);
6524         if (!@list) {
6525                 die_error(404, "No projects found");
6526         }
6527
6528         git_header_html();
6529         if (defined $home_text && -f $home_text) {
6530                 print "<div class=\"index_include\">\n";
6531                 insert_file($home_text);
6532                 print "</div>\n";
6533         }
6534
6535         git_project_search_form($searchtext, $search_use_regexp);
6536         git_project_list_body(\@list, $order);
6537         git_footer_html();
6538 }
6539
6540 sub git_forks {
6541         my $order = $input_params{'order'};
6542         if (defined $order && $order !~ m/none|project|descr|owner|age/) {
6543                 die_error(400, "Unknown order parameter");
6544         }
6545
6546         my $filter = $project;
6547         $filter =~ s/\.git$//;
6548         my @list = git_get_projects_list($filter);
6549         if (!@list) {
6550                 die_error(404, "No forks found");
6551         }
6552
6553         git_header_html();
6554         git_print_page_nav('','');
6555         git_print_header_div('summary', "$project forks");
6556         git_project_list_body(\@list, $order);
6557         git_footer_html();
6558 }
6559
6560 sub git_project_index {
6561         my @projects = git_get_projects_list($project_filter, $strict_export);
6562         if (!@projects) {
6563                 die_error(404, "No projects found");
6564         }
6565
6566         print $cgi->header(
6567                 -type => 'text/plain',
6568                 -charset => 'utf-8',
6569                 -content_disposition => 'inline; filename="index.aux"');
6570
6571         foreach my $pr (@projects) {
6572                 if (!exists $pr->{'owner'}) {
6573                         $pr->{'owner'} = git_get_project_owner("$pr->{'path'}");
6574                 }
6575
6576                 my ($path, $owner) = ($pr->{'path'}, $pr->{'owner'});
6577                 # quote as in CGI::Util::encode, but keep the slash, and use '+' for ' '
6578                 $path  =~ s/([^a-zA-Z0-9_.\-\/ ])/sprintf("%%%02X", ord($1))/eg;
6579                 $owner =~ s/([^a-zA-Z0-9_.\-\/ ])/sprintf("%%%02X", ord($1))/eg;
6580                 $path  =~ s/ /\+/g;
6581                 $owner =~ s/ /\+/g;
6582
6583                 print "$path $owner\n";
6584         }
6585 }
6586
6587 sub git_summary {
6588         my $descr = git_get_project_description($project) || "none";
6589         my %co = parse_commit("HEAD");
6590         my %cd = %co ? parse_date($co{'committer_epoch'}, $co{'committer_tz'}) : ();
6591         my $head = $co{'id'};
6592         my $remote_heads = gitweb_check_feature('remote_heads');
6593
6594         my $owner = git_get_project_owner($project);
6595
6596         my $refs = git_get_references();
6597         # These get_*_list functions return one more to allow us to see if
6598         # there are more ...
6599         my @taglist  = git_get_tags_list(16);
6600         my @headlist = git_get_heads_list(16);
6601         my %remotedata = $remote_heads ? git_get_remotes_list() : ();
6602         my @forklist;
6603         my $check_forks = gitweb_check_feature('forks');
6604
6605         if ($check_forks) {
6606                 # find forks of a project
6607                 my $filter = $project;
6608                 $filter =~ s/\.git$//;
6609                 @forklist = git_get_projects_list($filter);
6610                 # filter out forks of forks
6611                 @forklist = filter_forks_from_projects_list(\@forklist)
6612                         if (@forklist);
6613         }
6614
6615         git_header_html();
6616         git_print_page_nav('summary','', $head);
6617
6618         print "<div class=\"title\">&nbsp;</div>\n";
6619         print "<table class=\"projects_list\">\n" .
6620               "<tr id=\"metadata_desc\"><td>description</td><td>" . esc_html($descr) . "</td></tr>\n";
6621         if ($owner and not $omit_owner) {
6622                 print  "<tr id=\"metadata_owner\"><td>owner</td><td>" . esc_html($owner) . "</td></tr>\n";
6623         }
6624         if (defined $cd{'rfc2822'}) {
6625                 print "<tr id=\"metadata_lchange\"><td>last change</td>" .
6626                       "<td>".format_timestamp_html(\%cd)."</td></tr>\n";
6627         }
6628
6629         # use per project git URL list in $projectroot/$project/cloneurl
6630         # or make project git URL from git base URL and project name
6631         my $url_tag = "URL";
6632         my @url_list = git_get_project_url_list($project);
6633         @url_list = map { "$_/$project" } @git_base_url_list unless @url_list;
6634         foreach my $git_url (@url_list) {
6635                 next unless $git_url;
6636                 print format_repo_url($url_tag, $git_url);
6637                 $url_tag = "";
6638         }
6639
6640         # Tag cloud
6641         my $show_ctags = gitweb_check_feature('ctags');
6642         if ($show_ctags) {
6643                 my $ctags = git_get_project_ctags($project);
6644                 if (%$ctags) {
6645                         # without ability to add tags, don't show if there are none
6646                         my $cloud = git_populate_project_tagcloud($ctags);
6647                         print "<tr id=\"metadata_ctags\">" .
6648                               "<td>content tags</td>" .
6649                               "<td>".git_show_project_tagcloud($cloud, 48)."</td>" .
6650                               "</tr>\n";
6651                 }
6652         }
6653
6654         print "</table>\n";
6655
6656         # If XSS prevention is on, we don't include README.html.
6657         # TODO: Allow a readme in some safe format.
6658         if (!$prevent_xss && -s "$projectroot/$project/README.html") {
6659                 print "<div class=\"title\">readme</div>\n" .
6660                       "<div class=\"readme\">\n";
6661                 insert_file("$projectroot/$project/README.html");
6662                 print "\n</div>\n"; # class="readme"
6663         }
6664
6665         # we need to request one more than 16 (0..15) to check if
6666         # those 16 are all
6667         my @commitlist = $head ? parse_commits($head, 17) : ();
6668         if (@commitlist) {
6669                 git_print_header_div('shortlog');
6670                 git_shortlog_body(\@commitlist, 0, 15, $refs,
6671                                   $#commitlist <=  15 ? undef :
6672                                   $cgi->a({-href => href(action=>"shortlog")}, "..."));
6673         }
6674
6675         if (@taglist) {
6676                 git_print_header_div('tags');
6677                 git_tags_body(\@taglist, 0, 15,
6678                               $#taglist <=  15 ? undef :
6679                               $cgi->a({-href => href(action=>"tags")}, "..."));
6680         }
6681
6682         if (@headlist) {
6683                 git_print_header_div('heads');
6684                 git_heads_body(\@headlist, $head, 0, 15,
6685                                $#headlist <= 15 ? undef :
6686                                $cgi->a({-href => href(action=>"heads")}, "..."));
6687         }
6688
6689         if (%remotedata) {
6690                 git_print_header_div('remotes');
6691                 git_remotes_body(\%remotedata, 15, $head);
6692         }
6693
6694         if (@forklist) {
6695                 git_print_header_div('forks');
6696                 git_project_list_body(\@forklist, 'age', 0, 15,
6697                                       $#forklist <= 15 ? undef :
6698                                       $cgi->a({-href => href(action=>"forks")}, "..."),
6699                                       'no_header');
6700         }
6701
6702         git_footer_html();
6703 }
6704
6705 sub git_tag {
6706         my %tag = parse_tag($hash);
6707
6708         if (! %tag) {
6709                 die_error(404, "Unknown tag object");
6710         }
6711
6712         my $head = git_get_head_hash($project);
6713         git_header_html();
6714         git_print_page_nav('','', $head,undef,$head);
6715         git_print_header_div('commit', esc_html($tag{'name'}), $hash);
6716         print "<div class=\"title_text\">\n" .
6717               "<table class=\"object_header\">\n" .
6718               "<tr>\n" .
6719               "<td>object</td>\n" .
6720               "<td>" . $cgi->a({-class => "list", -href => href(action=>$tag{'type'}, hash=>$tag{'object'})},
6721                                $tag{'object'}) . "</td>\n" .
6722               "<td class=\"link\">" . $cgi->a({-href => href(action=>$tag{'type'}, hash=>$tag{'object'})},
6723                                               $tag{'type'}) . "</td>\n" .
6724               "</tr>\n";
6725         if (defined($tag{'author'})) {
6726                 git_print_authorship_rows(\%tag, 'author');
6727         }
6728         print "</table>\n\n" .
6729               "</div>\n";
6730         print "<div class=\"page_body\">";
6731         my $comment = $tag{'comment'};
6732         foreach my $line (@$comment) {
6733                 chomp $line;
6734                 print esc_html($line, -nbsp=>1) . "<br/>\n";
6735         }
6736         print "</div>\n";
6737         git_footer_html();
6738 }
6739
6740 sub git_blame_common {
6741         my $format = shift || 'porcelain';
6742         if ($format eq 'porcelain' && $input_params{'javascript'}) {
6743                 $format = 'incremental';
6744                 $action = 'blame_incremental'; # for page title etc
6745         }
6746
6747         # permissions
6748         gitweb_check_feature('blame')
6749                 or die_error(403, "Blame view not allowed");
6750
6751         # error checking
6752         die_error(400, "No file name given") unless $file_name;
6753         $hash_base ||= git_get_head_hash($project);
6754         die_error(404, "Couldn't find base commit") unless $hash_base;
6755         my %co = parse_commit($hash_base)
6756                 or die_error(404, "Commit not found");
6757         my $ftype = "blob";
6758         if (!defined $hash) {
6759                 $hash = git_get_hash_by_path($hash_base, $file_name, "blob")
6760                         or die_error(404, "Error looking up file");
6761         } else {
6762                 $ftype = git_get_type($hash);
6763                 if ($ftype !~ "blob") {
6764                         die_error(400, "Object is not a blob");
6765                 }
6766         }
6767
6768         my $fd;
6769         if ($format eq 'incremental') {
6770                 # get file contents (as base)
6771                 open $fd, "-|", git_cmd(), 'cat-file', 'blob', $hash
6772                         or die_error(500, "Open git-cat-file failed");
6773         } elsif ($format eq 'data') {
6774                 # run git-blame --incremental
6775                 open $fd, "-|", git_cmd(), "blame", "--incremental",
6776                         $hash_base, "--", $file_name
6777                         or die_error(500, "Open git-blame --incremental failed");
6778         } else {
6779                 # run git-blame --porcelain
6780                 open $fd, "-|", git_cmd(), "blame", '-p',
6781                         $hash_base, '--', $file_name
6782                         or die_error(500, "Open git-blame --porcelain failed");
6783         }
6784         binmode $fd, ':utf8';
6785
6786         # incremental blame data returns early
6787         if ($format eq 'data') {
6788                 print $cgi->header(
6789                         -type=>"text/plain", -charset => "utf-8",
6790                         -status=> "200 OK");
6791                 local $| = 1; # output autoflush
6792                 while (my $line = <$fd>) {
6793                         print to_utf8($line);
6794                 }
6795                 close $fd
6796                         or print "ERROR $!\n";
6797
6798                 print 'END';
6799                 if (defined $t0 && gitweb_check_feature('timed')) {
6800                         print ' '.
6801                               tv_interval($t0, [ gettimeofday() ]).
6802                               ' '.$number_of_git_cmds;
6803                 }
6804                 print "\n";
6805
6806                 return;
6807         }
6808
6809         # page header
6810         git_header_html();
6811         my $formats_nav =
6812                 $cgi->a({-href => href(action=>"blob", -replay=>1)},
6813                         "blob") .
6814                 " | ";
6815         if ($format eq 'incremental') {
6816                 $formats_nav .=
6817                         $cgi->a({-href => href(action=>"blame", javascript=>0, -replay=>1)},
6818                                 "blame") . " (non-incremental)";
6819         } else {
6820                 $formats_nav .=
6821                         $cgi->a({-href => href(action=>"blame_incremental", -replay=>1)},
6822                                 "blame") . " (incremental)";
6823         }
6824         $formats_nav .=
6825                 " | " .
6826                 $cgi->a({-href => href(action=>"history", -replay=>1)},
6827                         "history") .
6828                 " | " .
6829                 $cgi->a({-href => href(action=>$action, file_name=>$file_name)},
6830                         "HEAD");
6831         git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
6832         git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
6833         git_print_page_path($file_name, $ftype, $hash_base);
6834
6835         # page body
6836         if ($format eq 'incremental') {
6837                 print "<noscript>\n<div class=\"error\"><center><b>\n".
6838                       "This page requires JavaScript to run.\n Use ".
6839                       $cgi->a({-href => href(action=>'blame',javascript=>0,-replay=>1)},
6840                               'this page').
6841                       " instead.\n".
6842                       "</b></center></div>\n</noscript>\n";
6843
6844                 print qq!<div id="progress_bar" style="width: 100%; background-color: yellow"></div>\n!;
6845         }
6846
6847         print qq!<div class="page_body">\n!;
6848         print qq!<div id="progress_info">... / ...</div>\n!
6849                 if ($format eq 'incremental');
6850         print qq!<table id="blame_table" class="blame" width="100%">\n!.
6851               #qq!<col width="5.5em" /><col width="2.5em" /><col width="*" />\n!.
6852               qq!<thead>\n!.
6853               qq!<tr><th>Commit</th><th>Line</th><th>Data</th></tr>\n!.
6854               qq!</thead>\n!.
6855               qq!<tbody>\n!;
6856
6857         my @rev_color = qw(light dark);
6858         my $num_colors = scalar(@rev_color);
6859         my $current_color = 0;
6860
6861         if ($format eq 'incremental') {
6862                 my $color_class = $rev_color[$current_color];
6863
6864                 #contents of a file
6865                 my $linenr = 0;
6866         LINE:
6867                 while (my $line = <$fd>) {
6868                         chomp $line;
6869                         $linenr++;
6870
6871                         print qq!<tr id="l$linenr" class="$color_class">!.
6872                               qq!<td class="sha1"><a href=""> </a></td>!.
6873                               qq!<td class="linenr">!.
6874                               qq!<a class="linenr" href="">$linenr</a></td>!;
6875                         print qq!<td class="pre">! . esc_html($line) . "</td>\n";
6876                         print qq!</tr>\n!;
6877                 }
6878
6879         } else { # porcelain, i.e. ordinary blame
6880                 my %metainfo = (); # saves information about commits
6881
6882                 # blame data
6883         LINE:
6884                 while (my $line = <$fd>) {
6885                         chomp $line;
6886                         # the header: <SHA-1> <src lineno> <dst lineno> [<lines in group>]
6887                         # no <lines in group> for subsequent lines in group of lines
6888                         my ($full_rev, $orig_lineno, $lineno, $group_size) =
6889                            ($line =~ /^($oid_regex) (\d+) (\d+)(?: (\d+))?$/);
6890                         if (!exists $metainfo{$full_rev}) {
6891                                 $metainfo{$full_rev} = { 'nprevious' => 0 };
6892                         }
6893                         my $meta = $metainfo{$full_rev};
6894                         my $data;
6895                         while ($data = <$fd>) {
6896                                 chomp $data;
6897                                 last if ($data =~ s/^\t//); # contents of line
6898                                 if ($data =~ /^(\S+)(?: (.*))?$/) {
6899                                         $meta->{$1} = $2 unless exists $meta->{$1};
6900                                 }
6901                                 if ($data =~ /^previous /) {
6902                                         $meta->{'nprevious'}++;
6903                                 }
6904                         }
6905                         my $short_rev = substr($full_rev, 0, 8);
6906                         my $author = $meta->{'author'};
6907                         my %date =
6908                                 parse_date($meta->{'author-time'}, $meta->{'author-tz'});
6909                         my $date = $date{'iso-tz'};
6910                         if ($group_size) {
6911                                 $current_color = ($current_color + 1) % $num_colors;
6912                         }
6913                         my $tr_class = $rev_color[$current_color];
6914                         $tr_class .= ' boundary' if (exists $meta->{'boundary'});
6915                         $tr_class .= ' no-previous' if ($meta->{'nprevious'} == 0);
6916                         $tr_class .= ' multiple-previous' if ($meta->{'nprevious'} > 1);
6917                         print "<tr id=\"l$lineno\" class=\"$tr_class\">\n";
6918                         if ($group_size) {
6919                                 print "<td class=\"sha1\"";
6920                                 print " title=\"". esc_html($author) . ", $date\"";
6921                                 print " rowspan=\"$group_size\"" if ($group_size > 1);
6922                                 print ">";
6923                                 print $cgi->a({-href => href(action=>"commit",
6924                                                              hash=>$full_rev,
6925                                                              file_name=>$file_name)},
6926                                               esc_html($short_rev));
6927                                 if ($group_size >= 2) {
6928                                         my @author_initials = ($author =~ /\b([[:upper:]])\B/g);
6929                                         if (@author_initials) {
6930                                                 print "<br />" .
6931                                                       esc_html(join('', @author_initials));
6932                                                 #           or join('.', ...)
6933                                         }
6934                                 }
6935                                 print "</td>\n";
6936                         }
6937                         # 'previous' <sha1 of parent commit> <filename at commit>
6938                         if (exists $meta->{'previous'} &&
6939                             $meta->{'previous'} =~ /^($oid_regex) (.*)$/) {
6940                                 $meta->{'parent'} = $1;
6941                                 $meta->{'file_parent'} = unquote($2);
6942                         }
6943                         my $linenr_commit =
6944                                 exists($meta->{'parent'}) ?
6945                                 $meta->{'parent'} : $full_rev;
6946                         my $linenr_filename =
6947                                 exists($meta->{'file_parent'}) ?
6948                                 $meta->{'file_parent'} : unquote($meta->{'filename'});
6949                         my $blamed = href(action => 'blame',
6950                                           file_name => $linenr_filename,
6951                                           hash_base => $linenr_commit);
6952                         print "<td class=\"linenr\">";
6953                         print $cgi->a({ -href => "$blamed#l$orig_lineno",
6954                                         -class => "linenr" },
6955                                       esc_html($lineno));
6956                         print "</td>";
6957                         print "<td class=\"pre\">" . esc_html($data) . "</td>\n";
6958                         print "</tr>\n";
6959                 } # end while
6960
6961         }
6962
6963         # footer
6964         print "</tbody>\n".
6965               "</table>\n"; # class="blame"
6966         print "</div>\n";   # class="blame_body"
6967         close $fd
6968                 or print "Reading blob failed\n";
6969
6970         git_footer_html();
6971 }
6972
6973 sub git_blame {
6974         git_blame_common();
6975 }
6976
6977 sub git_blame_incremental {
6978         git_blame_common('incremental');
6979 }
6980
6981 sub git_blame_data {
6982         git_blame_common('data');
6983 }
6984
6985 sub git_tags {
6986         my $head = git_get_head_hash($project);
6987         git_header_html();
6988         git_print_page_nav('','', $head,undef,$head,format_ref_views('tags'));
6989         git_print_header_div('summary', $project);
6990
6991         my @tagslist = git_get_tags_list();
6992         if (@tagslist) {
6993                 git_tags_body(\@tagslist);
6994         }
6995         git_footer_html();
6996 }
6997
6998 sub git_heads {
6999         my $head = git_get_head_hash($project);
7000         git_header_html();
7001         git_print_page_nav('','', $head,undef,$head,format_ref_views('heads'));
7002         git_print_header_div('summary', $project);
7003
7004         my @headslist = git_get_heads_list();
7005         if (@headslist) {
7006                 git_heads_body(\@headslist, $head);
7007         }
7008         git_footer_html();
7009 }
7010
7011 # used both for single remote view and for list of all the remotes
7012 sub git_remotes {
7013         gitweb_check_feature('remote_heads')
7014                 or die_error(403, "Remote heads view is disabled");
7015
7016         my $head = git_get_head_hash($project);
7017         my $remote = $input_params{'hash'};
7018
7019         my $remotedata = git_get_remotes_list($remote);
7020         die_error(500, "Unable to get remote information") unless defined $remotedata;
7021
7022         unless (%$remotedata) {
7023                 die_error(404, defined $remote ?
7024                         "Remote $remote not found" :
7025                         "No remotes found");
7026         }
7027
7028         git_header_html(undef, undef, -action_extra => $remote);
7029         git_print_page_nav('', '',  $head, undef, $head,
7030                 format_ref_views($remote ? '' : 'remotes'));
7031
7032         fill_remote_heads($remotedata);
7033         if (defined $remote) {
7034                 git_print_header_div('remotes', "$remote remote for $project");
7035                 git_remote_block($remote, $remotedata->{$remote}, undef, $head);
7036         } else {
7037                 git_print_header_div('summary', "$project remotes");
7038                 git_remotes_body($remotedata, undef, $head);
7039         }
7040
7041         git_footer_html();
7042 }
7043
7044 sub git_blob_plain {
7045         my $type = shift;
7046         my $expires;
7047
7048         if (!defined $hash) {
7049                 if (defined $file_name) {
7050                         my $base = $hash_base || git_get_head_hash($project);
7051                         $hash = git_get_hash_by_path($base, $file_name, "blob")
7052                                 or die_error(404, "Cannot find file");
7053                 } else {
7054                         die_error(400, "No file name defined");
7055                 }
7056         } elsif ($hash =~ m/^$oid_regex$/) {
7057                 # blobs defined by non-textual hash id's can be cached
7058                 $expires = "+1d";
7059         }
7060
7061         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
7062                 or die_error(500, "Open git-cat-file blob '$hash' failed");
7063
7064         # content-type (can include charset)
7065         $type = blob_contenttype($fd, $file_name, $type);
7066
7067         # "save as" filename, even when no $file_name is given
7068         my $save_as = "$hash";
7069         if (defined $file_name) {
7070                 $save_as = $file_name;
7071         } elsif ($type =~ m/^text\//) {
7072                 $save_as .= '.txt';
7073         }
7074
7075         # With XSS prevention on, blobs of all types except a few known safe
7076         # ones are served with "Content-Disposition: attachment" to make sure
7077         # they don't run in our security domain.  For certain image types,
7078         # blob view writes an <img> tag referring to blob_plain view, and we
7079         # want to be sure not to break that by serving the image as an
7080         # attachment (though Firefox 3 doesn't seem to care).
7081         my $sandbox = $prevent_xss &&
7082                 $type !~ m!^(?:text/[a-z]+|image/(?:gif|png|jpeg))(?:[ ;]|$)!;
7083
7084         # serve text/* as text/plain
7085         if ($prevent_xss &&
7086             ($type =~ m!^text/[a-z]+\b(.*)$! ||
7087              ($type =~ m!^[a-z]+/[a-z]\+xml\b(.*)$! && -T $fd))) {
7088                 my $rest = $1;
7089                 $rest = defined $rest ? $rest : '';
7090                 $type = "text/plain$rest";
7091         }
7092
7093         print $cgi->header(
7094                 -type => $type,
7095                 -expires => $expires,
7096                 -content_disposition =>
7097                         ($sandbox ? 'attachment' : 'inline')
7098                         . '; filename="' . $save_as . '"');
7099         local $/ = undef;
7100         local *FCGI::Stream::PRINT = $FCGI_Stream_PRINT_raw;
7101         binmode STDOUT, ':raw';
7102         print <$fd>;
7103         binmode STDOUT, ':utf8'; # as set at the beginning of gitweb.cgi
7104         close $fd;
7105 }
7106
7107 sub git_blob {
7108         my $expires;
7109
7110         if (!defined $hash) {
7111                 if (defined $file_name) {
7112                         my $base = $hash_base || git_get_head_hash($project);
7113                         $hash = git_get_hash_by_path($base, $file_name, "blob")
7114                                 or die_error(404, "Cannot find file");
7115                 } else {
7116                         die_error(400, "No file name defined");
7117                 }
7118         } elsif ($hash =~ m/^$oid_regex$/) {
7119                 # blobs defined by non-textual hash id's can be cached
7120                 $expires = "+1d";
7121         }
7122
7123         my $have_blame = gitweb_check_feature('blame');
7124         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
7125                 or die_error(500, "Couldn't cat $file_name, $hash");
7126         my $mimetype = blob_mimetype($fd, $file_name);
7127         # use 'blob_plain' (aka 'raw') view for files that cannot be displayed
7128         if ($mimetype !~ m!^(?:text/|image/(?:gif|png|jpeg)$)! && -B $fd) {
7129                 close $fd;
7130                 return git_blob_plain($mimetype);
7131         }
7132         # we can have blame only for text/* mimetype
7133         $have_blame &&= ($mimetype =~ m!^text/!);
7134
7135         my $highlight = gitweb_check_feature('highlight');
7136         my $syntax = guess_file_syntax($highlight, $file_name);
7137         $fd = run_highlighter($fd, $highlight, $syntax);
7138
7139         git_header_html(undef, $expires);
7140         my $formats_nav = '';
7141         if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7142                 if (defined $file_name) {
7143                         if ($have_blame) {
7144                                 $formats_nav .=
7145                                         $cgi->a({-href => href(action=>"blame", -replay=>1)},
7146                                                 "blame") .
7147                                         " | ";
7148                         }
7149                         $formats_nav .=
7150                                 $cgi->a({-href => href(action=>"history", -replay=>1)},
7151                                         "history") .
7152                                 " | " .
7153                                 $cgi->a({-href => href(action=>"blob_plain", -replay=>1)},
7154                                         "raw") .
7155                                 " | " .
7156                                 $cgi->a({-href => href(action=>"blob",
7157                                                        hash_base=>"HEAD", file_name=>$file_name)},
7158                                         "HEAD");
7159                 } else {
7160                         $formats_nav .=
7161                                 $cgi->a({-href => href(action=>"blob_plain", -replay=>1)},
7162                                         "raw");
7163                 }
7164                 git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
7165                 git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
7166         } else {
7167                 print "<div class=\"page_nav\">\n" .
7168                       "<br/><br/></div>\n" .
7169                       "<div class=\"title\">".esc_html($hash)."</div>\n";
7170         }
7171         git_print_page_path($file_name, "blob", $hash_base);
7172         print "<div class=\"page_body\">\n";
7173         if ($mimetype =~ m!^image/!) {
7174                 print qq!<img class="blob" type="!.esc_attr($mimetype).qq!"!;
7175                 if ($file_name) {
7176                         print qq! alt="!.esc_attr($file_name).qq!" title="!.esc_attr($file_name).qq!"!;
7177                 }
7178                 print qq! src="! .
7179                       esc_attr(href(action=>"blob_plain", hash=>$hash,
7180                            hash_base=>$hash_base, file_name=>$file_name)) .
7181                       qq!" />\n!;
7182         } else {
7183                 my $nr;
7184                 while (my $line = <$fd>) {
7185                         chomp $line;
7186                         $nr++;
7187                         $line = untabify($line);
7188                         printf qq!<div class="pre"><a id="l%i" href="%s#l%i" class="linenr">%4i</a> %s</div>\n!,
7189                                $nr, esc_attr(href(-replay => 1)), $nr, $nr,
7190                                $highlight ? sanitize($line) : esc_html($line, -nbsp=>1);
7191                 }
7192         }
7193         close $fd
7194                 or print "Reading blob failed.\n";
7195         print "</div>";
7196         git_footer_html();
7197 }
7198
7199 sub git_tree {
7200         if (!defined $hash_base) {
7201                 $hash_base = "HEAD";
7202         }
7203         if (!defined $hash) {
7204                 if (defined $file_name) {
7205                         $hash = git_get_hash_by_path($hash_base, $file_name, "tree");
7206                 } else {
7207                         $hash = $hash_base;
7208                 }
7209         }
7210         die_error(404, "No such tree") unless defined($hash);
7211
7212         my $show_sizes = gitweb_check_feature('show-sizes');
7213         my $have_blame = gitweb_check_feature('blame');
7214
7215         my @entries = ();
7216         {
7217                 local $/ = "\0";
7218                 open my $fd, "-|", git_cmd(), "ls-tree", '-z',
7219                         ($show_sizes ? '-l' : ()), @extra_options, $hash
7220                         or die_error(500, "Open git-ls-tree failed");
7221                 @entries = map { chomp; $_ } <$fd>;
7222                 close $fd
7223                         or die_error(404, "Reading tree failed");
7224         }
7225
7226         my $refs = git_get_references();
7227         my $ref = format_ref_marker($refs, $hash_base);
7228         git_header_html();
7229         my $basedir = '';
7230         if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7231                 my @views_nav = ();
7232                 if (defined $file_name) {
7233                         push @views_nav,
7234                                 $cgi->a({-href => href(action=>"history", -replay=>1)},
7235                                         "history"),
7236                                 $cgi->a({-href => href(action=>"tree",
7237                                                        hash_base=>"HEAD", file_name=>$file_name)},
7238                                         "HEAD"),
7239                 }
7240                 my $snapshot_links = format_snapshot_links($hash);
7241                 if (defined $snapshot_links) {
7242                         # FIXME: Should be available when we have no hash base as well.
7243                         push @views_nav, $snapshot_links;
7244                 }
7245                 git_print_page_nav('tree','', $hash_base, undef, undef,
7246                                    join(' | ', @views_nav));
7247                 git_print_header_div('commit', esc_html($co{'title'}) . $ref, $hash_base);
7248         } else {
7249                 undef $hash_base;
7250                 print "<div class=\"page_nav\">\n";
7251                 print "<br/><br/></div>\n";
7252                 print "<div class=\"title\">".esc_html($hash)."</div>\n";
7253         }
7254         if (defined $file_name) {
7255                 $basedir = $file_name;
7256                 if ($basedir ne '' && substr($basedir, -1) ne '/') {
7257                         $basedir .= '/';
7258                 }
7259                 git_print_page_path($file_name, 'tree', $hash_base);
7260         }
7261         print "<div class=\"page_body\">\n";
7262         print "<table class=\"tree\">\n";
7263         my $alternate = 1;
7264         # '..' (top directory) link if possible
7265         if (defined $hash_base &&
7266             defined $file_name && $file_name =~ m![^/]+$!) {
7267                 if ($alternate) {
7268                         print "<tr class=\"dark\">\n";
7269                 } else {
7270                         print "<tr class=\"light\">\n";
7271                 }
7272                 $alternate ^= 1;
7273
7274                 my $up = $file_name;
7275                 $up =~ s!/?[^/]+$!!;
7276                 undef $up unless $up;
7277                 # based on git_print_tree_entry
7278                 print '<td class="mode">' . mode_str('040000') . "</td>\n";
7279                 print '<td class="size">&nbsp;</td>'."\n" if $show_sizes;
7280                 print '<td class="list">';
7281                 print $cgi->a({-href => href(action=>"tree",
7282                                              hash_base=>$hash_base,
7283                                              file_name=>$up)},
7284                               "..");
7285                 print "</td>\n";
7286                 print "<td class=\"link\"></td>\n";
7287
7288                 print "</tr>\n";
7289         }
7290         foreach my $line (@entries) {
7291                 my %t = parse_ls_tree_line($line, -z => 1, -l => $show_sizes);
7292
7293                 if ($alternate) {
7294                         print "<tr class=\"dark\">\n";
7295                 } else {
7296                         print "<tr class=\"light\">\n";
7297                 }
7298                 $alternate ^= 1;
7299
7300                 git_print_tree_entry(\%t, $basedir, $hash_base, $have_blame);
7301
7302                 print "</tr>\n";
7303         }
7304         print "</table>\n" .
7305               "</div>";
7306         git_footer_html();
7307 }
7308
7309 sub sanitize_for_filename {
7310     my $name = shift;
7311
7312     $name =~ s!/!-!g;
7313     $name =~ s/[^[:alnum:]_.-]//g;
7314
7315     return $name;
7316 }
7317
7318 sub snapshot_name {
7319         my ($project, $hash) = @_;
7320
7321         # path/to/project.git  -> project
7322         # path/to/project/.git -> project
7323         my $name = to_utf8($project);
7324         $name =~ s,([^/])/*\.git$,$1,;
7325         $name = sanitize_for_filename(basename($name));
7326
7327         my $ver = $hash;
7328         if ($hash =~ /^[0-9a-fA-F]+$/) {
7329                 # shorten SHA-1 hash
7330                 my $full_hash = git_get_full_hash($project, $hash);
7331                 if ($full_hash =~ /^$hash/ && length($hash) > 7) {
7332                         $ver = git_get_short_hash($project, $hash);
7333                 }
7334         } elsif ($hash =~ m!^refs/tags/(.*)$!) {
7335                 # tags don't need shortened SHA-1 hash
7336                 $ver = $1;
7337         } else {
7338                 # branches and other need shortened SHA-1 hash
7339                 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
7340                 if ($hash =~ m!^refs/($strip_refs|remotes)/(.*)$!) {
7341                         my $ref_dir = (defined $1) ? $1 : '';
7342                         $ver = $2;
7343
7344                         $ref_dir = sanitize_for_filename($ref_dir);
7345                         # for refs neither in heads nor remotes we want to
7346                         # add a ref dir to archive name
7347                         if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
7348                                 $ver = $ref_dir . '-' . $ver;
7349                         }
7350                 }
7351                 $ver .= '-' . git_get_short_hash($project, $hash);
7352         }
7353         # special case of sanitization for filename - we change
7354         # slashes to dots instead of dashes
7355         # in case of hierarchical branch names
7356         $ver =~ s!/!.!g;
7357         $ver =~ s/[^[:alnum:]_.-]//g;
7358
7359         # name = project-version_string
7360         $name = "$name-$ver";
7361
7362         return wantarray ? ($name, $name) : $name;
7363 }
7364
7365 sub exit_if_unmodified_since {
7366         my ($latest_epoch) = @_;
7367         our $cgi;
7368
7369         my $if_modified = $cgi->http('IF_MODIFIED_SINCE');
7370         if (defined $if_modified) {
7371                 my $since;
7372                 if (eval { require HTTP::Date; 1; }) {
7373                         $since = HTTP::Date::str2time($if_modified);
7374                 } elsif (eval { require Time::ParseDate; 1; }) {
7375                         $since = Time::ParseDate::parsedate($if_modified, GMT => 1);
7376                 }
7377                 if (defined $since && $latest_epoch <= $since) {
7378                         my %latest_date = parse_date($latest_epoch);
7379                         print $cgi->header(
7380                                 -last_modified => $latest_date{'rfc2822'},
7381                                 -status => '304 Not Modified');
7382                         goto DONE_GITWEB;
7383                 }
7384         }
7385 }
7386
7387 sub git_snapshot {
7388         my $format = $input_params{'snapshot_format'};
7389         if (!@snapshot_fmts) {
7390                 die_error(403, "Snapshots not allowed");
7391         }
7392         # default to first supported snapshot format
7393         $format ||= $snapshot_fmts[0];
7394         if ($format !~ m/^[a-z0-9]+$/) {
7395                 die_error(400, "Invalid snapshot format parameter");
7396         } elsif (!exists($known_snapshot_formats{$format})) {
7397                 die_error(400, "Unknown snapshot format");
7398         } elsif ($known_snapshot_formats{$format}{'disabled'}) {
7399                 die_error(403, "Snapshot format not allowed");
7400         } elsif (!grep($_ eq $format, @snapshot_fmts)) {
7401                 die_error(403, "Unsupported snapshot format");
7402         }
7403
7404         my $type = git_get_type("$hash^{}");
7405         if (!$type) {
7406                 die_error(404, 'Object does not exist');
7407         }  elsif ($type eq 'blob') {
7408                 die_error(400, 'Object is not a tree-ish');
7409         }
7410
7411         my ($name, $prefix) = snapshot_name($project, $hash);
7412         my $filename = "$name$known_snapshot_formats{$format}{'suffix'}";
7413
7414         my %co = parse_commit($hash);
7415         exit_if_unmodified_since($co{'committer_epoch'}) if %co;
7416
7417         my $cmd = quote_command(
7418                 git_cmd(), 'archive',
7419                 "--format=$known_snapshot_formats{$format}{'format'}",
7420                 "--prefix=$prefix/", $hash);
7421         if (exists $known_snapshot_formats{$format}{'compressor'}) {
7422                 $cmd .= ' | ' . quote_command(@{$known_snapshot_formats{$format}{'compressor'}});
7423         }
7424
7425         $filename =~ s/(["\\])/\\$1/g;
7426         my %latest_date;
7427         if (%co) {
7428                 %latest_date = parse_date($co{'committer_epoch'}, $co{'committer_tz'});
7429         }
7430
7431         print $cgi->header(
7432                 -type => $known_snapshot_formats{$format}{'type'},
7433                 -content_disposition => 'inline; filename="' . $filename . '"',
7434                 %co ? (-last_modified => $latest_date{'rfc2822'}) : (),
7435                 -status => '200 OK');
7436
7437         open my $fd, "-|", $cmd
7438                 or die_error(500, "Execute git-archive failed");
7439         local *FCGI::Stream::PRINT = $FCGI_Stream_PRINT_raw;
7440         binmode STDOUT, ':raw';
7441         print <$fd>;
7442         binmode STDOUT, ':utf8'; # as set at the beginning of gitweb.cgi
7443         close $fd;
7444 }
7445
7446 sub git_log_generic {
7447         my ($fmt_name, $body_subr, $base, $parent, $file_name, $file_hash) = @_;
7448
7449         my $head = git_get_head_hash($project);
7450         if (!defined $base) {
7451                 $base = $head;
7452         }
7453         if (!defined $page) {
7454                 $page = 0;
7455         }
7456         my $refs = git_get_references();
7457
7458         my $commit_hash = $base;
7459         if (defined $parent) {
7460                 $commit_hash = "$parent..$base";
7461         }
7462         my @commitlist =
7463                 parse_commits($commit_hash, 101, (100 * $page),
7464                               defined $file_name ? ($file_name, "--full-history") : ());
7465
7466         my $ftype;
7467         if (!defined $file_hash && defined $file_name) {
7468                 # some commits could have deleted file in question,
7469                 # and not have it in tree, but one of them has to have it
7470                 for (my $i = 0; $i < @commitlist; $i++) {
7471                         $file_hash = git_get_hash_by_path($commitlist[$i]{'id'}, $file_name);
7472                         last if defined $file_hash;
7473                 }
7474         }
7475         if (defined $file_hash) {
7476                 $ftype = git_get_type($file_hash);
7477         }
7478         if (defined $file_name && !defined $ftype) {
7479                 die_error(500, "Unknown type of object");
7480         }
7481         my %co;
7482         if (defined $file_name) {
7483                 %co = parse_commit($base)
7484                         or die_error(404, "Unknown commit object");
7485         }
7486
7487
7488         my $paging_nav = format_paging_nav($fmt_name, $page, $#commitlist >= 100);
7489         my $next_link = '';
7490         if ($#commitlist >= 100) {
7491                 $next_link =
7492                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
7493                                  -accesskey => "n", -title => "Alt-n"}, "next");
7494         }
7495         my $patch_max = gitweb_get_feature('patches');
7496         if ($patch_max && !defined $file_name &&
7497                 !gitweb_check_feature('email-privacy')) {
7498                 if ($patch_max < 0 || @commitlist <= $patch_max) {
7499                         $paging_nav .= " &sdot; " .
7500                                 $cgi->a({-href => href(action=>"patches", -replay=>1)},
7501                                         "patches");
7502                 }
7503         }
7504
7505         git_header_html();
7506         git_print_page_nav($fmt_name,'', $hash,$hash,$hash, $paging_nav);
7507         if (defined $file_name) {
7508                 git_print_header_div('commit', esc_html($co{'title'}), $base);
7509         } else {
7510                 git_print_header_div('summary', $project)
7511         }
7512         git_print_page_path($file_name, $ftype, $hash_base)
7513                 if (defined $file_name);
7514
7515         $body_subr->(\@commitlist, 0, 99, $refs, $next_link,
7516                      $file_name, $file_hash, $ftype);
7517
7518         git_footer_html();
7519 }
7520
7521 sub git_log {
7522         git_log_generic('log', \&git_log_body,
7523                         $hash, $hash_parent);
7524 }
7525
7526 sub git_commit {
7527         $hash ||= $hash_base || "HEAD";
7528         my %co = parse_commit($hash)
7529             or die_error(404, "Unknown commit object");
7530
7531         my $parent  = $co{'parent'};
7532         my $parents = $co{'parents'}; # listref
7533
7534         # we need to prepare $formats_nav before any parameter munging
7535         my $formats_nav;
7536         if (!defined $parent) {
7537                 # --root commitdiff
7538                 $formats_nav .= '(initial)';
7539         } elsif (@$parents == 1) {
7540                 # single parent commit
7541                 $formats_nav .=
7542                         '(parent: ' .
7543                         $cgi->a({-href => href(action=>"commit",
7544                                                hash=>$parent)},
7545                                 esc_html(substr($parent, 0, 7))) .
7546                         ')';
7547         } else {
7548                 # merge commit
7549                 $formats_nav .=
7550                         '(merge: ' .
7551                         join(' ', map {
7552                                 $cgi->a({-href => href(action=>"commit",
7553                                                        hash=>$_)},
7554                                         esc_html(substr($_, 0, 7)));
7555                         } @$parents ) .
7556                         ')';
7557         }
7558         if (gitweb_check_feature('patches') && @$parents <= 1 &&
7559                 !gitweb_check_feature('email-privacy')) {
7560                 $formats_nav .= " | " .
7561                         $cgi->a({-href => href(action=>"patch", -replay=>1)},
7562                                 "patch");
7563         }
7564
7565         if (!defined $parent) {
7566                 $parent = "--root";
7567         }
7568         my @difftree;
7569         open my $fd, "-|", git_cmd(), "diff-tree", '-r', "--no-commit-id",
7570                 @diff_opts,
7571                 (@$parents <= 1 ? $parent : '-c'),
7572                 $hash, "--"
7573                 or die_error(500, "Open git-diff-tree failed");
7574         @difftree = map { chomp; $_ } <$fd>;
7575         close $fd or die_error(404, "Reading git-diff-tree failed");
7576
7577         # non-textual hash id's can be cached
7578         my $expires;
7579         if ($hash =~ m/^$oid_regex$/) {
7580                 $expires = "+1d";
7581         }
7582         my $refs = git_get_references();
7583         my $ref = format_ref_marker($refs, $co{'id'});
7584
7585         git_header_html(undef, $expires);
7586         git_print_page_nav('commit', '',
7587                            $hash, $co{'tree'}, $hash,
7588                            $formats_nav);
7589
7590         if (defined $co{'parent'}) {
7591                 git_print_header_div('commitdiff', esc_html($co{'title'}) . $ref, $hash);
7592         } else {
7593                 git_print_header_div('tree', esc_html($co{'title'}) . $ref, $co{'tree'}, $hash);
7594         }
7595         print "<div class=\"title_text\">\n" .
7596               "<table class=\"object_header\">\n";
7597         git_print_authorship_rows(\%co);
7598         print "<tr><td>commit</td><td class=\"sha1\">$co{'id'}</td></tr>\n";
7599         print "<tr>" .
7600               "<td>tree</td>" .
7601               "<td class=\"sha1\">" .
7602               $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$hash),
7603                        class => "list"}, $co{'tree'}) .
7604               "</td>" .
7605               "<td class=\"link\">" .
7606               $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$hash)},
7607                       "tree");
7608         my $snapshot_links = format_snapshot_links($hash);
7609         if (defined $snapshot_links) {
7610                 print " | " . $snapshot_links;
7611         }
7612         print "</td>" .
7613               "</tr>\n";
7614
7615         foreach my $par (@$parents) {
7616                 print "<tr>" .
7617                       "<td>parent</td>" .
7618                       "<td class=\"sha1\">" .
7619                       $cgi->a({-href => href(action=>"commit", hash=>$par),
7620                                class => "list"}, $par) .
7621                       "</td>" .
7622                       "<td class=\"link\">" .
7623                       $cgi->a({-href => href(action=>"commit", hash=>$par)}, "commit") .
7624                       " | " .
7625                       $cgi->a({-href => href(action=>"commitdiff", hash=>$hash, hash_parent=>$par)}, "diff") .
7626                       "</td>" .
7627                       "</tr>\n";
7628         }
7629         print "</table>".
7630               "</div>\n";
7631
7632         print "<div class=\"page_body\">\n";
7633         git_print_log($co{'comment'});
7634         print "</div>\n";
7635
7636         git_difftree_body(\@difftree, $hash, @$parents);
7637
7638         git_footer_html();
7639 }
7640
7641 sub git_object {
7642         # object is defined by:
7643         # - hash or hash_base alone
7644         # - hash_base and file_name
7645         my $type;
7646
7647         # - hash or hash_base alone
7648         if ($hash || ($hash_base && !defined $file_name)) {
7649                 my $object_id = $hash || $hash_base;
7650
7651                 open my $fd, "-|", quote_command(
7652                         git_cmd(), 'cat-file', '-t', $object_id) . ' 2> /dev/null'
7653                         or die_error(404, "Object does not exist");
7654                 $type = <$fd>;
7655                 defined $type && chomp $type;
7656                 close $fd
7657                         or die_error(404, "Object does not exist");
7658
7659         # - hash_base and file_name
7660         } elsif ($hash_base && defined $file_name) {
7661                 $file_name =~ s,/+$,,;
7662
7663                 system(git_cmd(), "cat-file", '-e', $hash_base) == 0
7664                         or die_error(404, "Base object does not exist");
7665
7666                 # here errors should not happen
7667                 open my $fd, "-|", git_cmd(), "ls-tree", $hash_base, "--", $file_name
7668                         or die_error(500, "Open git-ls-tree failed");
7669                 my $line = <$fd>;
7670                 close $fd;
7671
7672                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
7673                 unless ($line && $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/) {
7674                         die_error(404, "File or directory for given base does not exist");
7675                 }
7676                 $type = $2;
7677                 $hash = $3;
7678         } else {
7679                 die_error(400, "Not enough information to find object");
7680         }
7681
7682         print $cgi->redirect(-uri => href(action=>$type, -full=>1,
7683                                           hash=>$hash, hash_base=>$hash_base,
7684                                           file_name=>$file_name),
7685                              -status => '302 Found');
7686 }
7687
7688 sub git_blobdiff {
7689         my $format = shift || 'html';
7690         my $diff_style = $input_params{'diff_style'} || 'inline';
7691
7692         my $fd;
7693         my @difftree;
7694         my %diffinfo;
7695         my $expires;
7696
7697         # preparing $fd and %diffinfo for git_patchset_body
7698         # new style URI
7699         if (defined $hash_base && defined $hash_parent_base) {
7700                 if (defined $file_name) {
7701                         # read raw output
7702                         open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7703                                 $hash_parent_base, $hash_base,
7704                                 "--", (defined $file_parent ? $file_parent : ()), $file_name
7705                                 or die_error(500, "Open git-diff-tree failed");
7706                         @difftree = map { chomp; $_ } <$fd>;
7707                         close $fd
7708                                 or die_error(404, "Reading git-diff-tree failed");
7709                         @difftree
7710                                 or die_error(404, "Blob diff not found");
7711
7712                 } elsif (defined $hash &&
7713                          $hash =~ $oid_regex) {
7714                         # try to find filename from $hash
7715
7716                         # read filtered raw output
7717                         open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7718                                 $hash_parent_base, $hash_base, "--"
7719                                 or die_error(500, "Open git-diff-tree failed");
7720                         @difftree =
7721                                 # ':100644 100644 03b21826... 3b93d5e7... M     ls-files.c'
7722                                 # $hash == to_id
7723                                 grep { /^:[0-7]{6} [0-7]{6} $oid_regex $hash/ }
7724                                 map { chomp; $_ } <$fd>;
7725                         close $fd
7726                                 or die_error(404, "Reading git-diff-tree failed");
7727                         @difftree
7728                                 or die_error(404, "Blob diff not found");
7729
7730                 } else {
7731                         die_error(400, "Missing one of the blob diff parameters");
7732                 }
7733
7734                 if (@difftree > 1) {
7735                         die_error(400, "Ambiguous blob diff specification");
7736                 }
7737
7738                 %diffinfo = parse_difftree_raw_line($difftree[0]);
7739                 $file_parent ||= $diffinfo{'from_file'} || $file_name;
7740                 $file_name   ||= $diffinfo{'to_file'};
7741
7742                 $hash_parent ||= $diffinfo{'from_id'};
7743                 $hash        ||= $diffinfo{'to_id'};
7744
7745                 # non-textual hash id's can be cached
7746                 if ($hash_base =~ m/^$oid_regex$/ &&
7747                     $hash_parent_base =~ m/^$oid_regex$/) {
7748                         $expires = '+1d';
7749                 }
7750
7751                 # open patch output
7752                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7753                         '-p', ($format eq 'html' ? "--full-index" : ()),
7754                         $hash_parent_base, $hash_base,
7755                         "--", (defined $file_parent ? $file_parent : ()), $file_name
7756                         or die_error(500, "Open git-diff-tree failed");
7757         }
7758
7759         # old/legacy style URI -- not generated anymore since 1.4.3.
7760         if (!%diffinfo) {
7761                 die_error('404 Not Found', "Missing one of the blob diff parameters")
7762         }
7763
7764         # header
7765         if ($format eq 'html') {
7766                 my $formats_nav =
7767                         $cgi->a({-href => href(action=>"blobdiff_plain", -replay=>1)},
7768                                 "raw");
7769                 $formats_nav .= diff_style_nav($diff_style);
7770                 git_header_html(undef, $expires);
7771                 if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7772                         git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
7773                         git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
7774                 } else {
7775                         print "<div class=\"page_nav\"><br/>$formats_nav<br/></div>\n";
7776                         print "<div class=\"title\">".esc_html("$hash vs $hash_parent")."</div>\n";
7777                 }
7778                 if (defined $file_name) {
7779                         git_print_page_path($file_name, "blob", $hash_base);
7780                 } else {
7781                         print "<div class=\"page_path\"></div>\n";
7782                 }
7783
7784         } elsif ($format eq 'plain') {
7785                 print $cgi->header(
7786                         -type => 'text/plain',
7787                         -charset => 'utf-8',
7788                         -expires => $expires,
7789                         -content_disposition => 'inline; filename="' . "$file_name" . '.patch"');
7790
7791                 print "X-Git-Url: " . $cgi->self_url() . "\n\n";
7792
7793         } else {
7794                 die_error(400, "Unknown blobdiff format");
7795         }
7796
7797         # patch
7798         if ($format eq 'html') {
7799                 print "<div class=\"page_body\">\n";
7800
7801                 git_patchset_body($fd, $diff_style,
7802                                   [ \%diffinfo ], $hash_base, $hash_parent_base);
7803                 close $fd;
7804
7805                 print "</div>\n"; # class="page_body"
7806                 git_footer_html();
7807
7808         } else {
7809                 while (my $line = <$fd>) {
7810                         $line =~ s!a/($hash|$hash_parent)!'a/'.esc_path($diffinfo{'from_file'})!eg;
7811                         $line =~ s!b/($hash|$hash_parent)!'b/'.esc_path($diffinfo{'to_file'})!eg;
7812
7813                         print $line;
7814
7815                         last if $line =~ m!^\+\+\+!;
7816                 }
7817                 local $/ = undef;
7818                 print <$fd>;
7819                 close $fd;
7820         }
7821 }
7822
7823 sub git_blobdiff_plain {
7824         git_blobdiff('plain');
7825 }
7826
7827 # assumes that it is added as later part of already existing navigation,
7828 # so it returns "| foo | bar" rather than just "foo | bar"
7829 sub diff_style_nav {
7830         my ($diff_style, $is_combined) = @_;
7831         $diff_style ||= 'inline';
7832
7833         return "" if ($is_combined);
7834
7835         my @styles = (inline => 'inline', 'sidebyside' => 'side by side');
7836         my %styles = @styles;
7837         @styles =
7838                 @styles[ map { $_ * 2 } 0..$#styles/2 ];
7839
7840         return join '',
7841                 map { " | ".$_ }
7842                 map {
7843                         $_ eq $diff_style ? $styles{$_} :
7844                         $cgi->a({-href => href(-replay=>1, diff_style => $_)}, $styles{$_})
7845                 } @styles;
7846 }
7847
7848 sub git_commitdiff {
7849         my %params = @_;
7850         my $format = $params{-format} || 'html';
7851         my $diff_style = $input_params{'diff_style'} || 'inline';
7852
7853         my ($patch_max) = gitweb_get_feature('patches');
7854         if ($format eq 'patch') {
7855                 die_error(403, "Patch view not allowed") unless $patch_max;
7856         }
7857
7858         $hash ||= $hash_base || "HEAD";
7859         my %co = parse_commit($hash)
7860             or die_error(404, "Unknown commit object");
7861
7862         # choose format for commitdiff for merge
7863         if (! defined $hash_parent && @{$co{'parents'}} > 1) {
7864                 $hash_parent = '--cc';
7865         }
7866         # we need to prepare $formats_nav before almost any parameter munging
7867         my $formats_nav;
7868         if ($format eq 'html') {
7869                 $formats_nav =
7870                         $cgi->a({-href => href(action=>"commitdiff_plain", -replay=>1)},
7871                                 "raw");
7872                 if ($patch_max && @{$co{'parents'}} <= 1 &&
7873                         !gitweb_check_feature('email-privacy')) {
7874                         $formats_nav .= " | " .
7875                                 $cgi->a({-href => href(action=>"patch", -replay=>1)},
7876                                         "patch");
7877                 }
7878                 $formats_nav .= diff_style_nav($diff_style, @{$co{'parents'}} > 1);
7879
7880                 if (defined $hash_parent &&
7881                     $hash_parent ne '-c' && $hash_parent ne '--cc') {
7882                         # commitdiff with two commits given
7883                         my $hash_parent_short = $hash_parent;
7884                         if ($hash_parent =~ m/^$oid_regex$/) {
7885                                 $hash_parent_short = substr($hash_parent, 0, 7);
7886                         }
7887                         $formats_nav .=
7888                                 ' (from';
7889                         for (my $i = 0; $i < @{$co{'parents'}}; $i++) {
7890                                 if ($co{'parents'}[$i] eq $hash_parent) {
7891                                         $formats_nav .= ' parent ' . ($i+1);
7892                                         last;
7893                                 }
7894                         }
7895                         $formats_nav .= ': ' .
7896                                 $cgi->a({-href => href(-replay=>1,
7897                                                        hash=>$hash_parent, hash_base=>undef)},
7898                                         esc_html($hash_parent_short)) .
7899                                 ')';
7900                 } elsif (!$co{'parent'}) {
7901                         # --root commitdiff
7902                         $formats_nav .= ' (initial)';
7903                 } elsif (scalar @{$co{'parents'}} == 1) {
7904                         # single parent commit
7905                         $formats_nav .=
7906                                 ' (parent: ' .
7907                                 $cgi->a({-href => href(-replay=>1,
7908                                                        hash=>$co{'parent'}, hash_base=>undef)},
7909                                         esc_html(substr($co{'parent'}, 0, 7))) .
7910                                 ')';
7911                 } else {
7912                         # merge commit
7913                         if ($hash_parent eq '--cc') {
7914                                 $formats_nav .= ' | ' .
7915                                         $cgi->a({-href => href(-replay=>1,
7916                                                                hash=>$hash, hash_parent=>'-c')},
7917                                                 'combined');
7918                         } else { # $hash_parent eq '-c'
7919                                 $formats_nav .= ' | ' .
7920                                         $cgi->a({-href => href(-replay=>1,
7921                                                                hash=>$hash, hash_parent=>'--cc')},
7922                                                 'compact');
7923                         }
7924                         $formats_nav .=
7925                                 ' (merge: ' .
7926                                 join(' ', map {
7927                                         $cgi->a({-href => href(-replay=>1,
7928                                                                hash=>$_, hash_base=>undef)},
7929                                                 esc_html(substr($_, 0, 7)));
7930                                 } @{$co{'parents'}} ) .
7931                                 ')';
7932                 }
7933         }
7934
7935         my $hash_parent_param = $hash_parent;
7936         if (!defined $hash_parent_param) {
7937                 # --cc for multiple parents, --root for parentless
7938                 $hash_parent_param =
7939                         @{$co{'parents'}} > 1 ? '--cc' : $co{'parent'} || '--root';
7940         }
7941
7942         # read commitdiff
7943         my $fd;
7944         my @difftree;
7945         if ($format eq 'html') {
7946                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7947                         "--no-commit-id", "--patch-with-raw", "--full-index",
7948                         $hash_parent_param, $hash, "--"
7949                         or die_error(500, "Open git-diff-tree failed");
7950
7951                 while (my $line = <$fd>) {
7952                         chomp $line;
7953                         # empty line ends raw part of diff-tree output
7954                         last unless $line;
7955                         push @difftree, scalar parse_difftree_raw_line($line);
7956                 }
7957
7958         } elsif ($format eq 'plain') {
7959                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7960                         '-p', $hash_parent_param, $hash, "--"
7961                         or die_error(500, "Open git-diff-tree failed");
7962         } elsif ($format eq 'patch') {
7963                 # For commit ranges, we limit the output to the number of
7964                 # patches specified in the 'patches' feature.
7965                 # For single commits, we limit the output to a single patch,
7966                 # diverging from the git-format-patch default.
7967                 my @commit_spec = ();
7968                 if ($hash_parent) {
7969                         if ($patch_max > 0) {
7970                                 push @commit_spec, "-$patch_max";
7971                         }
7972                         push @commit_spec, '-n', "$hash_parent..$hash";
7973                 } else {
7974                         if ($params{-single}) {
7975                                 push @commit_spec, '-1';
7976                         } else {
7977                                 if ($patch_max > 0) {
7978                                         push @commit_spec, "-$patch_max";
7979                                 }
7980                                 push @commit_spec, "-n";
7981                         }
7982                         push @commit_spec, '--root', $hash;
7983                 }
7984                 open $fd, "-|", git_cmd(), "format-patch", @diff_opts,
7985                         '--encoding=utf8', '--stdout', @commit_spec
7986                         or die_error(500, "Open git-format-patch failed");
7987         } else {
7988                 die_error(400, "Unknown commitdiff format");
7989         }
7990
7991         # non-textual hash id's can be cached
7992         my $expires;
7993         if ($hash =~ m/^$oid_regex$/) {
7994                 $expires = "+1d";
7995         }
7996
7997         # write commit message
7998         if ($format eq 'html') {
7999                 my $refs = git_get_references();
8000                 my $ref = format_ref_marker($refs, $co{'id'});
8001
8002                 git_header_html(undef, $expires);
8003                 git_print_page_nav('commitdiff','', $hash,$co{'tree'},$hash, $formats_nav);
8004                 git_print_header_div('commit', esc_html($co{'title'}) . $ref, $hash);
8005                 print "<div class=\"title_text\">\n" .
8006                       "<table class=\"object_header\">\n";
8007                 git_print_authorship_rows(\%co);
8008                 print "</table>".
8009                       "</div>\n";
8010                 print "<div class=\"page_body\">\n";
8011                 if (@{$co{'comment'}} > 1) {
8012                         print "<div class=\"log\">\n";
8013                         git_print_log($co{'comment'}, -final_empty_line=> 1, -remove_title => 1);
8014                         print "</div>\n"; # class="log"
8015                 }
8016
8017         } elsif ($format eq 'plain') {
8018                 my $refs = git_get_references("tags");
8019                 my $tagname = git_get_rev_name_tags($hash);
8020                 my $filename = basename($project) . "-$hash.patch";
8021
8022                 print $cgi->header(
8023                         -type => 'text/plain',
8024                         -charset => 'utf-8',
8025                         -expires => $expires,
8026                         -content_disposition => 'inline; filename="' . "$filename" . '"');
8027                 my %ad = parse_date($co{'author_epoch'}, $co{'author_tz'});
8028                 print "From: " . to_utf8($co{'author'}) . "\n";
8029                 print "Date: $ad{'rfc2822'} ($ad{'tz_local'})\n";
8030                 print "Subject: " . to_utf8($co{'title'}) . "\n";
8031
8032                 print "X-Git-Tag: $tagname\n" if $tagname;
8033                 print "X-Git-Url: " . $cgi->self_url() . "\n\n";
8034
8035                 foreach my $line (@{$co{'comment'}}) {
8036                         print to_utf8($line) . "\n";
8037                 }
8038                 print "---\n\n";
8039         } elsif ($format eq 'patch') {
8040                 my $filename = basename($project) . "-$hash.patch";
8041
8042                 print $cgi->header(
8043                         -type => 'text/plain',
8044                         -charset => 'utf-8',
8045                         -expires => $expires,
8046                         -content_disposition => 'inline; filename="' . "$filename" . '"');
8047         }
8048
8049         # write patch
8050         if ($format eq 'html') {
8051                 my $use_parents = !defined $hash_parent ||
8052                         $hash_parent eq '-c' || $hash_parent eq '--cc';
8053                 git_difftree_body(\@difftree, $hash,
8054                                   $use_parents ? @{$co{'parents'}} : $hash_parent);
8055                 print "<br/>\n";
8056
8057                 git_patchset_body($fd, $diff_style,
8058                                   \@difftree, $hash,
8059                                   $use_parents ? @{$co{'parents'}} : $hash_parent);
8060                 close $fd;
8061                 print "</div>\n"; # class="page_body"
8062                 git_footer_html();
8063
8064         } elsif ($format eq 'plain') {
8065                 local $/ = undef;
8066                 print <$fd>;
8067                 close $fd
8068                         or print "Reading git-diff-tree failed\n";
8069         } elsif ($format eq 'patch') {
8070                 local $/ = undef;
8071                 print <$fd>;
8072                 close $fd
8073                         or print "Reading git-format-patch failed\n";
8074         }
8075 }
8076
8077 sub git_commitdiff_plain {
8078         git_commitdiff(-format => 'plain');
8079 }
8080
8081 # format-patch-style patches
8082 sub git_patch {
8083         git_commitdiff(-format => 'patch', -single => 1);
8084 }
8085
8086 sub git_patches {
8087         git_commitdiff(-format => 'patch');
8088 }
8089
8090 sub git_history {
8091         git_log_generic('history', \&git_history_body,
8092                         $hash_base, $hash_parent_base,
8093                         $file_name, $hash);
8094 }
8095
8096 sub git_search {
8097         $searchtype ||= 'commit';
8098
8099         # check if appropriate features are enabled
8100         gitweb_check_feature('search')
8101                 or die_error(403, "Search is disabled");
8102         if ($searchtype eq 'pickaxe') {
8103                 # pickaxe may take all resources of your box and run for several minutes
8104                 # with every query - so decide by yourself how public you make this feature
8105                 gitweb_check_feature('pickaxe')
8106                         or die_error(403, "Pickaxe search is disabled");
8107         }
8108         if ($searchtype eq 'grep') {
8109                 # grep search might be potentially CPU-intensive, too
8110                 gitweb_check_feature('grep')
8111                         or die_error(403, "Grep search is disabled");
8112         }
8113
8114         if (!defined $searchtext) {
8115                 die_error(400, "Text field is empty");
8116         }
8117         if (!defined $hash) {
8118                 $hash = git_get_head_hash($project);
8119         }
8120         my %co = parse_commit($hash);
8121         if (!%co) {
8122                 die_error(404, "Unknown commit object");
8123         }
8124         if (!defined $page) {
8125                 $page = 0;
8126         }
8127
8128         if ($searchtype eq 'commit' ||
8129             $searchtype eq 'author' ||
8130             $searchtype eq 'committer') {
8131                 git_search_message(%co);
8132         } elsif ($searchtype eq 'pickaxe') {
8133                 git_search_changes(%co);
8134         } elsif ($searchtype eq 'grep') {
8135                 git_search_files(%co);
8136         } else {
8137                 die_error(400, "Unknown search type");
8138         }
8139 }
8140
8141 sub git_search_help {
8142         git_header_html();
8143         git_print_page_nav('','', $hash,$hash,$hash);
8144         print <<EOT;
8145 <p><strong>Pattern</strong> is by default a normal string that is matched precisely (but without
8146 regard to case, except in the case of pickaxe). However, when you check the <em>re</em> checkbox,
8147 the pattern entered is recognized as the POSIX extended
8148 <a href="https://en.wikipedia.org/wiki/Regular_expression">regular expression</a> (also case
8149 insensitive).</p>
8150 <dl>
8151 <dt><b>commit</b></dt>
8152 <dd>The commit messages and authorship information will be scanned for the given pattern.</dd>
8153 EOT
8154         my $have_grep = gitweb_check_feature('grep');
8155         if ($have_grep) {
8156                 print <<EOT;
8157 <dt><b>grep</b></dt>
8158 <dd>All files in the currently selected tree (HEAD unless you are explicitly browsing
8159     a different one) are searched for the given pattern. On large trees, this search can take
8160 a while and put some strain on the server, so please use it with some consideration. Note that
8161 due to git-grep peculiarity, currently if regexp mode is turned off, the matches are
8162 case-sensitive.</dd>
8163 EOT
8164         }
8165         print <<EOT;
8166 <dt><b>author</b></dt>
8167 <dd>Name and e-mail of the change author and date of birth of the patch will be scanned for the given pattern.</dd>
8168 <dt><b>committer</b></dt>
8169 <dd>Name and e-mail of the committer and date of commit will be scanned for the given pattern.</dd>
8170 EOT
8171         my $have_pickaxe = gitweb_check_feature('pickaxe');
8172         if ($have_pickaxe) {
8173                 print <<EOT;
8174 <dt><b>pickaxe</b></dt>
8175 <dd>All commits that caused the string to appear or disappear from any file (changes that
8176 added, removed or "modified" the string) will be listed. This search can take a while and
8177 takes a lot of strain on the server, so please use it wisely. Note that since you may be
8178 interested even in changes just changing the case as well, this search is case sensitive.</dd>
8179 EOT
8180         }
8181         print "</dl>\n";
8182         git_footer_html();
8183 }
8184
8185 sub git_shortlog {
8186         git_log_generic('shortlog', \&git_shortlog_body,
8187                         $hash, $hash_parent);
8188 }
8189
8190 ## ......................................................................
8191 ## feeds (RSS, Atom; OPML)
8192
8193 sub git_feed {
8194         my $format = shift || 'atom';
8195         my $have_blame = gitweb_check_feature('blame');
8196
8197         # Atom: http://www.atomenabled.org/developers/syndication/
8198         # RSS:  https://web.archive.org/web/20030729001534/http://www.notestips.com/80256B3A007F2692/1/NAMO5P9UPQ
8199         if ($format ne 'rss' && $format ne 'atom') {
8200                 die_error(400, "Unknown web feed format");
8201         }
8202
8203         # log/feed of current (HEAD) branch, log of given branch, history of file/directory
8204         my $head = $hash || 'HEAD';
8205         my @commitlist = parse_commits($head, 150, 0, $file_name);
8206
8207         my %latest_commit;
8208         my %latest_date;
8209         my $content_type = "application/$format+xml";
8210         if (defined $cgi->http('HTTP_ACCEPT') &&
8211                  $cgi->Accept('text/xml') > $cgi->Accept($content_type)) {
8212                 # browser (feed reader) prefers text/xml
8213                 $content_type = 'text/xml';
8214         }
8215         if (defined($commitlist[0])) {
8216                 %latest_commit = %{$commitlist[0]};
8217                 my $latest_epoch = $latest_commit{'committer_epoch'};
8218                 exit_if_unmodified_since($latest_epoch);
8219                 %latest_date = parse_date($latest_epoch, $latest_commit{'committer_tz'});
8220         }
8221         print $cgi->header(
8222                 -type => $content_type,
8223                 -charset => 'utf-8',
8224                 %latest_date ? (-last_modified => $latest_date{'rfc2822'}) : (),
8225                 -status => '200 OK');
8226
8227         # Optimization: skip generating the body if client asks only
8228         # for Last-Modified date.
8229         return if ($cgi->request_method() eq 'HEAD');
8230
8231         # header variables
8232         my $title = "$site_name - $project/$action";
8233         my $feed_type = 'log';
8234         if (defined $hash) {
8235                 $title .= " - '$hash'";
8236                 $feed_type = 'branch log';
8237                 if (defined $file_name) {
8238                         $title .= " :: $file_name";
8239                         $feed_type = 'history';
8240                 }
8241         } elsif (defined $file_name) {
8242                 $title .= " - $file_name";
8243                 $feed_type = 'history';
8244         }
8245         $title .= " $feed_type";
8246         $title = esc_html($title);
8247         my $descr = git_get_project_description($project);
8248         if (defined $descr) {
8249                 $descr = esc_html($descr);
8250         } else {
8251                 $descr = "$project " .
8252                          ($format eq 'rss' ? 'RSS' : 'Atom') .
8253                          " feed";
8254         }
8255         my $owner = git_get_project_owner($project);
8256         $owner = esc_html($owner);
8257
8258         #header
8259         my $alt_url;
8260         if (defined $file_name) {
8261                 $alt_url = href(-full=>1, action=>"history", hash=>$hash, file_name=>$file_name);
8262         } elsif (defined $hash) {
8263                 $alt_url = href(-full=>1, action=>"log", hash=>$hash);
8264         } else {
8265                 $alt_url = href(-full=>1, action=>"summary");
8266         }
8267         $alt_url = esc_attr($alt_url);
8268         print qq!<?xml version="1.0" encoding="utf-8"?>\n!;
8269         if ($format eq 'rss') {
8270                 print <<XML;
8271 <rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/">
8272 <channel>
8273 XML
8274                 print "<title>$title</title>\n" .
8275                       "<link>$alt_url</link>\n" .
8276                       "<description>$descr</description>\n" .
8277                       "<language>en</language>\n" .
8278                       # project owner is responsible for 'editorial' content
8279                       "<managingEditor>$owner</managingEditor>\n";
8280                 if (defined $logo || defined $favicon) {
8281                         # prefer the logo to the favicon, since RSS
8282                         # doesn't allow both
8283                         my $img = esc_url($logo || $favicon);
8284                         print "<image>\n" .
8285                               "<url>$img</url>\n" .
8286                               "<title>$title</title>\n" .
8287                               "<link>$alt_url</link>\n" .
8288                               "</image>\n";
8289                 }
8290                 if (%latest_date) {
8291                         print "<pubDate>$latest_date{'rfc2822'}</pubDate>\n";
8292                         print "<lastBuildDate>$latest_date{'rfc2822'}</lastBuildDate>\n";
8293                 }
8294                 print "<generator>gitweb v.$version/$git_version</generator>\n";
8295         } elsif ($format eq 'atom') {
8296                 print <<XML;
8297 <feed xmlns="http://www.w3.org/2005/Atom">
8298 XML
8299                 print "<title>$title</title>\n" .
8300                       "<subtitle>$descr</subtitle>\n" .
8301                       '<link rel="alternate" type="text/html" href="' .
8302                       $alt_url . '" />' . "\n" .
8303                       '<link rel="self" type="' . $content_type . '" href="' .
8304                       $cgi->self_url() . '" />' . "\n" .
8305                       "<id>" . esc_url(href(-full=>1)) . "</id>\n" .
8306                       # use project owner for feed author
8307                       "<author><name>$owner</name></author>\n";
8308                 if (defined $favicon) {
8309                         print "<icon>" . esc_url($favicon) . "</icon>\n";
8310                 }
8311                 if (defined $logo) {
8312                         # not twice as wide as tall: 72 x 27 pixels
8313                         print "<logo>" . esc_url($logo) . "</logo>\n";
8314                 }
8315                 if (! %latest_date) {
8316                         # dummy date to keep the feed valid until commits trickle in:
8317                         print "<updated>1970-01-01T00:00:00Z</updated>\n";
8318                 } else {
8319                         print "<updated>$latest_date{'iso-8601'}</updated>\n";
8320                 }
8321                 print "<generator version='$version/$git_version'>gitweb</generator>\n";
8322         }
8323
8324         # contents
8325         for (my $i = 0; $i <= $#commitlist; $i++) {
8326                 my %co = %{$commitlist[$i]};
8327                 my $commit = $co{'id'};
8328                 # we read 150, we always show 30 and the ones more recent than 48 hours
8329                 if (($i >= 20) && ((time - $co{'author_epoch'}) > 48*60*60)) {
8330                         last;
8331                 }
8332                 my %cd = parse_date($co{'author_epoch'}, $co{'author_tz'});
8333
8334                 # get list of changed files
8335                 open my $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
8336                         $co{'parent'} || "--root",
8337                         $co{'id'}, "--", (defined $file_name ? $file_name : ())
8338                         or next;
8339                 my @difftree = map { chomp; $_ } <$fd>;
8340                 close $fd
8341                         or next;
8342
8343                 # print element (entry, item)
8344                 my $co_url = href(-full=>1, action=>"commitdiff", hash=>$commit);
8345                 if ($format eq 'rss') {
8346                         print "<item>\n" .
8347                               "<title>" . esc_html($co{'title'}) . "</title>\n" .
8348                               "<author>" . esc_html($co{'author'}) . "</author>\n" .
8349                               "<pubDate>$cd{'rfc2822'}</pubDate>\n" .
8350                               "<guid isPermaLink=\"true\">$co_url</guid>\n" .
8351                               "<link>" . esc_html($co_url) . "</link>\n" .
8352                               "<description>" . esc_html($co{'title'}) . "</description>\n" .
8353                               "<content:encoded>" .
8354                               "<![CDATA[\n";
8355                 } elsif ($format eq 'atom') {
8356                         print "<entry>\n" .
8357                               "<title type=\"html\">" . esc_html($co{'title'}) . "</title>\n" .
8358                               "<updated>$cd{'iso-8601'}</updated>\n" .
8359                               "<author>\n" .
8360                               "  <name>" . esc_html($co{'author_name'}) . "</name>\n";
8361                         if ($co{'author_email'}) {
8362                                 print "  <email>" . esc_html($co{'author_email'}) . "</email>\n";
8363                         }
8364                         print "</author>\n" .
8365                               # use committer for contributor
8366                               "<contributor>\n" .
8367                               "  <name>" . esc_html($co{'committer_name'}) . "</name>\n";
8368                         if ($co{'committer_email'}) {
8369                                 print "  <email>" . esc_html($co{'committer_email'}) . "</email>\n";
8370                         }
8371                         print "</contributor>\n" .
8372                               "<published>$cd{'iso-8601'}</published>\n" .
8373                               "<link rel=\"alternate\" type=\"text/html\" href=\"" . esc_attr($co_url) . "\" />\n" .
8374                               "<id>" . esc_html($co_url) . "</id>\n" .
8375                               "<content type=\"xhtml\" xml:base=\"" . esc_url($my_url) . "\">\n" .
8376                               "<div xmlns=\"http://www.w3.org/1999/xhtml\">\n";
8377                 }
8378                 my $comment = $co{'comment'};
8379                 print "<pre>\n";
8380                 foreach my $line (@$comment) {
8381                         $line = esc_html($line);
8382                         print "$line\n";
8383                 }
8384                 print "</pre><ul>\n";
8385                 foreach my $difftree_line (@difftree) {
8386                         my %difftree = parse_difftree_raw_line($difftree_line);
8387                         next if !$difftree{'from_id'};
8388
8389                         my $file = $difftree{'file'} || $difftree{'to_file'};
8390
8391                         print "<li>" .
8392                               "[" .
8393                               $cgi->a({-href => href(-full=>1, action=>"blobdiff",
8394                                                      hash=>$difftree{'to_id'}, hash_parent=>$difftree{'from_id'},
8395                                                      hash_base=>$co{'id'}, hash_parent_base=>$co{'parent'},
8396                                                      file_name=>$file, file_parent=>$difftree{'from_file'}),
8397                                       -title => "diff"}, 'D');
8398                         if ($have_blame) {
8399                                 print $cgi->a({-href => href(-full=>1, action=>"blame",
8400                                                              file_name=>$file, hash_base=>$commit),
8401                                               -title => "blame"}, 'B');
8402                         }
8403                         # if this is not a feed of a file history
8404                         if (!defined $file_name || $file_name ne $file) {
8405                                 print $cgi->a({-href => href(-full=>1, action=>"history",
8406                                                              file_name=>$file, hash=>$commit),
8407                                               -title => "history"}, 'H');
8408                         }
8409                         $file = esc_path($file);
8410                         print "] ".
8411                               "$file</li>\n";
8412                 }
8413                 if ($format eq 'rss') {
8414                         print "</ul>]]>\n" .
8415                               "</content:encoded>\n" .
8416                               "</item>\n";
8417                 } elsif ($format eq 'atom') {
8418                         print "</ul>\n</div>\n" .
8419                               "</content>\n" .
8420                               "</entry>\n";
8421                 }
8422         }
8423
8424         # end of feed
8425         if ($format eq 'rss') {
8426                 print "</channel>\n</rss>\n";
8427         } elsif ($format eq 'atom') {
8428                 print "</feed>\n";
8429         }
8430 }
8431
8432 sub git_rss {
8433         git_feed('rss');
8434 }
8435
8436 sub git_atom {
8437         git_feed('atom');
8438 }
8439
8440 sub git_opml {
8441         my @list = git_get_projects_list($project_filter, $strict_export);
8442         if (!@list) {
8443                 die_error(404, "No projects found");
8444         }
8445
8446         print $cgi->header(
8447                 -type => 'text/xml',
8448                 -charset => 'utf-8',
8449                 -content_disposition => 'inline; filename="opml.xml"');
8450
8451         my $title = esc_html($site_name);
8452         my $filter = " within subdirectory ";
8453         if (defined $project_filter) {
8454                 $filter .= esc_html($project_filter);
8455         } else {
8456                 $filter = "";
8457         }
8458         print <<XML;
8459 <?xml version="1.0" encoding="utf-8"?>
8460 <opml version="1.0">
8461 <head>
8462   <title>$title OPML Export$filter</title>
8463 </head>
8464 <body>
8465 <outline text="git RSS feeds">
8466 XML
8467
8468         foreach my $pr (@list) {
8469                 my %proj = %$pr;
8470                 my $head = git_get_head_hash($proj{'path'});
8471                 if (!defined $head) {
8472                         next;
8473                 }
8474                 $git_dir = "$projectroot/$proj{'path'}";
8475                 my %co = parse_commit($head);
8476                 if (!%co) {
8477                         next;
8478                 }
8479
8480                 my $path = esc_html(chop_str($proj{'path'}, 25, 5));
8481                 my $rss  = esc_attr(href('project' => $proj{'path'}, 'action' => 'rss', -full => 1));
8482                 my $html = esc_attr(href('project' => $proj{'path'}, 'action' => 'summary', -full => 1));
8483                 print "<outline type=\"rss\" text=\"$path\" title=\"$path\" xmlUrl=\"$rss\" htmlUrl=\"$html\"/>\n";
8484         }
8485         print <<XML;
8486 </outline>
8487 </body>
8488 </opml>
8489 XML
8490 }