Fifth batch
[git.git] / gitweb / gitweb.perl
blob0959a782eccb60f1358f7abf5db534058c37e69e
1 #!/usr/bin/perl
3 # gitweb - simple web interface to track changes in git repositories
5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
6 # (C) 2005, Christian Gierke
8 # This program is licensed under the GPLv2
10 use 5.008;
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);
25 binmode STDOUT, ':utf8';
27 if (!defined($CGI::VERSION) || $CGI::VERSION < 4.08) {
28 eval 'sub CGI::multi_param { CGI::param(@_) }'
31 our $t0 = [ gettimeofday() ];
32 our $number_of_git_cmds = 0;
34 BEGIN {
35 CGI->compile() if $ENV{'MOD_PERL'};
38 our $version = "++GIT_VERSION++";
40 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
41 sub evaluate_uri {
42 our $cgi;
44 our $my_url = $cgi->url();
45 our $my_uri = $cgi->url(-absolute => 1);
47 # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
48 # needed and used only for URLs with nonempty PATH_INFO
49 our $base_url = $my_url;
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.
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'};
77 # target of the home link on top of all pages
78 our $home_link = $my_uri || "/";
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";
85 # absolute fs-path which will be prepended to the project path
86 #our $projectroot = "/pub/scm";
87 our $projectroot = "++GITWEB_PROJECTROOT++";
89 # fs traversing limit for getting project list
90 # the number is relative to the projectroot
91 our $project_maxdepth = "++GITWEB_PROJECT_MAXDEPTH++";
93 # string of the home link on top of all pages
94 our $home_link_str = "++GITWEB_HOME_LINK_STR++";
96 # extra breadcrumbs preceding the home link
97 our @extra_breadcrumbs = ();
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";
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++";
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++";
124 # URI and label (title) of GIT logo link
125 #our $logo_url = "http://www.kernel.org/pub/software/scm/git/docs/";
126 #our $logo_label = "git documentation";
127 our $logo_url = "http://git-scm.com/";
128 our $logo_label = "git homepage";
130 # source of projects list
131 our $projects_list = "++GITWEB_LIST++";
133 # the width (in characters) of the projects list "Description" column
134 our $projects_list_description_width = 25;
136 # group projects by category on the projects list
137 # (enabled if this variable evaluates to true)
138 our $projects_list_group_categories = 0;
140 # default category if none specified
141 # (leave the empty string for no category)
142 our $project_list_default_category = "";
144 # default order of projects list
145 # valid values are none, project, descr, owner, and age
146 our $default_projects_order = "project";
148 # show repository only if this file exists
149 # (only effective if this variable evaluates to true)
150 our $export_ok = "++GITWEB_EXPORT_OK++";
152 # don't generate age column on the projects list page
153 our $omit_age_column = 0;
155 # don't generate information about owners of repositories
156 our $omit_owner=0;
158 # show repository only if this subroutine returns true
159 # when given the path to the project, for example:
160 # sub { return -e "$_[0]/git-daemon-export-ok"; }
161 our $export_auth_hook = undef;
163 # only allow viewing of repositories also shown on the overview page
164 our $strict_export = "++GITWEB_STRICT_EXPORT++";
166 # list of git base URLs used for URL to where fetch project from,
167 # i.e. full URL is "$git_base_url/$project"
168 our @git_base_url_list = grep { $_ ne '' } ("++GITWEB_BASE_URL++");
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;
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;
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';
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
195 # Disables features that would allow repository owners to inject script into
196 # the gitweb domain.
197 our $prevent_xss = 0;
199 # Path to the highlight executable to use (must be the one from
200 # http://www.andre-simon.de 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++";
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)}
216 'tgz' => {
217 'display' => 'tar.gz',
218 'type' => 'application/x-gzip',
219 'suffix' => '.tar.gz',
220 'format' => 'tar',
221 'compressor' => ['gzip', '-n']},
223 'tbz2' => {
224 'display' => 'tar.bz2',
225 'type' => 'application/x-bzip2',
226 'suffix' => '.tar.bz2',
227 'format' => 'tar',
228 'compressor' => ['bzip2']},
230 'txz' => {
231 'display' => 'tar.xz',
232 'type' => 'application/x-xz',
233 'suffix' => '.tar.xz',
234 'format' => 'tar',
235 'compressor' => ['xz'],
236 'disabled' => 1},
238 'zip' => {
239 'display' => 'zip',
240 'type' => 'application/x-zip',
241 'suffix' => '.zip',
242 'format' => 'zip'},
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',
252 # backward compatibility: legacy gitweb config support
253 'x-gzip' => undef, 'gz' => undef,
254 'x-bzip2' => undef, 'bz2' => undef,
255 'x-zip' => undef, '' => undef,
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
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;
272 # configuration for 'highlight' (http://www.andre-simon.de/)
273 # match by basename
274 our %highlight_basename = (
275 #'Program' => 'py',
276 #'Library' => 'py',
277 'SConstruct' => 'py', # SCons equivalent of Makefile
278 'Makefile' => 'make',
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)),
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)}
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
307 # if there is no 'sub' key (no feature-sub), then feature cannot be
308 # overridden
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
314 # Enable the 'blame' blob view, showing the last commit that modified
315 # each line in the file. This can be very CPU-intensive.
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]},
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.
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']},
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.
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]},
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.
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]},
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.
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]},
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.
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]},
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.
405 # To enable system wide have in $GITWEB_CONFIG
406 # $feature{'pathinfo'}{'default'} = [1];
407 # Project specific override is not supported.
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]},
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.
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]},
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.
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 %.
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' => []},
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.
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.
463 # To enable system wide have in $GITWEB_CONFIG
464 # $feature{'ctags'}{'default'} = [1];
465 # Project specific override is not supported.
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]},
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.
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]},
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).
492 # Currently available providers are gravatar and picon.
493 # If an unknown provider is specified, the feature is disabled.
495 # Picon currently relies on the indiana.edu database.
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' => ['']},
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]},
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]},
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
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.
539 # To enable system wide have in $GITWEB_CONFIG
540 # $feature{'highlight'}{'default'} = [1];
542 'highlight' => {
543 'sub' => sub { feature_bool('highlight', @_) },
544 'override' => 0,
545 'default' => [0]},
547 # Enable displaying of remote heads in the heads list
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]},
559 # Enable showing branches under other refs in addition to heads
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 # $feature{'extra-branch-refs'}{'override'} = 1;
565 # and in project config gitweb.extrabranchrefs = dirs of choice
566 # Every directory is separated with whitespace.
568 'extra-branch-refs' => {
569 'sub' => \&feature_extra_branch_refs,
570 'override' => 0,
571 'default' => []},
574 sub gitweb_get_feature {
575 my ($name) = @_;
576 return unless exists $feature{$name};
577 my ($sub, $override, @defaults) = (
578 $feature{$name}{'sub'},
579 $feature{$name}{'override'},
580 @{$feature{$name}{'default'}});
581 # project specific override is possible only if we have project
582 our $git_dir; # global variable, declared later
583 if (!$override || !defined $git_dir) {
584 return @defaults;
586 if (!defined $sub) {
587 warn "feature $name is not overridable";
588 return @defaults;
590 return $sub->(@defaults);
593 # A wrapper to check if a given feature is enabled.
594 # With this, you can say
596 # my $bool_feat = gitweb_check_feature('bool_feat');
597 # gitweb_check_feature('bool_feat') or somecode;
599 # instead of
601 # my ($bool_feat) = gitweb_get_feature('bool_feat');
602 # (gitweb_get_feature('bool_feat'))[0] or somecode;
604 sub gitweb_check_feature {
605 return (gitweb_get_feature(@_))[0];
609 sub feature_bool {
610 my $key = shift;
611 my ($val) = git_get_project_config($key, '--bool');
613 if (!defined $val) {
614 return ($_[0]);
615 } elsif ($val eq 'true') {
616 return (1);
617 } elsif ($val eq 'false') {
618 return (0);
622 sub feature_snapshot {
623 my (@fmts) = @_;
625 my ($val) = git_get_project_config('snapshot');
627 if ($val) {
628 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
631 return @fmts;
634 sub feature_patches {
635 my @val = (git_get_project_config('patches', '--int'));
637 if (@val) {
638 return @val;
641 return ($_[0]);
644 sub feature_avatar {
645 my @val = (git_get_project_config('avatar'));
647 return @val ? @val : @_;
650 sub feature_extra_branch_refs {
651 my (@branch_refs) = @_;
652 my $values = git_get_project_config('extrabranchrefs');
654 if ($values) {
655 $values = config_to_multi ($values);
656 @branch_refs = ();
657 foreach my $value (@{$values}) {
658 push @branch_refs, split /\s+/, $value;
662 return @branch_refs;
665 # checking HEAD file with -e is fragile if the repository was
666 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
667 # and then pruned.
668 sub check_head_link {
669 my ($dir) = @_;
670 my $headfile = "$dir/HEAD";
671 return ((-e $headfile) ||
672 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
675 sub check_export_ok {
676 my ($dir) = @_;
677 return (check_head_link($dir) &&
678 (!$export_ok || -e "$dir/$export_ok") &&
679 (!$export_auth_hook || $export_auth_hook->($dir)));
682 # process alternate names for backward compatibility
683 # filter out unsupported (unknown) snapshot formats
684 sub filter_snapshot_fmts {
685 my @fmts = @_;
687 @fmts = map {
688 exists $known_snapshot_format_aliases{$_} ?
689 $known_snapshot_format_aliases{$_} : $_} @fmts;
690 @fmts = grep {
691 exists $known_snapshot_formats{$_} &&
692 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
695 sub filter_and_validate_refs {
696 my @refs = @_;
697 my %unique_refs = ();
699 foreach my $ref (@refs) {
700 die_error(500, "Invalid ref '$ref' in 'extra-branch-refs' feature") unless (is_valid_ref_format($ref));
701 # 'heads' are added implicitly in get_branch_refs().
702 $unique_refs{$ref} = 1 if ($ref ne 'heads');
704 return sort keys %unique_refs;
707 # If it is set to code reference, it is code that it is to be run once per
708 # request, allowing updating configurations that change with each request,
709 # while running other code in config file only once.
711 # Otherwise, if it is false then gitweb would process config file only once;
712 # if it is true then gitweb config would be run for each request.
713 our $per_request_config = 1;
715 # read and parse gitweb config file given by its parameter.
716 # returns true on success, false on recoverable error, allowing
717 # to chain this subroutine, using first file that exists.
718 # dies on errors during parsing config file, as it is unrecoverable.
719 sub read_config_file {
720 my $filename = shift;
721 return unless defined $filename;
722 # die if there are errors parsing config file
723 if (-e $filename) {
724 do $filename;
725 die $@ if $@;
726 return 1;
728 return;
731 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM, $GITWEB_CONFIG_COMMON);
732 sub evaluate_gitweb_config {
733 our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
734 our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
735 our $GITWEB_CONFIG_COMMON = $ENV{'GITWEB_CONFIG_COMMON'} || "++GITWEB_CONFIG_COMMON++";
737 # Protect against duplications of file names, to not read config twice.
738 # Only one of $GITWEB_CONFIG and $GITWEB_CONFIG_SYSTEM is used, so
739 # there possibility of duplication of filename there doesn't matter.
740 $GITWEB_CONFIG = "" if ($GITWEB_CONFIG eq $GITWEB_CONFIG_COMMON);
741 $GITWEB_CONFIG_SYSTEM = "" if ($GITWEB_CONFIG_SYSTEM eq $GITWEB_CONFIG_COMMON);
743 # Common system-wide settings for convenience.
744 # Those settings can be overridden by GITWEB_CONFIG or GITWEB_CONFIG_SYSTEM.
745 read_config_file($GITWEB_CONFIG_COMMON);
747 # Use first config file that exists. This means use the per-instance
748 # GITWEB_CONFIG if exists, otherwise use GITWEB_SYSTEM_CONFIG.
749 read_config_file($GITWEB_CONFIG) and return;
750 read_config_file($GITWEB_CONFIG_SYSTEM);
753 # Get loadavg of system, to compare against $maxload.
754 # Currently it requires '/proc/loadavg' present to get loadavg;
755 # if it is not present it returns 0, which means no load checking.
756 sub get_loadavg {
757 if( -e '/proc/loadavg' ){
758 open my $fd, '<', '/proc/loadavg'
759 or return 0;
760 my @load = split(/\s+/, scalar <$fd>);
761 close $fd;
763 # The first three columns measure CPU and IO utilization of the last one,
764 # five, and 10 minute periods. The fourth column shows the number of
765 # currently running processes and the total number of processes in the m/n
766 # format. The last column displays the last process ID used.
767 return $load[0] || 0;
769 # additional checks for load average should go here for things that don't export
770 # /proc/loadavg
772 return 0;
775 # version of the core git binary
776 our $git_version;
777 sub evaluate_git_version {
778 our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
779 $number_of_git_cmds++;
782 sub check_loadavg {
783 if (defined $maxload && get_loadavg() > $maxload) {
784 die_error(503, "The load average on the server is too high");
788 # ======================================================================
789 # input validation and dispatch
791 # Various hash size-related values.
792 my $sha1_len = 40;
793 my $sha256_extra_len = 24;
794 my $sha256_len = $sha1_len + $sha256_extra_len;
796 # A regex matching $len hex characters. $len may be a range (e.g. 7,64).
797 sub oid_nlen_regex {
798 my $len = shift;
799 my $hchr = qr/[0-9a-fA-F]/;
800 return qr/(?:(?:$hchr){$len})/;
803 # A regex matching two sets of $nlen hex characters, prefixed by the literal
804 # string $prefix and with the literal string $infix between them.
805 sub oid_nlen_prefix_infix_regex {
806 my $nlen = shift;
807 my $prefix = shift;
808 my $infix = shift;
810 my $rx = oid_nlen_regex($nlen);
812 return qr/^\Q$prefix\E$rx\Q$infix\E$rx$/;
815 # A regex matching a valid object ID.
816 our $oid_regex;
818 my $x = oid_nlen_regex($sha1_len);
819 my $y = oid_nlen_regex($sha256_extra_len);
820 $oid_regex = qr/(?:$x(?:$y)?)/;
823 # input parameters can be collected from a variety of sources (presently, CGI
824 # and PATH_INFO), so we define an %input_params hash that collects them all
825 # together during validation: this allows subsequent uses (e.g. href()) to be
826 # agnostic of the parameter origin
828 our %input_params = ();
830 # input parameters are stored with the long parameter name as key. This will
831 # also be used in the href subroutine to convert parameters to their CGI
832 # equivalent, and since the href() usage is the most frequent one, we store
833 # the name -> CGI key mapping here, instead of the reverse.
835 # XXX: Warning: If you touch this, check the search form for updating,
836 # too.
838 our @cgi_param_mapping = (
839 project => "p",
840 action => "a",
841 file_name => "f",
842 file_parent => "fp",
843 hash => "h",
844 hash_parent => "hp",
845 hash_base => "hb",
846 hash_parent_base => "hpb",
847 page => "pg",
848 order => "o",
849 searchtext => "s",
850 searchtype => "st",
851 snapshot_format => "sf",
852 extra_options => "opt",
853 search_use_regexp => "sr",
854 ctag => "by_tag",
855 diff_style => "ds",
856 project_filter => "pf",
857 # this must be last entry (for manipulation from JavaScript)
858 javascript => "js"
860 our %cgi_param_mapping = @cgi_param_mapping;
862 # we will also need to know the possible actions, for validation
863 our %actions = (
864 "blame" => \&git_blame,
865 "blame_incremental" => \&git_blame_incremental,
866 "blame_data" => \&git_blame_data,
867 "blobdiff" => \&git_blobdiff,
868 "blobdiff_plain" => \&git_blobdiff_plain,
869 "blob" => \&git_blob,
870 "blob_plain" => \&git_blob_plain,
871 "commitdiff" => \&git_commitdiff,
872 "commitdiff_plain" => \&git_commitdiff_plain,
873 "commit" => \&git_commit,
874 "forks" => \&git_forks,
875 "heads" => \&git_heads,
876 "history" => \&git_history,
877 "log" => \&git_log,
878 "patch" => \&git_patch,
879 "patches" => \&git_patches,
880 "remotes" => \&git_remotes,
881 "rss" => \&git_rss,
882 "atom" => \&git_atom,
883 "search" => \&git_search,
884 "search_help" => \&git_search_help,
885 "shortlog" => \&git_shortlog,
886 "summary" => \&git_summary,
887 "tag" => \&git_tag,
888 "tags" => \&git_tags,
889 "tree" => \&git_tree,
890 "snapshot" => \&git_snapshot,
891 "object" => \&git_object,
892 # those below don't need $project
893 "opml" => \&git_opml,
894 "project_list" => \&git_project_list,
895 "project_index" => \&git_project_index,
898 # finally, we have the hash of allowed extra_options for the commands that
899 # allow them
900 our %allowed_options = (
901 "--no-merges" => [ qw(rss atom log shortlog history) ],
904 # fill %input_params with the CGI parameters. All values except for 'opt'
905 # should be single values, but opt can be an array. We should probably
906 # build an array of parameters that can be multi-valued, but since for the time
907 # being it's only this one, we just single it out
908 sub evaluate_query_params {
909 our $cgi;
911 while (my ($name, $symbol) = each %cgi_param_mapping) {
912 if ($symbol eq 'opt') {
913 $input_params{$name} = [ map { decode_utf8($_) } $cgi->multi_param($symbol) ];
914 } else {
915 $input_params{$name} = decode_utf8($cgi->param($symbol));
920 # now read PATH_INFO and update the parameter list for missing parameters
921 sub evaluate_path_info {
922 return if defined $input_params{'project'};
923 return if !$path_info;
924 $path_info =~ s,^/+,,;
925 return if !$path_info;
927 # find which part of PATH_INFO is project
928 my $project = $path_info;
929 $project =~ s,/+$,,;
930 while ($project && !check_head_link("$projectroot/$project")) {
931 $project =~ s,/*[^/]*$,,;
933 return unless $project;
934 $input_params{'project'} = $project;
936 # do not change any parameters if an action is given using the query string
937 return if $input_params{'action'};
938 $path_info =~ s,^\Q$project\E/*,,;
940 # next, check if we have an action
941 my $action = $path_info;
942 $action =~ s,/.*$,,;
943 if (exists $actions{$action}) {
944 $path_info =~ s,^$action/*,,;
945 $input_params{'action'} = $action;
948 # list of actions that want hash_base instead of hash, but can have no
949 # pathname (f) parameter
950 my @wants_base = (
951 'tree',
952 'history',
955 # we want to catch, among others
956 # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
957 my ($parentrefname, $parentpathname, $refname, $pathname) =
958 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
960 # first, analyze the 'current' part
961 if (defined $pathname) {
962 # we got "branch:filename" or "branch:dir/"
963 # we could use git_get_type(branch:pathname), but:
964 # - it needs $git_dir
965 # - it does a git() call
966 # - the convention of terminating directories with a slash
967 # makes it superfluous
968 # - embedding the action in the PATH_INFO would make it even
969 # more superfluous
970 $pathname =~ s,^/+,,;
971 if (!$pathname || substr($pathname, -1) eq "/") {
972 $input_params{'action'} ||= "tree";
973 $pathname =~ s,/$,,;
974 } else {
975 # the default action depends on whether we had parent info
976 # or not
977 if ($parentrefname) {
978 $input_params{'action'} ||= "blobdiff_plain";
979 } else {
980 $input_params{'action'} ||= "blob_plain";
983 $input_params{'hash_base'} ||= $refname;
984 $input_params{'file_name'} ||= $pathname;
985 } elsif (defined $refname) {
986 # we got "branch". In this case we have to choose if we have to
987 # set hash or hash_base.
989 # Most of the actions without a pathname only want hash to be
990 # set, except for the ones specified in @wants_base that want
991 # hash_base instead. It should also be noted that hand-crafted
992 # links having 'history' as an action and no pathname or hash
993 # set will fail, but that happens regardless of PATH_INFO.
994 if (defined $parentrefname) {
995 # if there is parent let the default be 'shortlog' action
996 # (for http://git.example.com/repo.git/A..B links); if there
997 # is no parent, dispatch will detect type of object and set
998 # action appropriately if required (if action is not set)
999 $input_params{'action'} ||= "shortlog";
1001 if ($input_params{'action'} &&
1002 grep { $_ eq $input_params{'action'} } @wants_base) {
1003 $input_params{'hash_base'} ||= $refname;
1004 } else {
1005 $input_params{'hash'} ||= $refname;
1009 # next, handle the 'parent' part, if present
1010 if (defined $parentrefname) {
1011 # a missing pathspec defaults to the 'current' filename, allowing e.g.
1012 # someproject/blobdiff/oldrev..newrev:/filename
1013 if ($parentpathname) {
1014 $parentpathname =~ s,^/+,,;
1015 $parentpathname =~ s,/$,,;
1016 $input_params{'file_parent'} ||= $parentpathname;
1017 } else {
1018 $input_params{'file_parent'} ||= $input_params{'file_name'};
1020 # we assume that hash_parent_base is wanted if a path was specified,
1021 # or if the action wants hash_base instead of hash
1022 if (defined $input_params{'file_parent'} ||
1023 grep { $_ eq $input_params{'action'} } @wants_base) {
1024 $input_params{'hash_parent_base'} ||= $parentrefname;
1025 } else {
1026 $input_params{'hash_parent'} ||= $parentrefname;
1030 # for the snapshot action, we allow URLs in the form
1031 # $project/snapshot/$hash.ext
1032 # where .ext determines the snapshot and gets removed from the
1033 # passed $refname to provide the $hash.
1035 # To be able to tell that $refname includes the format extension, we
1036 # require the following two conditions to be satisfied:
1037 # - the hash input parameter MUST have been set from the $refname part
1038 # of the URL (i.e. they must be equal)
1039 # - the snapshot format MUST NOT have been defined already (e.g. from
1040 # CGI parameter sf)
1041 # It's also useless to try any matching unless $refname has a dot,
1042 # so we check for that too
1043 if (defined $input_params{'action'} &&
1044 $input_params{'action'} eq 'snapshot' &&
1045 defined $refname && index($refname, '.') != -1 &&
1046 $refname eq $input_params{'hash'} &&
1047 !defined $input_params{'snapshot_format'}) {
1048 # We loop over the known snapshot formats, checking for
1049 # extensions. Allowed extensions are both the defined suffix
1050 # (which includes the initial dot already) and the snapshot
1051 # format key itself, with a prepended dot
1052 while (my ($fmt, $opt) = each %known_snapshot_formats) {
1053 my $hash = $refname;
1054 unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
1055 next;
1057 my $sfx = $1;
1058 # a valid suffix was found, so set the snapshot format
1059 # and reset the hash parameter
1060 $input_params{'snapshot_format'} = $fmt;
1061 $input_params{'hash'} = $hash;
1062 # we also set the format suffix to the one requested
1063 # in the URL: this way a request for e.g. .tgz returns
1064 # a .tgz instead of a .tar.gz
1065 $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
1066 last;
1071 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
1072 $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
1073 $searchtext, $search_regexp, $project_filter);
1074 sub evaluate_and_validate_params {
1075 our $action = $input_params{'action'};
1076 if (defined $action) {
1077 if (!is_valid_action($action)) {
1078 die_error(400, "Invalid action parameter");
1082 # parameters which are pathnames
1083 our $project = $input_params{'project'};
1084 if (defined $project) {
1085 if (!is_valid_project($project)) {
1086 undef $project;
1087 die_error(404, "No such project");
1091 our $project_filter = $input_params{'project_filter'};
1092 if (defined $project_filter) {
1093 if (!is_valid_pathname($project_filter)) {
1094 die_error(404, "Invalid project_filter parameter");
1098 our $file_name = $input_params{'file_name'};
1099 if (defined $file_name) {
1100 if (!is_valid_pathname($file_name)) {
1101 die_error(400, "Invalid file parameter");
1105 our $file_parent = $input_params{'file_parent'};
1106 if (defined $file_parent) {
1107 if (!is_valid_pathname($file_parent)) {
1108 die_error(400, "Invalid file parent parameter");
1112 # parameters which are refnames
1113 our $hash = $input_params{'hash'};
1114 if (defined $hash) {
1115 if (!is_valid_refname($hash)) {
1116 die_error(400, "Invalid hash parameter");
1120 our $hash_parent = $input_params{'hash_parent'};
1121 if (defined $hash_parent) {
1122 if (!is_valid_refname($hash_parent)) {
1123 die_error(400, "Invalid hash parent parameter");
1127 our $hash_base = $input_params{'hash_base'};
1128 if (defined $hash_base) {
1129 if (!is_valid_refname($hash_base)) {
1130 die_error(400, "Invalid hash base parameter");
1134 our @extra_options = @{$input_params{'extra_options'}};
1135 # @extra_options is always defined, since it can only be (currently) set from
1136 # CGI, and $cgi->param() returns the empty array in array context if the param
1137 # is not set
1138 foreach my $opt (@extra_options) {
1139 if (not exists $allowed_options{$opt}) {
1140 die_error(400, "Invalid option parameter");
1142 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1143 die_error(400, "Invalid option parameter for this action");
1147 our $hash_parent_base = $input_params{'hash_parent_base'};
1148 if (defined $hash_parent_base) {
1149 if (!is_valid_refname($hash_parent_base)) {
1150 die_error(400, "Invalid hash parent base parameter");
1154 # other parameters
1155 our $page = $input_params{'page'};
1156 if (defined $page) {
1157 if ($page =~ m/[^0-9]/) {
1158 die_error(400, "Invalid page parameter");
1162 our $searchtype = $input_params{'searchtype'};
1163 if (defined $searchtype) {
1164 if ($searchtype =~ m/[^a-z]/) {
1165 die_error(400, "Invalid searchtype parameter");
1169 our $search_use_regexp = $input_params{'search_use_regexp'};
1171 our $searchtext = $input_params{'searchtext'};
1172 our $search_regexp = undef;
1173 if (defined $searchtext) {
1174 if (length($searchtext) < 2) {
1175 die_error(403, "At least two characters are required for search parameter");
1177 if ($search_use_regexp) {
1178 $search_regexp = $searchtext;
1179 if (!eval { qr/$search_regexp/; 1; }) {
1180 (my $error = $@) =~ s/ at \S+ line \d+.*\n?//;
1181 die_error(400, "Invalid search regexp '$search_regexp'",
1182 esc_html($error));
1184 } else {
1185 $search_regexp = quotemeta $searchtext;
1190 # path to the current git repository
1191 our $git_dir;
1192 sub evaluate_git_dir {
1193 our $git_dir = "$projectroot/$project" if $project;
1196 our (@snapshot_fmts, $git_avatar, @extra_branch_refs);
1197 sub configure_gitweb_features {
1198 # list of supported snapshot formats
1199 our @snapshot_fmts = gitweb_get_feature('snapshot');
1200 @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1202 our ($git_avatar) = gitweb_get_feature('avatar');
1203 $git_avatar = '' unless $git_avatar =~ /^(?:gravatar|picon)$/s;
1205 our @extra_branch_refs = gitweb_get_feature('extra-branch-refs');
1206 @extra_branch_refs = filter_and_validate_refs (@extra_branch_refs);
1209 sub get_branch_refs {
1210 return ('heads', @extra_branch_refs);
1213 # custom error handler: 'die <message>' is Internal Server Error
1214 sub handle_errors_html {
1215 my $msg = shift; # it is already HTML escaped
1217 # to avoid infinite loop where error occurs in die_error,
1218 # change handler to default handler, disabling handle_errors_html
1219 set_message("Error occurred when inside die_error:\n$msg");
1221 # you cannot jump out of die_error when called as error handler;
1222 # the subroutine set via CGI::Carp::set_message is called _after_
1223 # HTTP headers are already written, so it cannot write them itself
1224 die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1226 set_message(\&handle_errors_html);
1228 # dispatch
1229 sub dispatch {
1230 if (!defined $action) {
1231 if (defined $hash) {
1232 $action = git_get_type($hash);
1233 $action or die_error(404, "Object does not exist");
1234 } elsif (defined $hash_base && defined $file_name) {
1235 $action = git_get_type("$hash_base:$file_name");
1236 $action or die_error(404, "File or directory does not exist");
1237 } elsif (defined $project) {
1238 $action = 'summary';
1239 } else {
1240 $action = 'project_list';
1243 if (!defined($actions{$action})) {
1244 die_error(400, "Unknown action");
1246 if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1247 !$project) {
1248 die_error(400, "Project needed");
1250 $actions{$action}->();
1253 sub reset_timer {
1254 our $t0 = [ gettimeofday() ]
1255 if defined $t0;
1256 our $number_of_git_cmds = 0;
1259 our $first_request = 1;
1260 sub run_request {
1261 reset_timer();
1263 evaluate_uri();
1264 if ($first_request) {
1265 evaluate_gitweb_config();
1266 evaluate_git_version();
1268 if ($per_request_config) {
1269 if (ref($per_request_config) eq 'CODE') {
1270 $per_request_config->();
1271 } elsif (!$first_request) {
1272 evaluate_gitweb_config();
1275 check_loadavg();
1277 # $projectroot and $projects_list might be set in gitweb config file
1278 $projects_list ||= $projectroot;
1280 evaluate_query_params();
1281 evaluate_path_info();
1282 evaluate_and_validate_params();
1283 evaluate_git_dir();
1285 configure_gitweb_features();
1287 dispatch();
1290 our $is_last_request = sub { 1 };
1291 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1292 our $CGI = 'CGI';
1293 our $cgi;
1294 our $FCGI_Stream_PRINT_raw = \&FCGI::Stream::PRINT;
1295 sub configure_as_fcgi {
1296 require CGI::Fast;
1297 our $CGI = 'CGI::Fast';
1298 # FCGI is not Unicode aware hence the UTF-8 encoding must be done manually.
1299 # However no encoding must be done within git_blob_plain() and git_snapshot()
1300 # which must still output in raw binary mode.
1301 no warnings 'redefine';
1302 my $enc = Encode::find_encoding('UTF-8');
1303 *FCGI::Stream::PRINT = sub {
1304 my @OUTPUT = @_;
1305 for (my $i = 1; $i < @_; $i++) {
1306 $OUTPUT[$i] = $enc->encode($_[$i], Encode::FB_CROAK|Encode::LEAVE_SRC);
1308 @_ = @OUTPUT;
1309 goto $FCGI_Stream_PRINT_raw;
1312 my $request_number = 0;
1313 # let each child service 100 requests
1314 our $is_last_request = sub { ++$request_number > 100 };
1316 sub evaluate_argv {
1317 my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1318 configure_as_fcgi()
1319 if $script_name =~ /\.fcgi$/;
1321 return unless (@ARGV);
1323 require Getopt::Long;
1324 Getopt::Long::GetOptions(
1325 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1326 'nproc|n=i' => sub {
1327 my ($arg, $val) = @_;
1328 return unless eval { require FCGI::ProcManager; 1; };
1329 my $proc_manager = FCGI::ProcManager->new({
1330 n_processes => $val,
1332 our $pre_listen_hook = sub { $proc_manager->pm_manage() };
1333 our $pre_dispatch_hook = sub { $proc_manager->pm_pre_dispatch() };
1334 our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1339 sub run {
1340 evaluate_argv();
1342 $first_request = 1;
1343 $pre_listen_hook->()
1344 if $pre_listen_hook;
1346 REQUEST:
1347 while ($cgi = $CGI->new()) {
1348 $pre_dispatch_hook->()
1349 if $pre_dispatch_hook;
1351 run_request();
1353 $post_dispatch_hook->()
1354 if $post_dispatch_hook;
1355 $first_request = 0;
1357 last REQUEST if ($is_last_request->());
1360 DONE_GITWEB:
1364 run();
1366 if (defined caller) {
1367 # wrapped in a subroutine processing requests,
1368 # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1369 return;
1370 } else {
1371 # pure CGI script, serving single request
1372 exit;
1375 ## ======================================================================
1376 ## action links
1378 # possible values of extra options
1379 # -full => 0|1 - use absolute/full URL ($my_uri/$my_url as base)
1380 # -replay => 1 - start from a current view (replay with modifications)
1381 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1382 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1383 sub href {
1384 my %params = @_;
1385 # default is to use -absolute url() i.e. $my_uri
1386 my $href = $params{-full} ? $my_url : $my_uri;
1388 # implicit -replay, must be first of implicit params
1389 $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1391 $params{'project'} = $project unless exists $params{'project'};
1393 if ($params{-replay}) {
1394 while (my ($name, $symbol) = each %cgi_param_mapping) {
1395 if (!exists $params{$name}) {
1396 $params{$name} = $input_params{$name};
1401 my $use_pathinfo = gitweb_check_feature('pathinfo');
1402 if (defined $params{'project'} &&
1403 (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1404 # try to put as many parameters as possible in PATH_INFO:
1405 # - project name
1406 # - action
1407 # - hash_parent or hash_parent_base:/file_parent
1408 # - hash or hash_base:/filename
1409 # - the snapshot_format as an appropriate suffix
1411 # When the script is the root DirectoryIndex for the domain,
1412 # $href here would be something like http://gitweb.example.com/
1413 # Thus, we strip any trailing / from $href, to spare us double
1414 # slashes in the final URL
1415 $href =~ s,/$,,;
1417 # Then add the project name, if present
1418 $href .= "/".esc_path_info($params{'project'});
1419 delete $params{'project'};
1421 # since we destructively absorb parameters, we keep this
1422 # boolean that remembers if we're handling a snapshot
1423 my $is_snapshot = $params{'action'} eq 'snapshot';
1425 # Summary just uses the project path URL, any other action is
1426 # added to the URL
1427 if (defined $params{'action'}) {
1428 $href .= "/".esc_path_info($params{'action'})
1429 unless $params{'action'} eq 'summary';
1430 delete $params{'action'};
1433 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1434 # stripping nonexistent or useless pieces
1435 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1436 || $params{'hash_parent'} || $params{'hash'});
1437 if (defined $params{'hash_base'}) {
1438 if (defined $params{'hash_parent_base'}) {
1439 $href .= esc_path_info($params{'hash_parent_base'});
1440 # skip the file_parent if it's the same as the file_name
1441 if (defined $params{'file_parent'}) {
1442 if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1443 delete $params{'file_parent'};
1444 } elsif ($params{'file_parent'} !~ /\.\./) {
1445 $href .= ":/".esc_path_info($params{'file_parent'});
1446 delete $params{'file_parent'};
1449 $href .= "..";
1450 delete $params{'hash_parent'};
1451 delete $params{'hash_parent_base'};
1452 } elsif (defined $params{'hash_parent'}) {
1453 $href .= esc_path_info($params{'hash_parent'}). "..";
1454 delete $params{'hash_parent'};
1457 $href .= esc_path_info($params{'hash_base'});
1458 if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1459 $href .= ":/".esc_path_info($params{'file_name'});
1460 delete $params{'file_name'};
1462 delete $params{'hash'};
1463 delete $params{'hash_base'};
1464 } elsif (defined $params{'hash'}) {
1465 $href .= esc_path_info($params{'hash'});
1466 delete $params{'hash'};
1469 # If the action was a snapshot, we can absorb the
1470 # snapshot_format parameter too
1471 if ($is_snapshot) {
1472 my $fmt = $params{'snapshot_format'};
1473 # snapshot_format should always be defined when href()
1474 # is called, but just in case some code forgets, we
1475 # fall back to the default
1476 $fmt ||= $snapshot_fmts[0];
1477 $href .= $known_snapshot_formats{$fmt}{'suffix'};
1478 delete $params{'snapshot_format'};
1482 # now encode the parameters explicitly
1483 my @result = ();
1484 for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1485 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1486 if (defined $params{$name}) {
1487 if (ref($params{$name}) eq "ARRAY") {
1488 foreach my $par (@{$params{$name}}) {
1489 push @result, $symbol . "=" . esc_param($par);
1491 } else {
1492 push @result, $symbol . "=" . esc_param($params{$name});
1496 $href .= "?" . join(';', @result) if scalar @result;
1498 # final transformation: trailing spaces must be escaped (URI-encoded)
1499 $href =~ s/(\s+)$/CGI::escape($1)/e;
1501 if ($params{-anchor}) {
1502 $href .= "#".esc_param($params{-anchor});
1505 return $href;
1509 ## ======================================================================
1510 ## validation, quoting/unquoting and escaping
1512 sub is_valid_action {
1513 my $input = shift;
1514 return undef unless exists $actions{$input};
1515 return 1;
1518 sub is_valid_project {
1519 my $input = shift;
1521 return unless defined $input;
1522 if (!is_valid_pathname($input) ||
1523 !(-d "$projectroot/$input") ||
1524 !check_export_ok("$projectroot/$input") ||
1525 ($strict_export && !project_in_list($input))) {
1526 return undef;
1527 } else {
1528 return 1;
1532 sub is_valid_pathname {
1533 my $input = shift;
1535 return undef unless defined $input;
1536 # no '.' or '..' as elements of path, i.e. no '.' or '..'
1537 # at the beginning, at the end, and between slashes.
1538 # also this catches doubled slashes
1539 if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1540 return undef;
1542 # no null characters
1543 if ($input =~ m!\0!) {
1544 return undef;
1546 return 1;
1549 sub is_valid_ref_format {
1550 my $input = shift;
1552 return undef unless defined $input;
1553 # restrictions on ref name according to git-check-ref-format
1554 if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1555 return undef;
1557 return 1;
1560 sub is_valid_refname {
1561 my $input = shift;
1563 return undef unless defined $input;
1564 # textual hashes are O.K.
1565 if ($input =~ m/^$oid_regex$/) {
1566 return 1;
1568 # it must be correct pathname
1569 is_valid_pathname($input) or return undef;
1570 # check git-check-ref-format restrictions
1571 is_valid_ref_format($input) or return undef;
1572 return 1;
1575 # decode sequences of octets in utf8 into Perl's internal form,
1576 # which is utf-8 with utf8 flag set if needed. gitweb writes out
1577 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1578 sub to_utf8 {
1579 my $str = shift;
1580 return undef unless defined $str;
1582 if (utf8::is_utf8($str) || utf8::decode($str)) {
1583 return $str;
1584 } else {
1585 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1589 # quote unsafe chars, but keep the slash, even when it's not
1590 # correct, but quoted slashes look too horrible in bookmarks
1591 sub esc_param {
1592 my $str = shift;
1593 return undef unless defined $str;
1594 $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1595 $str =~ s/ /\+/g;
1596 return $str;
1599 # the quoting rules for path_info fragment are slightly different
1600 sub esc_path_info {
1601 my $str = shift;
1602 return undef unless defined $str;
1604 # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1605 $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1607 return $str;
1610 # quote unsafe chars in whole URL, so some characters cannot be quoted
1611 sub esc_url {
1612 my $str = shift;
1613 return undef unless defined $str;
1614 $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1615 $str =~ s/ /\+/g;
1616 return $str;
1619 # quote unsafe characters in HTML attributes
1620 sub esc_attr {
1622 # for XHTML conformance escaping '"' to '&quot;' is not enough
1623 return esc_html(@_);
1626 # replace invalid utf8 character with SUBSTITUTION sequence
1627 sub esc_html {
1628 my $str = shift;
1629 my %opts = @_;
1631 return undef unless defined $str;
1633 $str = to_utf8($str);
1634 $str = $cgi->escapeHTML($str);
1635 if ($opts{'-nbsp'}) {
1636 $str =~ s/ /&nbsp;/g;
1638 $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1639 return $str;
1642 # quote control characters and escape filename to HTML
1643 sub esc_path {
1644 my $str = shift;
1645 my %opts = @_;
1647 return undef unless defined $str;
1649 $str = to_utf8($str);
1650 $str = $cgi->escapeHTML($str);
1651 if ($opts{'-nbsp'}) {
1652 $str =~ s/ /&nbsp;/g;
1654 $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1655 return $str;
1658 # Sanitize for use in XHTML + application/xml+xhtml (valid XML 1.0)
1659 sub sanitize {
1660 my $str = shift;
1662 return undef unless defined $str;
1664 $str = to_utf8($str);
1665 $str =~ s|([[:cntrl:]])|(index("\t\n\r", $1) != -1 ? $1 : quot_cec($1))|eg;
1666 return $str;
1669 # Make control characters "printable", using character escape codes (CEC)
1670 sub quot_cec {
1671 my $cntrl = shift;
1672 my %opts = @_;
1673 my %es = ( # character escape codes, aka escape sequences
1674 "\t" => '\t', # tab (HT)
1675 "\n" => '\n', # line feed (LF)
1676 "\r" => '\r', # carriage return (CR)
1677 "\f" => '\f', # form feed (FF)
1678 "\b" => '\b', # backspace (BS)
1679 "\a" => '\a', # alarm (bell) (BEL)
1680 "\e" => '\e', # escape (ESC)
1681 "\013" => '\v', # vertical tab (VT)
1682 "\000" => '\0', # nul character (NUL)
1684 my $chr = ( (exists $es{$cntrl})
1685 ? $es{$cntrl}
1686 : sprintf('\%2x', ord($cntrl)) );
1687 if ($opts{-nohtml}) {
1688 return $chr;
1689 } else {
1690 return "<span class=\"cntrl\">$chr</span>";
1694 # Alternatively use unicode control pictures codepoints,
1695 # Unicode "printable representation" (PR)
1696 sub quot_upr {
1697 my $cntrl = shift;
1698 my %opts = @_;
1700 my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1701 if ($opts{-nohtml}) {
1702 return $chr;
1703 } else {
1704 return "<span class=\"cntrl\">$chr</span>";
1708 # git may return quoted and escaped filenames
1709 sub unquote {
1710 my $str = shift;
1712 sub unq {
1713 my $seq = shift;
1714 my %es = ( # character escape codes, aka escape sequences
1715 't' => "\t", # tab (HT, TAB)
1716 'n' => "\n", # newline (NL)
1717 'r' => "\r", # return (CR)
1718 'f' => "\f", # form feed (FF)
1719 'b' => "\b", # backspace (BS)
1720 'a' => "\a", # alarm (bell) (BEL)
1721 'e' => "\e", # escape (ESC)
1722 'v' => "\013", # vertical tab (VT)
1725 if ($seq =~ m/^[0-7]{1,3}$/) {
1726 # octal char sequence
1727 return chr(oct($seq));
1728 } elsif (exists $es{$seq}) {
1729 # C escape sequence, aka character escape code
1730 return $es{$seq};
1732 # quoted ordinary character
1733 return $seq;
1736 if ($str =~ m/^"(.*)"$/) {
1737 # needs unquoting
1738 $str = $1;
1739 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1741 return $str;
1744 # escape tabs (convert tabs to spaces)
1745 sub untabify {
1746 my $line = shift;
1748 while ((my $pos = index($line, "\t")) != -1) {
1749 if (my $count = (8 - ($pos % 8))) {
1750 my $spaces = ' ' x $count;
1751 $line =~ s/\t/$spaces/;
1755 return $line;
1758 sub project_in_list {
1759 my $project = shift;
1760 my @list = git_get_projects_list();
1761 return @list && scalar(grep { $_->{'path'} eq $project } @list);
1764 ## ----------------------------------------------------------------------
1765 ## HTML aware string manipulation
1767 # Try to chop given string on a word boundary between position
1768 # $len and $len+$add_len. If there is no word boundary there,
1769 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1770 # (marking chopped part) would be longer than given string.
1771 sub chop_str {
1772 my $str = shift;
1773 my $len = shift;
1774 my $add_len = shift || 10;
1775 my $where = shift || 'right'; # 'left' | 'center' | 'right'
1777 # Make sure perl knows it is utf8 encoded so we don't
1778 # cut in the middle of a utf8 multibyte char.
1779 $str = to_utf8($str);
1781 # allow only $len chars, but don't cut a word if it would fit in $add_len
1782 # if it doesn't fit, cut it if it's still longer than the dots we would add
1783 # remove chopped character entities entirely
1785 # when chopping in the middle, distribute $len into left and right part
1786 # return early if chopping wouldn't make string shorter
1787 if ($where eq 'center') {
1788 return $str if ($len + 5 >= length($str)); # filler is length 5
1789 $len = int($len/2);
1790 } else {
1791 return $str if ($len + 4 >= length($str)); # filler is length 4
1794 # regexps: ending and beginning with word part up to $add_len
1795 my $endre = qr/.{$len}\w{0,$add_len}/;
1796 my $begre = qr/\w{0,$add_len}.{$len}/;
1798 if ($where eq 'left') {
1799 $str =~ m/^(.*?)($begre)$/;
1800 my ($lead, $body) = ($1, $2);
1801 if (length($lead) > 4) {
1802 $lead = " ...";
1804 return "$lead$body";
1806 } elsif ($where eq 'center') {
1807 $str =~ m/^($endre)(.*)$/;
1808 my ($left, $str) = ($1, $2);
1809 $str =~ m/^(.*?)($begre)$/;
1810 my ($mid, $right) = ($1, $2);
1811 if (length($mid) > 5) {
1812 $mid = " ... ";
1814 return "$left$mid$right";
1816 } else {
1817 $str =~ m/^($endre)(.*)$/;
1818 my $body = $1;
1819 my $tail = $2;
1820 if (length($tail) > 4) {
1821 $tail = "... ";
1823 return "$body$tail";
1827 # takes the same arguments as chop_str, but also wraps a <span> around the
1828 # result with a title attribute if it does get chopped. Additionally, the
1829 # string is HTML-escaped.
1830 sub chop_and_escape_str {
1831 my ($str) = @_;
1833 my $chopped = chop_str(@_);
1834 $str = to_utf8($str);
1835 if ($chopped eq $str) {
1836 return esc_html($chopped);
1837 } else {
1838 $str =~ s/[[:cntrl:]]/?/g;
1839 return $cgi->span({-title=>$str}, esc_html($chopped));
1843 # Highlight selected fragments of string, using given CSS class,
1844 # and escape HTML. It is assumed that fragments do not overlap.
1845 # Regions are passed as list of pairs (array references).
1847 # Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
1848 # '<span class="mark">foo</span>bar'
1849 sub esc_html_hl_regions {
1850 my ($str, $css_class, @sel) = @_;
1851 my %opts = grep { ref($_) ne 'ARRAY' } @sel;
1852 @sel = grep { ref($_) eq 'ARRAY' } @sel;
1853 return esc_html($str, %opts) unless @sel;
1855 my $out = '';
1856 my $pos = 0;
1858 for my $s (@sel) {
1859 my ($begin, $end) = @$s;
1861 # Don't create empty <span> elements.
1862 next if $end <= $begin;
1864 my $escaped = esc_html(substr($str, $begin, $end - $begin),
1865 %opts);
1867 $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
1868 if ($begin - $pos > 0);
1869 $out .= $cgi->span({-class => $css_class}, $escaped);
1871 $pos = $end;
1873 $out .= esc_html(substr($str, $pos), %opts)
1874 if ($pos < length($str));
1876 return $out;
1879 # return positions of beginning and end of each match
1880 sub matchpos_list {
1881 my ($str, $regexp) = @_;
1882 return unless (defined $str && defined $regexp);
1884 my @matches;
1885 while ($str =~ /$regexp/g) {
1886 push @matches, [$-[0], $+[0]];
1888 return @matches;
1891 # highlight match (if any), and escape HTML
1892 sub esc_html_match_hl {
1893 my ($str, $regexp) = @_;
1894 return esc_html($str) unless defined $regexp;
1896 my @matches = matchpos_list($str, $regexp);
1897 return esc_html($str) unless @matches;
1899 return esc_html_hl_regions($str, 'match', @matches);
1903 # highlight match (if any) of shortened string, and escape HTML
1904 sub esc_html_match_hl_chopped {
1905 my ($str, $chopped, $regexp) = @_;
1906 return esc_html_match_hl($str, $regexp) unless defined $chopped;
1908 my @matches = matchpos_list($str, $regexp);
1909 return esc_html($chopped) unless @matches;
1911 # filter matches so that we mark chopped string
1912 my $tail = "... "; # see chop_str
1913 unless ($chopped =~ s/\Q$tail\E$//) {
1914 $tail = '';
1916 my $chop_len = length($chopped);
1917 my $tail_len = length($tail);
1918 my @filtered;
1920 for my $m (@matches) {
1921 if ($m->[0] > $chop_len) {
1922 push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
1923 last;
1924 } elsif ($m->[1] > $chop_len) {
1925 push @filtered, [ $m->[0], $chop_len + $tail_len ];
1926 last;
1928 push @filtered, $m;
1931 return esc_html_hl_regions($chopped . $tail, 'match', @filtered);
1934 ## ----------------------------------------------------------------------
1935 ## functions returning short strings
1937 # CSS class for given age value (in seconds)
1938 sub age_class {
1939 my $age = shift;
1941 if (!defined $age) {
1942 return "noage";
1943 } elsif ($age < 60*60*2) {
1944 return "age0";
1945 } elsif ($age < 60*60*24*2) {
1946 return "age1";
1947 } else {
1948 return "age2";
1952 # convert age in seconds to "nn units ago" string
1953 sub age_string {
1954 my $age = shift;
1955 my $age_str;
1957 if ($age > 60*60*24*365*2) {
1958 $age_str = (int $age/60/60/24/365);
1959 $age_str .= " years ago";
1960 } elsif ($age > 60*60*24*(365/12)*2) {
1961 $age_str = int $age/60/60/24/(365/12);
1962 $age_str .= " months ago";
1963 } elsif ($age > 60*60*24*7*2) {
1964 $age_str = int $age/60/60/24/7;
1965 $age_str .= " weeks ago";
1966 } elsif ($age > 60*60*24*2) {
1967 $age_str = int $age/60/60/24;
1968 $age_str .= " days ago";
1969 } elsif ($age > 60*60*2) {
1970 $age_str = int $age/60/60;
1971 $age_str .= " hours ago";
1972 } elsif ($age > 60*2) {
1973 $age_str = int $age/60;
1974 $age_str .= " min ago";
1975 } elsif ($age > 2) {
1976 $age_str = int $age;
1977 $age_str .= " sec ago";
1978 } else {
1979 $age_str .= " right now";
1981 return $age_str;
1984 use constant {
1985 S_IFINVALID => 0030000,
1986 S_IFGITLINK => 0160000,
1989 # submodule/subproject, a commit object reference
1990 sub S_ISGITLINK {
1991 my $mode = shift;
1993 return (($mode & S_IFMT) == S_IFGITLINK)
1996 # convert file mode in octal to symbolic file mode string
1997 sub mode_str {
1998 my $mode = oct shift;
2000 if (S_ISGITLINK($mode)) {
2001 return 'm---------';
2002 } elsif (S_ISDIR($mode & S_IFMT)) {
2003 return 'drwxr-xr-x';
2004 } elsif (S_ISLNK($mode)) {
2005 return 'lrwxrwxrwx';
2006 } elsif (S_ISREG($mode)) {
2007 # git cares only about the executable bit
2008 if ($mode & S_IXUSR) {
2009 return '-rwxr-xr-x';
2010 } else {
2011 return '-rw-r--r--';
2013 } else {
2014 return '----------';
2018 # convert file mode in octal to file type string
2019 sub file_type {
2020 my $mode = shift;
2022 if ($mode !~ m/^[0-7]+$/) {
2023 return $mode;
2024 } else {
2025 $mode = oct $mode;
2028 if (S_ISGITLINK($mode)) {
2029 return "submodule";
2030 } elsif (S_ISDIR($mode & S_IFMT)) {
2031 return "directory";
2032 } elsif (S_ISLNK($mode)) {
2033 return "symlink";
2034 } elsif (S_ISREG($mode)) {
2035 return "file";
2036 } else {
2037 return "unknown";
2041 # convert file mode in octal to file type description string
2042 sub file_type_long {
2043 my $mode = shift;
2045 if ($mode !~ m/^[0-7]+$/) {
2046 return $mode;
2047 } else {
2048 $mode = oct $mode;
2051 if (S_ISGITLINK($mode)) {
2052 return "submodule";
2053 } elsif (S_ISDIR($mode & S_IFMT)) {
2054 return "directory";
2055 } elsif (S_ISLNK($mode)) {
2056 return "symlink";
2057 } elsif (S_ISREG($mode)) {
2058 if ($mode & S_IXUSR) {
2059 return "executable";
2060 } else {
2061 return "file";
2063 } else {
2064 return "unknown";
2069 ## ----------------------------------------------------------------------
2070 ## functions returning short HTML fragments, or transforming HTML fragments
2071 ## which don't belong to other sections
2073 # format line of commit message.
2074 sub format_log_line_html {
2075 my $line = shift;
2077 # Potentially abbreviated OID.
2078 my $regex = oid_nlen_regex("7,64");
2080 $line = esc_html($line, -nbsp=>1);
2081 $line =~ s{
2084 # The output of "git describe", e.g. v2.10.0-297-gf6727b0
2085 # or hadoop-20160921-113441-20-g094fb7d
2086 (?<!-) # see strbuf_check_tag_ref(). Tags can't start with -
2087 [A-Za-z0-9.-]+
2088 (?!\.) # refs can't end with ".", see check_refname_format()
2089 -g$regex
2091 # Just a normal looking Git SHA1
2092 $regex
2096 $cgi->a({-href => href(action=>"object", hash=>$1),
2097 -class => "text"}, $1);
2098 }egx;
2100 return $line;
2103 # format marker of refs pointing to given object
2105 # the destination action is chosen based on object type and current context:
2106 # - for annotated tags, we choose the tag view unless it's the current view
2107 # already, in which case we go to shortlog view
2108 # - for other refs, we keep the current view if we're in history, shortlog or
2109 # log view, and select shortlog otherwise
2110 sub format_ref_marker {
2111 my ($refs, $id) = @_;
2112 my $markers = '';
2114 if (defined $refs->{$id}) {
2115 foreach my $ref (@{$refs->{$id}}) {
2116 # this code exploits the fact that non-lightweight tags are the
2117 # only indirect objects, and that they are the only objects for which
2118 # we want to use tag instead of shortlog as action
2119 my ($type, $name) = qw();
2120 my $indirect = ($ref =~ s/\^\{\}$//);
2121 # e.g. tags/v2.6.11 or heads/next
2122 if ($ref =~ m!^(.*?)s?/(.*)$!) {
2123 $type = $1;
2124 $name = $2;
2125 } else {
2126 $type = "ref";
2127 $name = $ref;
2130 my $class = $type;
2131 $class .= " indirect" if $indirect;
2133 my $dest_action = "shortlog";
2135 if ($indirect) {
2136 $dest_action = "tag" unless $action eq "tag";
2137 } elsif ($action =~ /^(history|(short)?log)$/) {
2138 $dest_action = $action;
2141 my $dest = "";
2142 $dest .= "refs/" unless $ref =~ m!^refs/!;
2143 $dest .= $ref;
2145 my $link = $cgi->a({
2146 -href => href(
2147 action=>$dest_action,
2148 hash=>$dest
2149 )}, esc_html($name));
2151 $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
2152 $link . "</span>";
2156 if ($markers) {
2157 return ' <span class="refs">'. $markers . '</span>';
2158 } else {
2159 return "";
2163 # format, perhaps shortened and with markers, title line
2164 sub format_subject_html {
2165 my ($long, $short, $href, $extra) = @_;
2166 $extra = '' unless defined($extra);
2168 if (length($short) < length($long)) {
2169 $long =~ s/[[:cntrl:]]/?/g;
2170 return $cgi->a({-href => $href, -class => "list subject",
2171 -title => to_utf8($long)},
2172 esc_html($short)) . $extra;
2173 } else {
2174 return $cgi->a({-href => $href, -class => "list subject"},
2175 esc_html($long)) . $extra;
2179 # Rather than recomputing the url for an email multiple times, we cache it
2180 # after the first hit. This gives a visible benefit in views where the avatar
2181 # for the same email is used repeatedly (e.g. shortlog).
2182 # The cache is shared by all avatar engines (currently gravatar only), which
2183 # are free to use it as preferred. Since only one avatar engine is used for any
2184 # given page, there's no risk for cache conflicts.
2185 our %avatar_cache = ();
2187 # Compute the picon url for a given email, by using the picon search service over at
2188 # http://www.cs.indiana.edu/picons/search.html
2189 sub picon_url {
2190 my $email = lc shift;
2191 if (!$avatar_cache{$email}) {
2192 my ($user, $domain) = split('@', $email);
2193 $avatar_cache{$email} =
2194 "//www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
2195 "$domain/$user/" .
2196 "users+domains+unknown/up/single";
2198 return $avatar_cache{$email};
2201 # Compute the gravatar url for a given email, if it's not in the cache already.
2202 # Gravatar stores only the part of the URL before the size, since that's the
2203 # one computationally more expensive. This also allows reuse of the cache for
2204 # different sizes (for this particular engine).
2205 sub gravatar_url {
2206 my $email = lc shift;
2207 my $size = shift;
2208 $avatar_cache{$email} ||=
2209 "//www.gravatar.com/avatar/" .
2210 md5_hex($email) . "?s=";
2211 return $avatar_cache{$email} . $size;
2214 # Insert an avatar for the given $email at the given $size if the feature
2215 # is enabled.
2216 sub git_get_avatar {
2217 my ($email, %opts) = @_;
2218 my $pre_white = ($opts{-pad_before} ? "&nbsp;" : "");
2219 my $post_white = ($opts{-pad_after} ? "&nbsp;" : "");
2220 $opts{-size} ||= 'default';
2221 my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
2222 my $url = "";
2223 if ($git_avatar eq 'gravatar') {
2224 $url = gravatar_url($email, $size);
2225 } elsif ($git_avatar eq 'picon') {
2226 $url = picon_url($email);
2228 # Other providers can be added by extending the if chain, defining $url
2229 # as needed. If no variant puts something in $url, we assume avatars
2230 # are completely disabled/unavailable.
2231 if ($url) {
2232 return $pre_white .
2233 "<img width=\"$size\" " .
2234 "class=\"avatar\" " .
2235 "src=\"".esc_url($url)."\" " .
2236 "alt=\"\" " .
2237 "/>" . $post_white;
2238 } else {
2239 return "";
2243 sub format_search_author {
2244 my ($author, $searchtype, $displaytext) = @_;
2245 my $have_search = gitweb_check_feature('search');
2247 if ($have_search) {
2248 my $performed = "";
2249 if ($searchtype eq 'author') {
2250 $performed = "authored";
2251 } elsif ($searchtype eq 'committer') {
2252 $performed = "committed";
2255 return $cgi->a({-href => href(action=>"search", hash=>$hash,
2256 searchtext=>$author,
2257 searchtype=>$searchtype), class=>"list",
2258 title=>"Search for commits $performed by $author"},
2259 $displaytext);
2261 } else {
2262 return $displaytext;
2266 # format the author name of the given commit with the given tag
2267 # the author name is chopped and escaped according to the other
2268 # optional parameters (see chop_str).
2269 sub format_author_html {
2270 my $tag = shift;
2271 my $co = shift;
2272 my $author = chop_and_escape_str($co->{'author_name'}, @_);
2273 return "<$tag class=\"author\">" .
2274 format_search_author($co->{'author_name'}, "author",
2275 git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2276 $author) .
2277 "</$tag>";
2280 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2281 sub format_git_diff_header_line {
2282 my $line = shift;
2283 my $diffinfo = shift;
2284 my ($from, $to) = @_;
2286 if ($diffinfo->{'nparents'}) {
2287 # combined diff
2288 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2289 if ($to->{'href'}) {
2290 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2291 esc_path($to->{'file'}));
2292 } else { # file was deleted (no href)
2293 $line .= esc_path($to->{'file'});
2295 } else {
2296 # "ordinary" diff
2297 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2298 if ($from->{'href'}) {
2299 $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2300 'a/' . esc_path($from->{'file'}));
2301 } else { # file was added (no href)
2302 $line .= 'a/' . esc_path($from->{'file'});
2304 $line .= ' ';
2305 if ($to->{'href'}) {
2306 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2307 'b/' . esc_path($to->{'file'}));
2308 } else { # file was deleted
2309 $line .= 'b/' . esc_path($to->{'file'});
2313 return "<div class=\"diff header\">$line</div>\n";
2316 # format extended diff header line, before patch itself
2317 sub format_extended_diff_header_line {
2318 my $line = shift;
2319 my $diffinfo = shift;
2320 my ($from, $to) = @_;
2322 # match <path>
2323 if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2324 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2325 esc_path($from->{'file'}));
2327 if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2328 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2329 esc_path($to->{'file'}));
2331 # match single <mode>
2332 if ($line =~ m/\s(\d{6})$/) {
2333 $line .= '<span class="info"> (' .
2334 file_type_long($1) .
2335 ')</span>';
2337 # match <hash>
2338 if ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", ",") |
2339 $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", ",")) {
2340 # can match only for combined diff
2341 $line = 'index ';
2342 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2343 if ($from->{'href'}[$i]) {
2344 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2345 -class=>"hash"},
2346 substr($diffinfo->{'from_id'}[$i],0,7));
2347 } else {
2348 $line .= '0' x 7;
2350 # separator
2351 $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2353 $line .= '..';
2354 if ($to->{'href'}) {
2355 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2356 substr($diffinfo->{'to_id'},0,7));
2357 } else {
2358 $line .= '0' x 7;
2361 } elsif ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", "..") |
2362 $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", "..")) {
2363 # can match only for ordinary diff
2364 my ($from_link, $to_link);
2365 if ($from->{'href'}) {
2366 $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2367 substr($diffinfo->{'from_id'},0,7));
2368 } else {
2369 $from_link = '0' x 7;
2371 if ($to->{'href'}) {
2372 $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2373 substr($diffinfo->{'to_id'},0,7));
2374 } else {
2375 $to_link = '0' x 7;
2377 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2378 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2381 return $line . "<br/>\n";
2384 # format from-file/to-file diff header
2385 sub format_diff_from_to_header {
2386 my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2387 my $line;
2388 my $result = '';
2390 $line = $from_line;
2391 #assert($line =~ m/^---/) if DEBUG;
2392 # no extra formatting for "^--- /dev/null"
2393 if (! $diffinfo->{'nparents'}) {
2394 # ordinary (single parent) diff
2395 if ($line =~ m!^--- "?a/!) {
2396 if ($from->{'href'}) {
2397 $line = '--- a/' .
2398 $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2399 esc_path($from->{'file'}));
2400 } else {
2401 $line = '--- a/' .
2402 esc_path($from->{'file'});
2405 $result .= qq!<div class="diff from_file">$line</div>\n!;
2407 } else {
2408 # combined diff (merge commit)
2409 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2410 if ($from->{'href'}[$i]) {
2411 $line = '--- ' .
2412 $cgi->a({-href=>href(action=>"blobdiff",
2413 hash_parent=>$diffinfo->{'from_id'}[$i],
2414 hash_parent_base=>$parents[$i],
2415 file_parent=>$from->{'file'}[$i],
2416 hash=>$diffinfo->{'to_id'},
2417 hash_base=>$hash,
2418 file_name=>$to->{'file'}),
2419 -class=>"path",
2420 -title=>"diff" . ($i+1)},
2421 $i+1) .
2422 '/' .
2423 $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2424 esc_path($from->{'file'}[$i]));
2425 } else {
2426 $line = '--- /dev/null';
2428 $result .= qq!<div class="diff from_file">$line</div>\n!;
2432 $line = $to_line;
2433 #assert($line =~ m/^\+\+\+/) if DEBUG;
2434 # no extra formatting for "^+++ /dev/null"
2435 if ($line =~ m!^\+\+\+ "?b/!) {
2436 if ($to->{'href'}) {
2437 $line = '+++ b/' .
2438 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2439 esc_path($to->{'file'}));
2440 } else {
2441 $line = '+++ b/' .
2442 esc_path($to->{'file'});
2445 $result .= qq!<div class="diff to_file">$line</div>\n!;
2447 return $result;
2450 # create note for patch simplified by combined diff
2451 sub format_diff_cc_simplified {
2452 my ($diffinfo, @parents) = @_;
2453 my $result = '';
2455 $result .= "<div class=\"diff header\">" .
2456 "diff --cc ";
2457 if (!is_deleted($diffinfo)) {
2458 $result .= $cgi->a({-href => href(action=>"blob",
2459 hash_base=>$hash,
2460 hash=>$diffinfo->{'to_id'},
2461 file_name=>$diffinfo->{'to_file'}),
2462 -class => "path"},
2463 esc_path($diffinfo->{'to_file'}));
2464 } else {
2465 $result .= esc_path($diffinfo->{'to_file'});
2467 $result .= "</div>\n" . # class="diff header"
2468 "<div class=\"diff nodifferences\">" .
2469 "Simple merge" .
2470 "</div>\n"; # class="diff nodifferences"
2472 return $result;
2475 sub diff_line_class {
2476 my ($line, $from, $to) = @_;
2478 # ordinary diff
2479 my $num_sign = 1;
2480 # combined diff
2481 if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2482 $num_sign = scalar @{$from->{'href'}};
2485 my @diff_line_classifier = (
2486 { regexp => qr/^\@\@{$num_sign} /, class => "chunk_header"},
2487 { regexp => qr/^\\/, class => "incomplete" },
2488 { regexp => qr/^ {$num_sign}/, class => "ctx" },
2489 # classifier for context must come before classifier add/rem,
2490 # or we would have to use more complicated regexp, for example
2491 # qr/(?= {0,$m}\+)[+ ]{$num_sign}/, where $m = $num_sign - 1;
2492 { regexp => qr/^[+ ]{$num_sign}/, class => "add" },
2493 { regexp => qr/^[- ]{$num_sign}/, class => "rem" },
2495 for my $clsfy (@diff_line_classifier) {
2496 return $clsfy->{'class'}
2497 if ($line =~ $clsfy->{'regexp'});
2500 # fallback
2501 return "";
2504 # assumes that $from and $to are defined and correctly filled,
2505 # and that $line holds a line of chunk header for unified diff
2506 sub format_unidiff_chunk_header {
2507 my ($line, $from, $to) = @_;
2509 my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2510 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2512 $from_lines = 0 unless defined $from_lines;
2513 $to_lines = 0 unless defined $to_lines;
2515 if ($from->{'href'}) {
2516 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2517 -class=>"list"}, $from_text);
2519 if ($to->{'href'}) {
2520 $to_text = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2521 -class=>"list"}, $to_text);
2523 $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2524 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2525 return $line;
2528 # assumes that $from and $to are defined and correctly filled,
2529 # and that $line holds a line of chunk header for combined diff
2530 sub format_cc_diff_chunk_header {
2531 my ($line, $from, $to) = @_;
2533 my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2534 my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2536 @from_text = split(' ', $ranges);
2537 for (my $i = 0; $i < @from_text; ++$i) {
2538 ($from_start[$i], $from_nlines[$i]) =
2539 (split(',', substr($from_text[$i], 1)), 0);
2542 $to_text = pop @from_text;
2543 $to_start = pop @from_start;
2544 $to_nlines = pop @from_nlines;
2546 $line = "<span class=\"chunk_info\">$prefix ";
2547 for (my $i = 0; $i < @from_text; ++$i) {
2548 if ($from->{'href'}[$i]) {
2549 $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2550 -class=>"list"}, $from_text[$i]);
2551 } else {
2552 $line .= $from_text[$i];
2554 $line .= " ";
2556 if ($to->{'href'}) {
2557 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2558 -class=>"list"}, $to_text);
2559 } else {
2560 $line .= $to_text;
2562 $line .= " $prefix</span>" .
2563 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2564 return $line;
2567 # process patch (diff) line (not to be used for diff headers),
2568 # returning HTML-formatted (but not wrapped) line.
2569 # If the line is passed as a reference, it is treated as HTML and not
2570 # esc_html()'ed.
2571 sub format_diff_line {
2572 my ($line, $diff_class, $from, $to) = @_;
2574 if (ref($line)) {
2575 $line = $$line;
2576 } else {
2577 chomp $line;
2578 $line = untabify($line);
2580 if ($from && $to && $line =~ m/^\@{2} /) {
2581 $line = format_unidiff_chunk_header($line, $from, $to);
2582 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2583 $line = format_cc_diff_chunk_header($line, $from, $to);
2584 } else {
2585 $line = esc_html($line, -nbsp=>1);
2589 my $diff_classes = "diff";
2590 $diff_classes .= " $diff_class" if ($diff_class);
2591 $line = "<div class=\"$diff_classes\">$line</div>\n";
2593 return $line;
2596 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2597 # linked. Pass the hash of the tree/commit to snapshot.
2598 sub format_snapshot_links {
2599 my ($hash) = @_;
2600 my $num_fmts = @snapshot_fmts;
2601 if ($num_fmts > 1) {
2602 # A parenthesized list of links bearing format names.
2603 # e.g. "snapshot (_tar.gz_ _zip_)"
2604 return "snapshot (" . join(' ', map
2605 $cgi->a({
2606 -href => href(
2607 action=>"snapshot",
2608 hash=>$hash,
2609 snapshot_format=>$_
2611 }, $known_snapshot_formats{$_}{'display'})
2612 , @snapshot_fmts) . ")";
2613 } elsif ($num_fmts == 1) {
2614 # A single "snapshot" link whose tooltip bears the format name.
2615 # i.e. "_snapshot_"
2616 my ($fmt) = @snapshot_fmts;
2617 return
2618 $cgi->a({
2619 -href => href(
2620 action=>"snapshot",
2621 hash=>$hash,
2622 snapshot_format=>$fmt
2624 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2625 }, "snapshot");
2626 } else { # $num_fmts == 0
2627 return undef;
2631 ## ......................................................................
2632 ## functions returning values to be passed, perhaps after some
2633 ## transformation, to other functions; e.g. returning arguments to href()
2635 # returns hash to be passed to href to generate gitweb URL
2636 # in -title key it returns description of link
2637 sub get_feed_info {
2638 my $format = shift || 'Atom';
2639 my %res = (action => lc($format));
2640 my $matched_ref = 0;
2642 # feed links are possible only for project views
2643 return unless (defined $project);
2644 # some views should link to OPML, or to generic project feed,
2645 # or don't have specific feed yet (so they should use generic)
2646 return if (!$action || $action =~ /^(?:tags|heads|forks|tag|search)$/x);
2648 my $branch = undef;
2649 # branches refs uses 'refs/' + $get_branch_refs()[x] + '/' prefix
2650 # (fullname) to differentiate from tag links; this also makes
2651 # possible to detect branch links
2652 for my $ref (get_branch_refs()) {
2653 if ((defined $hash_base && $hash_base =~ m!^refs/\Q$ref\E/(.*)$!) ||
2654 (defined $hash && $hash =~ m!^refs/\Q$ref\E/(.*)$!)) {
2655 $branch = $1;
2656 $matched_ref = $ref;
2657 last;
2660 # find log type for feed description (title)
2661 my $type = 'log';
2662 if (defined $file_name) {
2663 $type = "history of $file_name";
2664 $type .= "/" if ($action eq 'tree');
2665 $type .= " on '$branch'" if (defined $branch);
2666 } else {
2667 $type = "log of $branch" if (defined $branch);
2670 $res{-title} = $type;
2671 $res{'hash'} = (defined $branch ? "refs/$matched_ref/$branch" : undef);
2672 $res{'file_name'} = $file_name;
2674 return %res;
2677 ## ----------------------------------------------------------------------
2678 ## git utility subroutines, invoking git commands
2680 # returns path to the core git executable and the --git-dir parameter as list
2681 sub git_cmd {
2682 $number_of_git_cmds++;
2683 return $GIT, '--git-dir='.$git_dir;
2686 # quote the given arguments for passing them to the shell
2687 # quote_command("command", "arg 1", "arg with ' and ! characters")
2688 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2689 # Try to avoid using this function wherever possible.
2690 sub quote_command {
2691 return join(' ',
2692 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2695 # get HEAD ref of given project as hash
2696 sub git_get_head_hash {
2697 return git_get_full_hash(shift, 'HEAD');
2700 sub git_get_full_hash {
2701 return git_get_hash(@_);
2704 sub git_get_short_hash {
2705 return git_get_hash(@_, '--short=7');
2708 sub git_get_hash {
2709 my ($project, $hash, @options) = @_;
2710 my $o_git_dir = $git_dir;
2711 my $retval = undef;
2712 $git_dir = "$projectroot/$project";
2713 if (open my $fd, '-|', git_cmd(), 'rev-parse',
2714 '--verify', '-q', @options, $hash) {
2715 $retval = <$fd>;
2716 chomp $retval if defined $retval;
2717 close $fd;
2719 if (defined $o_git_dir) {
2720 $git_dir = $o_git_dir;
2722 return $retval;
2725 # get type of given object
2726 sub git_get_type {
2727 my $hash = shift;
2729 open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2730 my $type = <$fd>;
2731 close $fd or return;
2732 chomp $type;
2733 return $type;
2736 # repository configuration
2737 our $config_file = '';
2738 our %config;
2740 # store multiple values for single key as anonymous array reference
2741 # single values stored directly in the hash, not as [ <value> ]
2742 sub hash_set_multi {
2743 my ($hash, $key, $value) = @_;
2745 if (!exists $hash->{$key}) {
2746 $hash->{$key} = $value;
2747 } elsif (!ref $hash->{$key}) {
2748 $hash->{$key} = [ $hash->{$key}, $value ];
2749 } else {
2750 push @{$hash->{$key}}, $value;
2754 # return hash of git project configuration
2755 # optionally limited to some section, e.g. 'gitweb'
2756 sub git_parse_project_config {
2757 my $section_regexp = shift;
2758 my %config;
2760 local $/ = "\0";
2762 open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2763 or return;
2765 while (my $keyval = <$fh>) {
2766 chomp $keyval;
2767 my ($key, $value) = split(/\n/, $keyval, 2);
2769 hash_set_multi(\%config, $key, $value)
2770 if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2772 close $fh;
2774 return %config;
2777 # convert config value to boolean: 'true' or 'false'
2778 # no value, number > 0, 'true' and 'yes' values are true
2779 # rest of values are treated as false (never as error)
2780 sub config_to_bool {
2781 my $val = shift;
2783 return 1 if !defined $val; # section.key
2785 # strip leading and trailing whitespace
2786 $val =~ s/^\s+//;
2787 $val =~ s/\s+$//;
2789 return (($val =~ /^\d+$/ && $val) || # section.key = 1
2790 ($val =~ /^(?:true|yes)$/i)); # section.key = true
2793 # convert config value to simple decimal number
2794 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2795 # to be multiplied by 1024, 1048576, or 1073741824
2796 sub config_to_int {
2797 my $val = shift;
2799 # strip leading and trailing whitespace
2800 $val =~ s/^\s+//;
2801 $val =~ s/\s+$//;
2803 if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2804 $unit = lc($unit);
2805 # unknown unit is treated as 1
2806 return $num * ($unit eq 'g' ? 1073741824 :
2807 $unit eq 'm' ? 1048576 :
2808 $unit eq 'k' ? 1024 : 1);
2810 return $val;
2813 # convert config value to array reference, if needed
2814 sub config_to_multi {
2815 my $val = shift;
2817 return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2820 sub git_get_project_config {
2821 my ($key, $type) = @_;
2823 return unless defined $git_dir;
2825 # key sanity check
2826 return unless ($key);
2827 # only subsection, if exists, is case sensitive,
2828 # and not lowercased by 'git config -z -l'
2829 if (my ($hi, $mi, $lo) = ($key =~ /^([^.]*)\.(.*)\.([^.]*)$/)) {
2830 $lo =~ s/_//g;
2831 $key = join(".", lc($hi), $mi, lc($lo));
2832 return if ($lo =~ /\W/ || $hi =~ /\W/);
2833 } else {
2834 $key = lc($key);
2835 $key =~ s/_//g;
2836 return if ($key =~ /\W/);
2838 $key =~ s/^gitweb\.//;
2840 # type sanity check
2841 if (defined $type) {
2842 $type =~ s/^--//;
2843 $type = undef
2844 unless ($type eq 'bool' || $type eq 'int');
2847 # get config
2848 if (!defined $config_file ||
2849 $config_file ne "$git_dir/config") {
2850 %config = git_parse_project_config('gitweb');
2851 $config_file = "$git_dir/config";
2854 # check if config variable (key) exists
2855 return unless exists $config{"gitweb.$key"};
2857 # ensure given type
2858 if (!defined $type) {
2859 return $config{"gitweb.$key"};
2860 } elsif ($type eq 'bool') {
2861 # backward compatibility: 'git config --bool' returns true/false
2862 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2863 } elsif ($type eq 'int') {
2864 return config_to_int($config{"gitweb.$key"});
2866 return $config{"gitweb.$key"};
2869 # get hash of given path at given ref
2870 sub git_get_hash_by_path {
2871 my $base = shift;
2872 my $path = shift || return undef;
2873 my $type = shift;
2875 $path =~ s,/+$,,;
2877 open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2878 or die_error(500, "Open git-ls-tree failed");
2879 my $line = <$fd>;
2880 close $fd or return undef;
2882 if (!defined $line) {
2883 # there is no tree or hash given by $path at $base
2884 return undef;
2887 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
2888 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/;
2889 if (defined $type && $type ne $2) {
2890 # type doesn't match
2891 return undef;
2893 return $3;
2896 # get path of entry with given hash at given tree-ish (ref)
2897 # used to get 'from' filename for combined diff (merge commit) for renames
2898 sub git_get_path_by_hash {
2899 my $base = shift || return;
2900 my $hash = shift || return;
2902 local $/ = "\0";
2904 open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2905 or return undef;
2906 while (my $line = <$fd>) {
2907 chomp $line;
2909 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423 gitweb'
2910 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f gitweb/README'
2911 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2912 close $fd;
2913 return $1;
2916 close $fd;
2917 return undef;
2920 ## ......................................................................
2921 ## git utility functions, directly accessing git repository
2923 # get the value of config variable either from file named as the variable
2924 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2925 # configuration variable in the repository config file.
2926 sub git_get_file_or_project_config {
2927 my ($path, $name) = @_;
2929 $git_dir = "$projectroot/$path";
2930 open my $fd, '<', "$git_dir/$name"
2931 or return git_get_project_config($name);
2932 my $conf = <$fd>;
2933 close $fd;
2934 if (defined $conf) {
2935 chomp $conf;
2937 return $conf;
2940 sub git_get_project_description {
2941 my $path = shift;
2942 return git_get_file_or_project_config($path, 'description');
2945 sub git_get_project_category {
2946 my $path = shift;
2947 return git_get_file_or_project_config($path, 'category');
2951 # supported formats:
2952 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2953 # - if its contents is a number, use it as tag weight,
2954 # - otherwise add a tag with weight 1
2955 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2956 # the same value multiple times increases tag weight
2957 # * `gitweb.ctag' multi-valued repo config variable
2958 sub git_get_project_ctags {
2959 my $project = shift;
2960 my $ctags = {};
2962 $git_dir = "$projectroot/$project";
2963 if (opendir my $dh, "$git_dir/ctags") {
2964 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2965 foreach my $tagfile (@files) {
2966 open my $ct, '<', $tagfile
2967 or next;
2968 my $val = <$ct>;
2969 chomp $val if $val;
2970 close $ct;
2972 (my $ctag = $tagfile) =~ s#.*/##;
2973 if ($val =~ /^\d+$/) {
2974 $ctags->{$ctag} = $val;
2975 } else {
2976 $ctags->{$ctag} = 1;
2979 closedir $dh;
2981 } elsif (open my $fh, '<', "$git_dir/ctags") {
2982 while (my $line = <$fh>) {
2983 chomp $line;
2984 $ctags->{$line}++ if $line;
2986 close $fh;
2988 } else {
2989 my $taglist = config_to_multi(git_get_project_config('ctag'));
2990 foreach my $tag (@$taglist) {
2991 $ctags->{$tag}++;
2995 return $ctags;
2998 # return hash, where keys are content tags ('ctags'),
2999 # and values are sum of weights of given tag in every project
3000 sub git_gather_all_ctags {
3001 my $projects = shift;
3002 my $ctags = {};
3004 foreach my $p (@$projects) {
3005 foreach my $ct (keys %{$p->{'ctags'}}) {
3006 $ctags->{$ct} += $p->{'ctags'}->{$ct};
3010 return $ctags;
3013 sub git_populate_project_tagcloud {
3014 my $ctags = shift;
3016 # First, merge different-cased tags; tags vote on casing
3017 my %ctags_lc;
3018 foreach (keys %$ctags) {
3019 $ctags_lc{lc $_}->{count} += $ctags->{$_};
3020 if (not $ctags_lc{lc $_}->{topcount}
3021 or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
3022 $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
3023 $ctags_lc{lc $_}->{topname} = $_;
3027 my $cloud;
3028 my $matched = $input_params{'ctag'};
3029 if (eval { require HTML::TagCloud; 1; }) {
3030 $cloud = HTML::TagCloud->new;
3031 foreach my $ctag (sort keys %ctags_lc) {
3032 # Pad the title with spaces so that the cloud looks
3033 # less crammed.
3034 my $title = esc_html($ctags_lc{$ctag}->{topname});
3035 $title =~ s/ /&nbsp;/g;
3036 $title =~ s/^/&nbsp;/g;
3037 $title =~ s/$/&nbsp;/g;
3038 if (defined $matched && $matched eq $ctag) {
3039 $title = qq(<span class="match">$title</span>);
3041 $cloud->add($title, href(project=>undef, ctag=>$ctag),
3042 $ctags_lc{$ctag}->{count});
3044 } else {
3045 $cloud = {};
3046 foreach my $ctag (keys %ctags_lc) {
3047 my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
3048 if (defined $matched && $matched eq $ctag) {
3049 $title = qq(<span class="match">$title</span>);
3051 $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
3052 $cloud->{$ctag}{ctag} =
3053 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
3056 return $cloud;
3059 sub git_show_project_tagcloud {
3060 my ($cloud, $count) = @_;
3061 if (ref $cloud eq 'HTML::TagCloud') {
3062 return $cloud->html_and_css($count);
3063 } else {
3064 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
3065 return
3066 '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
3067 join (', ', map {
3068 $cloud->{$_}->{'ctag'}
3069 } splice(@tags, 0, $count)) .
3070 '</div>';
3074 sub git_get_project_url_list {
3075 my $path = shift;
3077 $git_dir = "$projectroot/$path";
3078 open my $fd, '<', "$git_dir/cloneurl"
3079 or return wantarray ?
3080 @{ config_to_multi(git_get_project_config('url')) } :
3081 config_to_multi(git_get_project_config('url'));
3082 my @git_project_url_list = map { chomp; $_ } <$fd>;
3083 close $fd;
3085 return wantarray ? @git_project_url_list : \@git_project_url_list;
3088 sub git_get_projects_list {
3089 my $filter = shift || '';
3090 my $paranoid = shift;
3091 my @list;
3093 if (-d $projects_list) {
3094 # search in directory
3095 my $dir = $projects_list;
3096 # remove the trailing "/"
3097 $dir =~ s!/+$!!;
3098 my $pfxlen = length("$dir");
3099 my $pfxdepth = ($dir =~ tr!/!!);
3100 # when filtering, search only given subdirectory
3101 if ($filter && !$paranoid) {
3102 $dir .= "/$filter";
3103 $dir =~ s!/+$!!;
3106 File::Find::find({
3107 follow_fast => 1, # follow symbolic links
3108 follow_skip => 2, # ignore duplicates
3109 dangling_symlinks => 0, # ignore dangling symlinks, silently
3110 wanted => sub {
3111 # global variables
3112 our $project_maxdepth;
3113 our $projectroot;
3114 # skip project-list toplevel, if we get it.
3115 return if (m!^[/.]$!);
3116 # only directories can be git repositories
3117 return unless (-d $_);
3118 # need search permission
3119 return unless (-x $_);
3120 # don't traverse too deep (Find is super slow on os x)
3121 # $project_maxdepth excludes depth of $projectroot
3122 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
3123 $File::Find::prune = 1;
3124 return;
3127 my $path = substr($File::Find::name, $pfxlen + 1);
3128 # paranoidly only filter here
3129 if ($paranoid && $filter && $path !~ m!^\Q$filter\E/!) {
3130 next;
3132 # we check related file in $projectroot
3133 if (check_export_ok("$projectroot/$path")) {
3134 push @list, { path => $path };
3135 $File::Find::prune = 1;
3138 }, "$dir");
3140 } elsif (-f $projects_list) {
3141 # read from file(url-encoded):
3142 # 'git%2Fgit.git Linus+Torvalds'
3143 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3144 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3145 open my $fd, '<', $projects_list or return;
3146 PROJECT:
3147 while (my $line = <$fd>) {
3148 chomp $line;
3149 my ($path, $owner) = split ' ', $line;
3150 $path = unescape($path);
3151 $owner = unescape($owner);
3152 if (!defined $path) {
3153 next;
3155 # if $filter is rpovided, check if $path begins with $filter
3156 if ($filter && $path !~ m!^\Q$filter\E/!) {
3157 next;
3159 if (check_export_ok("$projectroot/$path")) {
3160 my $pr = {
3161 path => $path
3163 if ($owner) {
3164 $pr->{'owner'} = to_utf8($owner);
3166 push @list, $pr;
3169 close $fd;
3171 return @list;
3174 # written with help of Tree::Trie module (Perl Artistic License, GPL compatible)
3175 # as side effects it sets 'forks' field to list of forks for forked projects
3176 sub filter_forks_from_projects_list {
3177 my $projects = shift;
3179 my %trie; # prefix tree of directories (path components)
3180 # generate trie out of those directories that might contain forks
3181 foreach my $pr (@$projects) {
3182 my $path = $pr->{'path'};
3183 $path =~ s/\.git$//; # forks of 'repo.git' are in 'repo/' directory
3184 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
3185 next unless ($path); # skip '.git' repository: tests, git-instaweb
3186 next unless (-d "$projectroot/$path"); # containing directory exists
3187 $pr->{'forks'} = []; # there can be 0 or more forks of project
3189 # add to trie
3190 my @dirs = split('/', $path);
3191 # walk the trie, until either runs out of components or out of trie
3192 my $ref = \%trie;
3193 while (scalar @dirs &&
3194 exists($ref->{$dirs[0]})) {
3195 $ref = $ref->{shift @dirs};
3197 # create rest of trie structure from rest of components
3198 foreach my $dir (@dirs) {
3199 $ref = $ref->{$dir} = {};
3201 # create end marker, store $pr as a data
3202 $ref->{''} = $pr if (!exists $ref->{''});
3205 # filter out forks, by finding shortest prefix match for paths
3206 my @filtered;
3207 PROJECT:
3208 foreach my $pr (@$projects) {
3209 # trie lookup
3210 my $ref = \%trie;
3211 DIR:
3212 foreach my $dir (split('/', $pr->{'path'})) {
3213 if (exists $ref->{''}) {
3214 # found [shortest] prefix, is a fork - skip it
3215 push @{$ref->{''}{'forks'}}, $pr;
3216 next PROJECT;
3218 if (!exists $ref->{$dir}) {
3219 # not in trie, cannot have prefix, not a fork
3220 push @filtered, $pr;
3221 next PROJECT;
3223 # If the dir is there, we just walk one step down the trie.
3224 $ref = $ref->{$dir};
3226 # we ran out of trie
3227 # (shouldn't happen: it's either no match, or end marker)
3228 push @filtered, $pr;
3231 return @filtered;
3234 # note: fill_project_list_info must be run first,
3235 # for 'descr_long' and 'ctags' to be filled
3236 sub search_projects_list {
3237 my ($projlist, %opts) = @_;
3238 my $tagfilter = $opts{'tagfilter'};
3239 my $search_re = $opts{'search_regexp'};
3241 return @$projlist
3242 unless ($tagfilter || $search_re);
3244 # searching projects require filling to be run before it;
3245 fill_project_list_info($projlist,
3246 $tagfilter ? 'ctags' : (),
3247 $search_re ? ('path', 'descr') : ());
3248 my @projects;
3249 PROJECT:
3250 foreach my $pr (@$projlist) {
3252 if ($tagfilter) {
3253 next unless ref($pr->{'ctags'}) eq 'HASH';
3254 next unless
3255 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
3258 if ($search_re) {
3259 next unless
3260 $pr->{'path'} =~ /$search_re/ ||
3261 $pr->{'descr_long'} =~ /$search_re/;
3264 push @projects, $pr;
3267 return @projects;
3270 our $gitweb_project_owner = undef;
3271 sub git_get_project_list_from_file {
3273 return if (defined $gitweb_project_owner);
3275 $gitweb_project_owner = {};
3276 # read from file (url-encoded):
3277 # 'git%2Fgit.git Linus+Torvalds'
3278 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3279 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3280 if (-f $projects_list) {
3281 open(my $fd, '<', $projects_list);
3282 while (my $line = <$fd>) {
3283 chomp $line;
3284 my ($pr, $ow) = split ' ', $line;
3285 $pr = unescape($pr);
3286 $ow = unescape($ow);
3287 $gitweb_project_owner->{$pr} = to_utf8($ow);
3289 close $fd;
3293 sub git_get_project_owner {
3294 my $project = shift;
3295 my $owner;
3297 return undef unless $project;
3298 $git_dir = "$projectroot/$project";
3300 if (!defined $gitweb_project_owner) {
3301 git_get_project_list_from_file();
3304 if (exists $gitweb_project_owner->{$project}) {
3305 $owner = $gitweb_project_owner->{$project};
3307 if (!defined $owner){
3308 $owner = git_get_project_config('owner');
3310 if (!defined $owner) {
3311 $owner = get_file_owner("$git_dir");
3314 return $owner;
3317 sub git_get_last_activity {
3318 my ($path) = @_;
3319 my $fd;
3321 $git_dir = "$projectroot/$path";
3322 open($fd, "-|", git_cmd(), 'for-each-ref',
3323 '--format=%(committer)',
3324 '--sort=-committerdate',
3325 '--count=1',
3326 map { "refs/$_" } get_branch_refs ()) or return;
3327 my $most_recent = <$fd>;
3328 close $fd or return;
3329 if (defined $most_recent &&
3330 $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
3331 my $timestamp = $1;
3332 my $age = time - $timestamp;
3333 return ($age, age_string($age));
3335 return (undef, undef);
3338 # Implementation note: when a single remote is wanted, we cannot use 'git
3339 # remote show -n' because that command always work (assuming it's a remote URL
3340 # if it's not defined), and we cannot use 'git remote show' because that would
3341 # try to make a network roundtrip. So the only way to find if that particular
3342 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3343 # and when we find what we want.
3344 sub git_get_remotes_list {
3345 my $wanted = shift;
3346 my %remotes = ();
3348 open my $fd, '-|' , git_cmd(), 'remote', '-v';
3349 return unless $fd;
3350 while (my $remote = <$fd>) {
3351 chomp $remote;
3352 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3353 next if $wanted and not $remote eq $wanted;
3354 my ($url, $key) = ($1, $2);
3356 $remotes{$remote} ||= { 'heads' => () };
3357 $remotes{$remote}{$key} = $url;
3359 close $fd or return;
3360 return wantarray ? %remotes : \%remotes;
3363 # Takes a hash of remotes as first parameter and fills it by adding the
3364 # available remote heads for each of the indicated remotes.
3365 sub fill_remote_heads {
3366 my $remotes = shift;
3367 my @heads = map { "remotes/$_" } keys %$remotes;
3368 my @remoteheads = git_get_heads_list(undef, @heads);
3369 foreach my $remote (keys %$remotes) {
3370 $remotes->{$remote}{'heads'} = [ grep {
3371 $_->{'name'} =~ s!^$remote/!!
3372 } @remoteheads ];
3376 sub git_get_references {
3377 my $type = shift || "";
3378 my %refs;
3379 # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3380 # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3381 open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3382 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3383 or return;
3385 while (my $line = <$fd>) {
3386 chomp $line;
3387 if ($line =~ m!^($oid_regex)\srefs/($type.*)$!) {
3388 if (defined $refs{$1}) {
3389 push @{$refs{$1}}, $2;
3390 } else {
3391 $refs{$1} = [ $2 ];
3395 close $fd or return;
3396 return \%refs;
3399 sub git_get_rev_name_tags {
3400 my $hash = shift || return undef;
3402 open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3403 or return;
3404 my $name_rev = <$fd>;
3405 close $fd;
3407 if ($name_rev =~ m|^$hash tags/(.*)$|) {
3408 return $1;
3409 } else {
3410 # catches also '$hash undefined' output
3411 return undef;
3415 ## ----------------------------------------------------------------------
3416 ## parse to hash functions
3418 sub parse_date {
3419 my $epoch = shift;
3420 my $tz = shift || "-0000";
3422 my %date;
3423 my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3424 my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3425 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3426 $date{'hour'} = $hour;
3427 $date{'minute'} = $min;
3428 $date{'mday'} = $mday;
3429 $date{'day'} = $days[$wday];
3430 $date{'month'} = $months[$mon];
3431 $date{'rfc2822'} = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3432 $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3433 $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3434 $mday, $months[$mon], $hour ,$min;
3435 $date{'iso-8601'} = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3436 1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3438 my ($tz_sign, $tz_hour, $tz_min) =
3439 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3440 $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3441 my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3442 ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3443 $date{'hour_local'} = $hour;
3444 $date{'minute_local'} = $min;
3445 $date{'tz_local'} = $tz;
3446 $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3447 1900+$year, $mon+1, $mday,
3448 $hour, $min, $sec, $tz);
3449 return %date;
3452 sub parse_tag {
3453 my $tag_id = shift;
3454 my %tag;
3455 my @comment;
3457 open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3458 $tag{'id'} = $tag_id;
3459 while (my $line = <$fd>) {
3460 chomp $line;
3461 if ($line =~ m/^object ($oid_regex)$/) {
3462 $tag{'object'} = $1;
3463 } elsif ($line =~ m/^type (.+)$/) {
3464 $tag{'type'} = $1;
3465 } elsif ($line =~ m/^tag (.+)$/) {
3466 $tag{'name'} = $1;
3467 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3468 $tag{'author'} = $1;
3469 $tag{'author_epoch'} = $2;
3470 $tag{'author_tz'} = $3;
3471 if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3472 $tag{'author_name'} = $1;
3473 $tag{'author_email'} = $2;
3474 } else {
3475 $tag{'author_name'} = $tag{'author'};
3477 } elsif ($line =~ m/--BEGIN/) {
3478 push @comment, $line;
3479 last;
3480 } elsif ($line eq "") {
3481 last;
3484 push @comment, <$fd>;
3485 $tag{'comment'} = \@comment;
3486 close $fd or return;
3487 if (!defined $tag{'name'}) {
3488 return
3490 return %tag
3493 sub parse_commit_text {
3494 my ($commit_text, $withparents) = @_;
3495 my @commit_lines = split '\n', $commit_text;
3496 my %co;
3498 pop @commit_lines; # Remove '\0'
3500 if (! @commit_lines) {
3501 return;
3504 my $header = shift @commit_lines;
3505 if ($header !~ m/^$oid_regex/) {
3506 return;
3508 ($co{'id'}, my @parents) = split ' ', $header;
3509 while (my $line = shift @commit_lines) {
3510 last if $line eq "\n";
3511 if ($line =~ m/^tree ($oid_regex)$/) {
3512 $co{'tree'} = $1;
3513 } elsif ((!defined $withparents) && ($line =~ m/^parent ($oid_regex)$/)) {
3514 push @parents, $1;
3515 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3516 $co{'author'} = to_utf8($1);
3517 $co{'author_epoch'} = $2;
3518 $co{'author_tz'} = $3;
3519 if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3520 $co{'author_name'} = $1;
3521 $co{'author_email'} = $2;
3522 } else {
3523 $co{'author_name'} = $co{'author'};
3525 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3526 $co{'committer'} = to_utf8($1);
3527 $co{'committer_epoch'} = $2;
3528 $co{'committer_tz'} = $3;
3529 if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3530 $co{'committer_name'} = $1;
3531 $co{'committer_email'} = $2;
3532 } else {
3533 $co{'committer_name'} = $co{'committer'};
3537 if (!defined $co{'tree'}) {
3538 return;
3540 $co{'parents'} = \@parents;
3541 $co{'parent'} = $parents[0];
3543 foreach my $title (@commit_lines) {
3544 $title =~ s/^ //;
3545 if ($title ne "") {
3546 $co{'title'} = chop_str($title, 80, 5);
3547 # remove leading stuff of merges to make the interesting part visible
3548 if (length($title) > 50) {
3549 $title =~ s/^Automatic //;
3550 $title =~ s/^merge (of|with) /Merge ... /i;
3551 if (length($title) > 50) {
3552 $title =~ s/(http|rsync):\/\///;
3554 if (length($title) > 50) {
3555 $title =~ s/(master|www|rsync)\.//;
3557 if (length($title) > 50) {
3558 $title =~ s/kernel.org:?//;
3560 if (length($title) > 50) {
3561 $title =~ s/\/pub\/scm//;
3564 $co{'title_short'} = chop_str($title, 50, 5);
3565 last;
3568 if (! defined $co{'title'} || $co{'title'} eq "") {
3569 $co{'title'} = $co{'title_short'} = '(no commit message)';
3571 # remove added spaces
3572 foreach my $line (@commit_lines) {
3573 $line =~ s/^ //;
3575 $co{'comment'} = \@commit_lines;
3577 my $age = time - $co{'committer_epoch'};
3578 $co{'age'} = $age;
3579 $co{'age_string'} = age_string($age);
3580 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3581 if ($age > 60*60*24*7*2) {
3582 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3583 $co{'age_string_age'} = $co{'age_string'};
3584 } else {
3585 $co{'age_string_date'} = $co{'age_string'};
3586 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3588 return %co;
3591 sub parse_commit {
3592 my ($commit_id) = @_;
3593 my %co;
3595 local $/ = "\0";
3597 open my $fd, "-|", git_cmd(), "rev-list",
3598 "--parents",
3599 "--header",
3600 "--max-count=1",
3601 $commit_id,
3602 "--",
3603 or die_error(500, "Open git-rev-list failed");
3604 %co = parse_commit_text(<$fd>, 1);
3605 close $fd;
3607 return %co;
3610 sub parse_commits {
3611 my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3612 my @cos;
3614 $maxcount ||= 1;
3615 $skip ||= 0;
3617 local $/ = "\0";
3619 open my $fd, "-|", git_cmd(), "rev-list",
3620 "--header",
3621 @args,
3622 ("--max-count=" . $maxcount),
3623 ("--skip=" . $skip),
3624 @extra_options,
3625 $commit_id,
3626 "--",
3627 ($filename ? ($filename) : ())
3628 or die_error(500, "Open git-rev-list failed");
3629 while (my $line = <$fd>) {
3630 my %co = parse_commit_text($line);
3631 push @cos, \%co;
3633 close $fd;
3635 return wantarray ? @cos : \@cos;
3638 # parse line of git-diff-tree "raw" output
3639 sub parse_difftree_raw_line {
3640 my $line = shift;
3641 my %res;
3643 # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M ls-files.c'
3644 # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M rev-tree.c'
3645 if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ($oid_regex) ($oid_regex) (.)([0-9]{0,3})\t(.*)$/) {
3646 $res{'from_mode'} = $1;
3647 $res{'to_mode'} = $2;
3648 $res{'from_id'} = $3;
3649 $res{'to_id'} = $4;
3650 $res{'status'} = $5;
3651 $res{'similarity'} = $6;
3652 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3653 ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3654 } else {
3655 $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3658 # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3659 # combined diff (for merge commit)
3660 elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:$oid_regex )+)([a-zA-Z]+)\t(.*)$//) {
3661 $res{'nparents'} = length($1);
3662 $res{'from_mode'} = [ split(' ', $2) ];
3663 $res{'to_mode'} = pop @{$res{'from_mode'}};
3664 $res{'from_id'} = [ split(' ', $3) ];
3665 $res{'to_id'} = pop @{$res{'from_id'}};
3666 $res{'status'} = [ split('', $4) ];
3667 $res{'to_file'} = unquote($5);
3669 # 'c512b523472485aef4fff9e57b229d9d243c967f'
3670 elsif ($line =~ m/^($oid_regex)$/) {
3671 $res{'commit'} = $1;
3674 return wantarray ? %res : \%res;
3677 # wrapper: return parsed line of git-diff-tree "raw" output
3678 # (the argument might be raw line, or parsed info)
3679 sub parsed_difftree_line {
3680 my $line_or_ref = shift;
3682 if (ref($line_or_ref) eq "HASH") {
3683 # pre-parsed (or generated by hand)
3684 return $line_or_ref;
3685 } else {
3686 return parse_difftree_raw_line($line_or_ref);
3690 # parse line of git-ls-tree output
3691 sub parse_ls_tree_line {
3692 my $line = shift;
3693 my %opts = @_;
3694 my %res;
3696 if ($opts{'-l'}) {
3697 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa 16717 panic.c'
3698 $line =~ m/^([0-9]+) (.+) ($oid_regex) +(-|[0-9]+)\t(.+)$/s;
3700 $res{'mode'} = $1;
3701 $res{'type'} = $2;
3702 $res{'hash'} = $3;
3703 $res{'size'} = $4;
3704 if ($opts{'-z'}) {
3705 $res{'name'} = $5;
3706 } else {
3707 $res{'name'} = unquote($5);
3709 } else {
3710 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
3711 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t(.+)$/s;
3713 $res{'mode'} = $1;
3714 $res{'type'} = $2;
3715 $res{'hash'} = $3;
3716 if ($opts{'-z'}) {
3717 $res{'name'} = $4;
3718 } else {
3719 $res{'name'} = unquote($4);
3723 return wantarray ? %res : \%res;
3726 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3727 sub parse_from_to_diffinfo {
3728 my ($diffinfo, $from, $to, @parents) = @_;
3730 if ($diffinfo->{'nparents'}) {
3731 # combined diff
3732 $from->{'file'} = [];
3733 $from->{'href'} = [];
3734 fill_from_file_info($diffinfo, @parents)
3735 unless exists $diffinfo->{'from_file'};
3736 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3737 $from->{'file'}[$i] =
3738 defined $diffinfo->{'from_file'}[$i] ?
3739 $diffinfo->{'from_file'}[$i] :
3740 $diffinfo->{'to_file'};
3741 if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3742 $from->{'href'}[$i] = href(action=>"blob",
3743 hash_base=>$parents[$i],
3744 hash=>$diffinfo->{'from_id'}[$i],
3745 file_name=>$from->{'file'}[$i]);
3746 } else {
3747 $from->{'href'}[$i] = undef;
3750 } else {
3751 # ordinary (not combined) diff
3752 $from->{'file'} = $diffinfo->{'from_file'};
3753 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3754 $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3755 hash=>$diffinfo->{'from_id'},
3756 file_name=>$from->{'file'});
3757 } else {
3758 delete $from->{'href'};
3762 $to->{'file'} = $diffinfo->{'to_file'};
3763 if (!is_deleted($diffinfo)) { # file exists in result
3764 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3765 hash=>$diffinfo->{'to_id'},
3766 file_name=>$to->{'file'});
3767 } else {
3768 delete $to->{'href'};
3772 ## ......................................................................
3773 ## parse to array of hashes functions
3775 sub git_get_heads_list {
3776 my ($limit, @classes) = @_;
3777 @classes = get_branch_refs() unless @classes;
3778 my @patterns = map { "refs/$_" } @classes;
3779 my @headslist;
3781 open my $fd, '-|', git_cmd(), 'for-each-ref',
3782 ($limit ? '--count='.($limit+1) : ()), '--sort=-committerdate',
3783 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3784 @patterns
3785 or return;
3786 while (my $line = <$fd>) {
3787 my %ref_item;
3789 chomp $line;
3790 my ($refinfo, $committerinfo) = split(/\0/, $line);
3791 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3792 my ($committer, $epoch, $tz) =
3793 ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3794 $ref_item{'fullname'} = $name;
3795 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
3796 $name =~ s!^refs/($strip_refs|remotes)/!!;
3797 $ref_item{'name'} = $name;
3798 # for refs neither in 'heads' nor 'remotes' we want to
3799 # show their ref dir
3800 my $ref_dir = (defined $1) ? $1 : '';
3801 if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
3802 $ref_item{'name'} .= ' (' . $ref_dir . ')';
3805 $ref_item{'id'} = $hash;
3806 $ref_item{'title'} = $title || '(no commit message)';
3807 $ref_item{'epoch'} = $epoch;
3808 if ($epoch) {
3809 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3810 } else {
3811 $ref_item{'age'} = "unknown";
3814 push @headslist, \%ref_item;
3816 close $fd;
3818 return wantarray ? @headslist : \@headslist;
3821 sub git_get_tags_list {
3822 my $limit = shift;
3823 my @tagslist;
3825 open my $fd, '-|', git_cmd(), 'for-each-ref',
3826 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3827 '--format=%(objectname) %(objecttype) %(refname) '.
3828 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3829 'refs/tags'
3830 or return;
3831 while (my $line = <$fd>) {
3832 my %ref_item;
3834 chomp $line;
3835 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3836 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3837 my ($creator, $epoch, $tz) =
3838 ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3839 $ref_item{'fullname'} = $name;
3840 $name =~ s!^refs/tags/!!;
3842 $ref_item{'type'} = $type;
3843 $ref_item{'id'} = $id;
3844 $ref_item{'name'} = $name;
3845 if ($type eq "tag") {
3846 $ref_item{'subject'} = $title;
3847 $ref_item{'reftype'} = $reftype;
3848 $ref_item{'refid'} = $refid;
3849 } else {
3850 $ref_item{'reftype'} = $type;
3851 $ref_item{'refid'} = $id;
3854 if ($type eq "tag" || $type eq "commit") {
3855 $ref_item{'epoch'} = $epoch;
3856 if ($epoch) {
3857 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3858 } else {
3859 $ref_item{'age'} = "unknown";
3863 push @tagslist, \%ref_item;
3865 close $fd;
3867 return wantarray ? @tagslist : \@tagslist;
3870 ## ----------------------------------------------------------------------
3871 ## filesystem-related functions
3873 sub get_file_owner {
3874 my $path = shift;
3876 my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3877 my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3878 if (!defined $gcos) {
3879 return undef;
3881 my $owner = $gcos;
3882 $owner =~ s/[,;].*$//;
3883 return to_utf8($owner);
3886 # assume that file exists
3887 sub insert_file {
3888 my $filename = shift;
3890 open my $fd, '<', $filename;
3891 print map { to_utf8($_) } <$fd>;
3892 close $fd;
3895 ## ......................................................................
3896 ## mimetype related functions
3898 sub mimetype_guess_file {
3899 my $filename = shift;
3900 my $mimemap = shift;
3901 -r $mimemap or return undef;
3903 my %mimemap;
3904 open(my $mh, '<', $mimemap) or return undef;
3905 while (<$mh>) {
3906 next if m/^#/; # skip comments
3907 my ($mimetype, @exts) = split(/\s+/);
3908 foreach my $ext (@exts) {
3909 $mimemap{$ext} = $mimetype;
3912 close($mh);
3914 $filename =~ /\.([^.]*)$/;
3915 return $mimemap{$1};
3918 sub mimetype_guess {
3919 my $filename = shift;
3920 my $mime;
3921 $filename =~ /\./ or return undef;
3923 if ($mimetypes_file) {
3924 my $file = $mimetypes_file;
3925 if ($file !~ m!^/!) { # if it is relative path
3926 # it is relative to project
3927 $file = "$projectroot/$project/$file";
3929 $mime = mimetype_guess_file($filename, $file);
3931 $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3932 return $mime;
3935 sub blob_mimetype {
3936 my $fd = shift;
3937 my $filename = shift;
3939 if ($filename) {
3940 my $mime = mimetype_guess($filename);
3941 $mime and return $mime;
3944 # just in case
3945 return $default_blob_plain_mimetype unless $fd;
3947 if (-T $fd) {
3948 return 'text/plain';
3949 } elsif (! $filename) {
3950 return 'application/octet-stream';
3951 } elsif ($filename =~ m/\.png$/i) {
3952 return 'image/png';
3953 } elsif ($filename =~ m/\.gif$/i) {
3954 return 'image/gif';
3955 } elsif ($filename =~ m/\.jpe?g$/i) {
3956 return 'image/jpeg';
3957 } else {
3958 return 'application/octet-stream';
3962 sub blob_contenttype {
3963 my ($fd, $file_name, $type) = @_;
3965 $type ||= blob_mimetype($fd, $file_name);
3966 if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3967 $type .= "; charset=$default_text_plain_charset";
3970 return $type;
3973 # guess file syntax for syntax highlighting; return undef if no highlighting
3974 # the name of syntax can (in the future) depend on syntax highlighter used
3975 sub guess_file_syntax {
3976 my ($highlight, $file_name) = @_;
3977 return undef unless ($highlight && defined $file_name);
3978 my $basename = basename($file_name, '.in');
3979 return $highlight_basename{$basename}
3980 if exists $highlight_basename{$basename};
3982 $basename =~ /\.([^.]*)$/;
3983 my $ext = $1 or return undef;
3984 return $highlight_ext{$ext}
3985 if exists $highlight_ext{$ext};
3987 return undef;
3990 # run highlighter and return FD of its output,
3991 # or return original FD if no highlighting
3992 sub run_highlighter {
3993 my ($fd, $highlight, $syntax) = @_;
3994 return $fd unless ($highlight);
3996 close $fd;
3997 my $syntax_arg = (defined $syntax) ? "--syntax $syntax" : "--force";
3998 open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
3999 quote_command($^X, '-CO', '-MEncode=decode,FB_DEFAULT', '-pse',
4000 '$_ = decode($fe, $_, FB_DEFAULT) if !utf8::decode($_);',
4001 '--', "-fe=$fallback_encoding")." | ".
4002 quote_command($highlight_bin).
4003 " --replace-tabs=8 --fragment $syntax_arg |"
4004 or die_error(500, "Couldn't open file or run syntax highlighter");
4005 return $fd;
4008 ## ======================================================================
4009 ## functions printing HTML: header, footer, error page
4011 sub get_page_title {
4012 my $title = to_utf8($site_name);
4014 unless (defined $project) {
4015 if (defined $project_filter) {
4016 $title .= " - projects in '" . esc_path($project_filter) . "'";
4018 return $title;
4020 $title .= " - " . to_utf8($project);
4022 return $title unless (defined $action);
4023 $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
4025 return $title unless (defined $file_name);
4026 $title .= " - " . esc_path($file_name);
4027 if ($action eq "tree" && $file_name !~ m|/$|) {
4028 $title .= "/";
4031 return $title;
4034 sub get_content_type_html {
4035 # require explicit support from the UA if we are to send the page as
4036 # 'application/xhtml+xml', otherwise send it as plain old 'text/html'.
4037 # we have to do this because MSIE sometimes globs '*/*', pretending to
4038 # support xhtml+xml but choking when it gets what it asked for.
4039 if (defined $cgi->http('HTTP_ACCEPT') &&
4040 $cgi->http('HTTP_ACCEPT') =~ m/(,|;|\s|^)application\/xhtml\+xml(,|;|\s|$)/ &&
4041 $cgi->Accept('application/xhtml+xml') != 0) {
4042 return 'application/xhtml+xml';
4043 } else {
4044 return 'text/html';
4048 sub print_feed_meta {
4049 if (defined $project) {
4050 my %href_params = get_feed_info();
4051 if (!exists $href_params{'-title'}) {
4052 $href_params{'-title'} = 'log';
4055 foreach my $format (qw(RSS Atom)) {
4056 my $type = lc($format);
4057 my %link_attr = (
4058 '-rel' => 'alternate',
4059 '-title' => esc_attr("$project - $href_params{'-title'} - $format feed"),
4060 '-type' => "application/$type+xml"
4063 $href_params{'extra_options'} = undef;
4064 $href_params{'action'} = $type;
4065 $link_attr{'-href'} = esc_attr(href(%href_params));
4066 print "<link ".
4067 "rel=\"$link_attr{'-rel'}\" ".
4068 "title=\"$link_attr{'-title'}\" ".
4069 "href=\"$link_attr{'-href'}\" ".
4070 "type=\"$link_attr{'-type'}\" ".
4071 "/>\n";
4073 $href_params{'extra_options'} = '--no-merges';
4074 $link_attr{'-href'} = esc_attr(href(%href_params));
4075 $link_attr{'-title'} .= ' (no merges)';
4076 print "<link ".
4077 "rel=\"$link_attr{'-rel'}\" ".
4078 "title=\"$link_attr{'-title'}\" ".
4079 "href=\"$link_attr{'-href'}\" ".
4080 "type=\"$link_attr{'-type'}\" ".
4081 "/>\n";
4084 } else {
4085 printf('<link rel="alternate" title="%s projects list" '.
4086 'href="%s" type="text/plain; charset=utf-8" />'."\n",
4087 esc_attr($site_name),
4088 esc_attr(href(project=>undef, action=>"project_index")));
4089 printf('<link rel="alternate" title="%s projects feeds" '.
4090 'href="%s" type="text/x-opml" />'."\n",
4091 esc_attr($site_name),
4092 esc_attr(href(project=>undef, action=>"opml")));
4096 sub print_header_links {
4097 my $status = shift;
4099 # print out each stylesheet that exist, providing backwards capability
4100 # for those people who defined $stylesheet in a config file
4101 if (defined $stylesheet) {
4102 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4103 } else {
4104 foreach my $stylesheet (@stylesheets) {
4105 next unless $stylesheet;
4106 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4109 print_feed_meta()
4110 if ($status eq '200 OK');
4111 if (defined $favicon) {
4112 print qq(<link rel="shortcut icon" href=").esc_url($favicon).qq(" type="image/png" />\n);
4116 sub print_nav_breadcrumbs_path {
4117 my $dirprefix = undef;
4118 while (my $part = shift) {
4119 $dirprefix .= "/" if defined $dirprefix;
4120 $dirprefix .= $part;
4121 print $cgi->a({-href => href(project => undef,
4122 project_filter => $dirprefix,
4123 action => "project_list")},
4124 esc_html($part)) . " / ";
4128 sub print_nav_breadcrumbs {
4129 my %opts = @_;
4131 for my $crumb (@extra_breadcrumbs, [ $home_link_str => $home_link ]) {
4132 print $cgi->a({-href => esc_url($crumb->[1])}, $crumb->[0]) . " / ";
4134 if (defined $project) {
4135 my @dirname = split '/', $project;
4136 my $projectbasename = pop @dirname;
4137 print_nav_breadcrumbs_path(@dirname);
4138 print $cgi->a({-href => href(action=>"summary")}, esc_html($projectbasename));
4139 if (defined $action) {
4140 my $action_print = $action ;
4141 if (defined $opts{-action_extra}) {
4142 $action_print = $cgi->a({-href => href(action=>$action)},
4143 $action);
4145 print " / $action_print";
4147 if (defined $opts{-action_extra}) {
4148 print " / $opts{-action_extra}";
4150 print "\n";
4151 } elsif (defined $project_filter) {
4152 print_nav_breadcrumbs_path(split '/', $project_filter);
4156 sub print_search_form {
4157 if (!defined $searchtext) {
4158 $searchtext = "";
4160 my $search_hash;
4161 if (defined $hash_base) {
4162 $search_hash = $hash_base;
4163 } elsif (defined $hash) {
4164 $search_hash = $hash;
4165 } else {
4166 $search_hash = "HEAD";
4168 my $action = $my_uri;
4169 my $use_pathinfo = gitweb_check_feature('pathinfo');
4170 if ($use_pathinfo) {
4171 $action .= "/".esc_url($project);
4173 print $cgi->start_form(-method => "get", -action => $action) .
4174 "<div class=\"search\">\n" .
4175 (!$use_pathinfo &&
4176 $cgi->input({-name=>"p", -value=>$project, -type=>"hidden"}) . "\n") .
4177 $cgi->input({-name=>"a", -value=>"search", -type=>"hidden"}) . "\n" .
4178 $cgi->input({-name=>"h", -value=>$search_hash, -type=>"hidden"}) . "\n" .
4179 $cgi->popup_menu(-name => 'st', -default => 'commit',
4180 -values => ['commit', 'grep', 'author', 'committer', 'pickaxe']) .
4181 " " . $cgi->a({-href => href(action=>"search_help"),
4182 -title => "se