Update draft release notes to 2.2
[git.git] / gitweb / gitweb.perl
1 #!/usr/bin/perl
2
3 # gitweb - simple web interface to track changes in git repositories
4 #
5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
6 # (C) 2005, Christian Gierke
7 #
8 # This program is licensed under the GPLv2
9
10 use 5.008;
11 use strict;
12 use warnings;
13 use CGI qw(:standard :escapeHTML -nosticky);
14 use CGI::Util qw(unescape);
15 use CGI::Carp qw(fatalsToBrowser set_message);
16 use Encode;
17 use Fcntl ':mode';
18 use File::Find qw();
19 use File::Basename qw(basename);
20 use Time::HiRes qw(gettimeofday tv_interval);
21 binmode STDOUT, ':utf8';
22
23 our $t0 = [ gettimeofday() ];
24 our $number_of_git_cmds = 0;
25
26 BEGIN {
27 CGI->compile() if $ENV{'MOD_PERL'};
28 }
29
30 our $version = "++GIT_VERSION++";
31
32 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
33 sub evaluate_uri {
34 our $cgi;
35
36 our $my_url = $cgi->url();
37 our $my_uri = $cgi->url(-absolute => 1);
38
39 # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
40 # needed and used only for URLs with nonempty PATH_INFO
41 our $base_url = $my_url;
42
43 # When the script is used as DirectoryIndex, the URL does not contain the name
44 # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
45 # have to do it ourselves. We make $path_info global because it's also used
46 # later on.
47 #
48 # Another issue with the script being the DirectoryIndex is that the resulting
49 # $my_url data is not the full script URL: this is good, because we want
50 # generated links to keep implying the script name if it wasn't explicitly
51 # indicated in the URL we're handling, but it means that $my_url cannot be used
52 # as base URL.
53 # Therefore, if we needed to strip PATH_INFO, then we know that we have
54 # to build the base URL ourselves:
55 our $path_info = decode_utf8($ENV{"PATH_INFO"});
56 if ($path_info) {
57 # $path_info has already been URL-decoded by the web server, but
58 # $my_url and $my_uri have not. URL-decode them so we can properly
59 # strip $path_info.
60 $my_url = unescape($my_url);
61 $my_uri = unescape($my_uri);
62 if ($my_url =~ s,\Q$path_info\E$,, &&
63 $my_uri =~ s,\Q$path_info\E$,, &&
64 defined $ENV{'SCRIPT_NAME'}) {
65 $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
66 }
67 }
68
69 # target of the home link on top of all pages
70 our $home_link = $my_uri || "/";
71 }
72
73 # core git executable to use
74 # this can just be "git" if your webserver has a sensible PATH
75 our $GIT = "++GIT_BINDIR++/git";
76
77 # absolute fs-path which will be prepended to the project path
78 #our $projectroot = "/pub/scm";
79 our $projectroot = "++GITWEB_PROJECTROOT++";
80
81 # fs traversing limit for getting project list
82 # the number is relative to the projectroot
83 our $project_maxdepth = "++GITWEB_PROJECT_MAXDEPTH++";
84
85 # string of the home link on top of all pages
86 our $home_link_str = "++GITWEB_HOME_LINK_STR++";
87
88 # extra breadcrumbs preceding the home link
89 our @extra_breadcrumbs = ();
90
91 # name of your site or organization to appear in page titles
92 # replace this with something more descriptive for clearer bookmarks
93 our $site_name = "++GITWEB_SITENAME++"
94 || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
95
96 # html snippet to include in the <head> section of each page
97 our $site_html_head_string = "++GITWEB_SITE_HTML_HEAD_STRING++";
98 # filename of html text to include at top of each page
99 our $site_header = "++GITWEB_SITE_HEADER++";
100 # html text to include at home page
101 our $home_text = "++GITWEB_HOMETEXT++";
102 # filename of html text to include at bottom of each page
103 our $site_footer = "++GITWEB_SITE_FOOTER++";
104
105 # URI of stylesheets
106 our @stylesheets = ("++GITWEB_CSS++");
107 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
108 our $stylesheet = undef;
109 # URI of GIT logo (72x27 size)
110 our $logo = "++GITWEB_LOGO++";
111 # URI of GIT favicon, assumed to be image/png type
112 our $favicon = "++GITWEB_FAVICON++";
113 # URI of gitweb.js (JavaScript code for gitweb)
114 our $javascript = "++GITWEB_JS++";
115
116 # URI and label (title) of GIT logo link
117 #our $logo_url = "http://www.kernel.org/pub/software/scm/git/docs/";
118 #our $logo_label = "git documentation";
119 our $logo_url = "http://git-scm.com/";
120 our $logo_label = "git homepage";
121
122 # source of projects list
123 our $projects_list = "++GITWEB_LIST++";
124
125 # the width (in characters) of the projects list "Description" column
126 our $projects_list_description_width = 25;
127
128 # group projects by category on the projects list
129 # (enabled if this variable evaluates to true)
130 our $projects_list_group_categories = 0;
131
132 # default category if none specified
133 # (leave the empty string for no category)
134 our $project_list_default_category = "";
135
136 # default order of projects list
137 # valid values are none, project, descr, owner, and age
138 our $default_projects_order = "project";
139
140 # show repository only if this file exists
141 # (only effective if this variable evaluates to true)
142 our $export_ok = "++GITWEB_EXPORT_OK++";
143
144 # don't generate age column on the projects list page
145 our $omit_age_column = 0;
146
147 # don't generate information about owners of repositories
148 our $omit_owner=0;
149
150 # show repository only if this subroutine returns true
151 # when given the path to the project, for example:
152 # sub { return -e "$_[0]/git-daemon-export-ok"; }
153 our $export_auth_hook = undef;
154
155 # only allow viewing of repositories also shown on the overview page
156 our $strict_export = "++GITWEB_STRICT_EXPORT++";
157
158 # list of git base URLs used for URL to where fetch project from,
159 # i.e. full URL is "$git_base_url/$project"
160 our @git_base_url_list = grep { $_ ne '' } ("++GITWEB_BASE_URL++");
161
162 # default blob_plain mimetype and default charset for text/plain blob
163 our $default_blob_plain_mimetype = 'text/plain';
164 our $default_text_plain_charset = undef;
165
166 # file to use for guessing MIME types before trying /etc/mime.types
167 # (relative to the current git repository)
168 our $mimetypes_file = undef;
169
170 # assume this charset if line contains non-UTF-8 characters;
171 # it should be valid encoding (see Encoding::Supported(3pm) for list),
172 # for which encoding all byte sequences are valid, for example
173 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
174 # could be even 'utf-8' for the old behavior)
175 our $fallback_encoding = 'latin1';
176
177 # rename detection options for git-diff and git-diff-tree
178 # - default is '-M', with the cost proportional to
179 # (number of removed files) * (number of new files).
180 # - more costly is '-C' (which implies '-M'), with the cost proportional to
181 # (number of changed files + number of removed files) * (number of new files)
182 # - even more costly is '-C', '--find-copies-harder' with cost
183 # (number of files in the original tree) * (number of new files)
184 # - one might want to include '-B' option, e.g. '-B', '-M'
185 our @diff_opts = ('-M'); # taken from git_commit
186
187 # Disables features that would allow repository owners to inject script into
188 # the gitweb domain.
189 our $prevent_xss = 0;
190
191 # Path to the highlight executable to use (must be the one from
192 # http://www.andre-simon.de due to assumptions about parameters and output).
193 # Useful if highlight is not installed on your webserver's PATH.
194 # [Default: highlight]
195 our $highlight_bin = "++HIGHLIGHT_BIN++";
196
197 # information about snapshot formats that gitweb is capable of serving
198 our %known_snapshot_formats = (
199 # name => {
200 # 'display' => display name,
201 # 'type' => mime type,
202 # 'suffix' => filename suffix,
203 # 'format' => --format for git-archive,
204 # 'compressor' => [compressor command and arguments]
205 # (array reference, optional)
206 # 'disabled' => boolean (optional)}
207 #
208 'tgz' => {
209 'display' => 'tar.gz',
210 'type' => 'application/x-gzip',
211 'suffix' => '.tar.gz',
212 'format' => 'tar',
213 'compressor' => ['gzip', '-n']},
214
215 'tbz2' => {
216 'display' => 'tar.bz2',
217 'type' => 'application/x-bzip2',
218 'suffix' => '.tar.bz2',
219 'format' => 'tar',
220 'compressor' => ['bzip2']},
221
222 'txz' => {
223 'display' => 'tar.xz',
224 'type' => 'application/x-xz',
225 'suffix' => '.tar.xz',
226 'format' => 'tar',
227 'compressor' => ['xz'],
228 'disabled' => 1},
229
230 'zip' => {
231 'display' => 'zip',
232 'type' => 'application/x-zip',
233 'suffix' => '.zip',
234 'format' => 'zip'},
235 );
236
237 # Aliases so we understand old gitweb.snapshot values in repository
238 # configuration.
239 our %known_snapshot_format_aliases = (
240 'gzip' => 'tgz',
241 'bzip2' => 'tbz2',
242 'xz' => 'txz',
243
244 # backward compatibility: legacy gitweb config support
245 'x-gzip' => undef, 'gz' => undef,
246 'x-bzip2' => undef, 'bz2' => undef,
247 'x-zip' => undef, '' => undef,
248 );
249
250 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
251 # are changed, it may be appropriate to change these values too via
252 # $GITWEB_CONFIG.
253 our %avatar_size = (
254 'default' => 16,
255 'double' => 32
256 );
257
258 # Used to set the maximum load that we will still respond to gitweb queries.
259 # If server load exceed this value then return "503 server busy" error.
260 # If gitweb cannot determined server load, it is taken to be 0.
261 # Leave it undefined (or set to 'undef') to turn off load checking.
262 our $maxload = 300;
263
264 # configuration for 'highlight' (http://www.andre-simon.de/)
265 # match by basename
266 our %highlight_basename = (
267 #'Program' => 'py',
268 #'Library' => 'py',
269 'SConstruct' => 'py', # SCons equivalent of Makefile
270 'Makefile' => 'make',
271 );
272 # match by extension
273 our %highlight_ext = (
274 # main extensions, defining name of syntax;
275 # see files in /usr/share/highlight/langDefs/ directory
276 (map { $_ => $_ } qw(py rb java css js tex bib xml awk bat ini spec tcl sql)),
277 # alternate extensions, see /etc/highlight/filetypes.conf
278 (map { $_ => 'c' } qw(c h)),
279 (map { $_ => 'sh' } qw(sh bash zsh ksh)),
280 (map { $_ => 'cpp' } qw(cpp cxx c++ cc)),
281 (map { $_ => 'php' } qw(php php3 php4 php5 phps)),
282 (map { $_ => 'pl' } qw(pl perl pm)), # perhaps also 'cgi'
283 (map { $_ => 'make'} qw(make mak mk)),
284 (map { $_ => 'xml' } qw(xml xhtml html htm)),
285 );
286
287 # You define site-wide feature defaults here; override them with
288 # $GITWEB_CONFIG as necessary.
289 our %feature = (
290 # feature => {
291 # 'sub' => feature-sub (subroutine),
292 # 'override' => allow-override (boolean),
293 # 'default' => [ default options...] (array reference)}
294 #
295 # if feature is overridable (it means that allow-override has true value),
296 # then feature-sub will be called with default options as parameters;
297 # return value of feature-sub indicates if to enable specified feature
298 #
299 # if there is no 'sub' key (no feature-sub), then feature cannot be
300 # overridden
301 #
302 # use gitweb_get_feature(<feature>) to retrieve the <feature> value
303 # (an array) or gitweb_check_feature(<feature>) to check if <feature>
304 # is enabled
305
306 # Enable the 'blame' blob view, showing the last commit that modified
307 # each line in the file. This can be very CPU-intensive.
308
309 # To enable system wide have in $GITWEB_CONFIG
310 # $feature{'blame'}{'default'} = [1];
311 # To have project specific config enable override in $GITWEB_CONFIG
312 # $feature{'blame'}{'override'} = 1;
313 # and in project config gitweb.blame = 0|1;
314 'blame' => {
315 'sub' => sub { feature_bool('blame', @_) },
316 'override' => 0,
317 'default' => [0]},
318
319 # Enable the 'snapshot' link, providing a compressed archive of any
320 # tree. This can potentially generate high traffic if you have large
321 # project.
322
323 # Value is a list of formats defined in %known_snapshot_formats that
324 # you wish to offer.
325 # To disable system wide have in $GITWEB_CONFIG
326 # $feature{'snapshot'}{'default'} = [];
327 # To have project specific config enable override in $GITWEB_CONFIG
328 # $feature{'snapshot'}{'override'} = 1;
329 # and in project config, a comma-separated list of formats or "none"
330 # to disable. Example: gitweb.snapshot = tbz2,zip;
331 'snapshot' => {
332 'sub' => \&feature_snapshot,
333 'override' => 0,
334 'default' => ['tgz']},
335
336 # Enable text search, which will list the commits which match author,
337 # committer or commit text to a given string. Enabled by default.
338 # Project specific override is not supported.
339 #
340 # Note that this controls all search features, which means that if
341 # it is disabled, then 'grep' and 'pickaxe' search would also be
342 # disabled.
343 'search' => {
344 'override' => 0,
345 'default' => [1]},
346
347 # Enable grep search, which will list the files in currently selected
348 # tree containing the given string. Enabled by default. This can be
349 # potentially CPU-intensive, of course.
350 # Note that you need to have 'search' feature enabled too.
351
352 # To enable system wide have in $GITWEB_CONFIG
353 # $feature{'grep'}{'default'} = [1];
354 # To have project specific config enable override in $GITWEB_CONFIG
355 # $feature{'grep'}{'override'} = 1;
356 # and in project config gitweb.grep = 0|1;
357 'grep' => {
358 'sub' => sub { feature_bool('grep', @_) },
359 'override' => 0,
360 'default' => [1]},
361
362 # Enable the pickaxe search, which will list the commits that modified
363 # a given string in a file. This can be practical and quite faster
364 # alternative to 'blame', but still potentially CPU-intensive.
365 # Note that you need to have 'search' feature enabled too.
366
367 # To enable system wide have in $GITWEB_CONFIG
368 # $feature{'pickaxe'}{'default'} = [1];
369 # To have project specific config enable override in $GITWEB_CONFIG
370 # $feature{'pickaxe'}{'override'} = 1;
371 # and in project config gitweb.pickaxe = 0|1;
372 'pickaxe' => {
373 'sub' => sub { feature_bool('pickaxe', @_) },
374 'override' => 0,
375 'default' => [1]},
376
377 # Enable showing size of blobs in a 'tree' view, in a separate
378 # column, similar to what 'ls -l' does. This cost a bit of IO.
379
380 # To disable system wide have in $GITWEB_CONFIG
381 # $feature{'show-sizes'}{'default'} = [0];
382 # To have project specific config enable override in $GITWEB_CONFIG
383 # $feature{'show-sizes'}{'override'} = 1;
384 # and in project config gitweb.showsizes = 0|1;
385 'show-sizes' => {
386 'sub' => sub { feature_bool('showsizes', @_) },
387 'override' => 0,
388 'default' => [1]},
389
390 # Make gitweb use an alternative format of the URLs which can be
391 # more readable and natural-looking: project name is embedded
392 # directly in the path and the query string contains other
393 # auxiliary information. All gitweb installations recognize
394 # URL in either format; this configures in which formats gitweb
395 # generates links.
396
397 # To enable system wide have in $GITWEB_CONFIG
398 # $feature{'pathinfo'}{'default'} = [1];
399 # Project specific override is not supported.
400
401 # Note that you will need to change the default location of CSS,
402 # favicon, logo and possibly other files to an absolute URL. Also,
403 # if gitweb.cgi serves as your indexfile, you will need to force
404 # $my_uri to contain the script name in your $GITWEB_CONFIG.
405 'pathinfo' => {
406 'override' => 0,
407 'default' => [0]},
408
409 # Make gitweb consider projects in project root subdirectories
410 # to be forks of existing projects. Given project $projname.git,
411 # projects matching $projname/*.git will not be shown in the main
412 # projects list, instead a '+' mark will be added to $projname
413 # there and a 'forks' view will be enabled for the project, listing
414 # all the forks. If project list is taken from a file, forks have
415 # to be listed after the main project.
416
417 # To enable system wide have in $GITWEB_CONFIG
418 # $feature{'forks'}{'default'} = [1];
419 # Project specific override is not supported.
420 'forks' => {
421 'override' => 0,
422 'default' => [0]},
423
424 # Insert custom links to the action bar of all project pages.
425 # This enables you mainly to link to third-party scripts integrating
426 # into gitweb; e.g. git-browser for graphical history representation
427 # or custom web-based repository administration interface.
428
429 # The 'default' value consists of a list of triplets in the form
430 # (label, link, position) where position is the label after which
431 # to insert the link and link is a format string where %n expands
432 # to the project name, %f to the project path within the filesystem,
433 # %h to the current hash (h gitweb parameter) and %b to the current
434 # hash base (hb gitweb parameter); %% expands to %.
435
436 # To enable system wide have in $GITWEB_CONFIG e.g.
437 # $feature{'actions'}{'default'} = [('graphiclog',
438 # '/git-browser/by-commit.html?r=%n', 'summary')];
439 # Project specific override is not supported.
440 'actions' => {
441 'override' => 0,
442 'default' => []},
443
444 # Allow gitweb scan project content tags of project repository,
445 # and display the popular Web 2.0-ish "tag cloud" near the projects
446 # list. Note that this is something COMPLETELY different from the
447 # normal Git tags.
448
449 # gitweb by itself can show existing tags, but it does not handle
450 # tagging itself; you need to do it externally, outside gitweb.
451 # The format is described in git_get_project_ctags() subroutine.
452 # You may want to install the HTML::TagCloud Perl module to get
453 # a pretty tag cloud instead of just a list of tags.
454
455 # To enable system wide have in $GITWEB_CONFIG
456 # $feature{'ctags'}{'default'} = [1];
457 # Project specific override is not supported.
458
459 # In the future whether ctags editing is enabled might depend
460 # on the value, but using 1 should always mean no editing of ctags.
461 'ctags' => {
462 'override' => 0,
463 'default' => [0]},
464
465 # The maximum number of patches in a patchset generated in patch
466 # view. Set this to 0 or undef to disable patch view, or to a
467 # negative number to remove any limit.
468
469 # To disable system wide have in $GITWEB_CONFIG
470 # $feature{'patches'}{'default'} = [0];
471 # To have project specific config enable override in $GITWEB_CONFIG
472 # $feature{'patches'}{'override'} = 1;
473 # and in project config gitweb.patches = 0|n;
474 # where n is the maximum number of patches allowed in a patchset.
475 'patches' => {
476 'sub' => \&feature_patches,
477 'override' => 0,
478 'default' => [16]},
479
480 # Avatar support. When this feature is enabled, views such as
481 # shortlog or commit will display an avatar associated with
482 # the email of the committer(s) and/or author(s).
483
484 # Currently available providers are gravatar and picon.
485 # If an unknown provider is specified, the feature is disabled.
486
487 # Gravatar depends on Digest::MD5.
488 # Picon currently relies on the indiana.edu database.
489
490 # To enable system wide have in $GITWEB_CONFIG
491 # $feature{'avatar'}{'default'} = ['<provider>'];
492 # where <provider> is either gravatar or picon.
493 # To have project specific config enable override in $GITWEB_CONFIG
494 # $feature{'avatar'}{'override'} = 1;
495 # and in project config gitweb.avatar = <provider>;
496 'avatar' => {
497 'sub' => \&feature_avatar,
498 'override' => 0,
499 'default' => ['']},
500
501 # Enable displaying how much time and how many git commands
502 # it took to generate and display page. Disabled by default.
503 # Project specific override is not supported.
504 'timed' => {
505 'override' => 0,
506 'default' => [0]},
507
508 # Enable turning some links into links to actions which require
509 # JavaScript to run (like 'blame_incremental'). Not enabled by
510 # default. Project specific override is currently not supported.
511 'javascript-actions' => {
512 'override' => 0,
513 'default' => [0]},
514
515 # Enable and configure ability to change common timezone for dates
516 # in gitweb output via JavaScript. Enabled by default.
517 # Project specific override is not supported.
518 'javascript-timezone' => {
519 'override' => 0,
520 'default' => [
521 'local', # default timezone: 'utc', 'local', or '(-|+)HHMM' format,
522 # or undef to turn off this feature
523 'gitweb_tz', # name of cookie where to store selected timezone
524 'datetime', # CSS class used to mark up dates for manipulation
525 ]},
526
527 # Syntax highlighting support. This is based on Daniel Svensson's
528 # and Sham Chukoury's work in gitweb-xmms2.git.
529 # It requires the 'highlight' program present in $PATH,
530 # and therefore is disabled by default.
531
532 # To enable system wide have in $GITWEB_CONFIG
533 # $feature{'highlight'}{'default'} = [1];
534
535 'highlight' => {
536 'sub' => sub { feature_bool('highlight', @_) },
537 'override' => 0,
538 'default' => [0]},
539
540 # Enable displaying of remote heads in the heads list
541
542 # To enable system wide have in $GITWEB_CONFIG
543 # $feature{'remote_heads'}{'default'} = [1];
544 # To have project specific config enable override in $GITWEB_CONFIG
545 # $feature{'remote_heads'}{'override'} = 1;
546 # and in project config gitweb.remoteheads = 0|1;
547 'remote_heads' => {
548 'sub' => sub { feature_bool('remote_heads', @_) },
549 'override' => 0,
550 'default' => [0]},
551
552 # Enable showing branches under other refs in addition to heads
553
554 # To set system wide extra branch refs have in $GITWEB_CONFIG
555 # $feature{'extra-branch-refs'}{'default'} = ['dirs', 'of', 'choice'];
556 # To have project specific config enable override in $GITWEB_CONFIG
557 # $feature{'extra-branch-refs'}{'override'} = 1;
558 # and in project config gitweb.extrabranchrefs = dirs of choice
559 # Every directory is separated with whitespace.
560
561 'extra-branch-refs' => {
562 'sub' => \&feature_extra_branch_refs,
563 'override' => 0,
564 'default' => []},
565 );
566
567 sub gitweb_get_feature {
568 my ($name) = @_;
569 return unless exists $feature{$name};
570 my ($sub, $override, @defaults) = (
571 $feature{$name}{'sub'},
572 $feature{$name}{'override'},
573 @{$feature{$name}{'default'}});
574 # project specific override is possible only if we have project
575 our $git_dir; # global variable, declared later
576 if (!$override || !defined $git_dir) {
577 return @defaults;
578 }
579 if (!defined $sub) {
580 warn "feature $name is not overridable";
581 return @defaults;
582 }
583 return $sub->(@defaults);
584 }
585
586 # A wrapper to check if a given feature is enabled.
587 # With this, you can say
588 #
589 # my $bool_feat = gitweb_check_feature('bool_feat');
590 # gitweb_check_feature('bool_feat') or somecode;
591 #
592 # instead of
593 #
594 # my ($bool_feat) = gitweb_get_feature('bool_feat');
595 # (gitweb_get_feature('bool_feat'))[0] or somecode;
596 #
597 sub gitweb_check_feature {
598 return (gitweb_get_feature(@_))[0];
599 }
600
601
602 sub feature_bool {
603 my $key = shift;
604 my ($val) = git_get_project_config($key, '--bool');
605
606 if (!defined $val) {
607 return ($_[0]);
608 } elsif ($val eq 'true') {
609 return (1);
610 } elsif ($val eq 'false') {
611 return (0);
612 }
613 }
614
615 sub feature_snapshot {
616 my (@fmts) = @_;
617
618 my ($val) = git_get_project_config('snapshot');
619
620 if ($val) {
621 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
622 }
623
624 return @fmts;
625 }
626
627 sub feature_patches {
628 my @val = (git_get_project_config('patches', '--int'));
629
630 if (@val) {
631 return @val;
632 }
633
634 return ($_[0]);
635 }
636
637 sub feature_avatar {
638 my @val = (git_get_project_config('avatar'));
639
640 return @val ? @val : @_;
641 }
642
643 sub feature_extra_branch_refs {
644 my (@branch_refs) = @_;
645 my $values = git_get_project_config('extrabranchrefs');
646
647 if ($values) {
648 $values = config_to_multi ($values);
649 @branch_refs = ();
650 foreach my $value (@{$values}) {
651 push @branch_refs, split /\s+/, $value;
652 }
653 }
654
655 return @branch_refs;
656 }
657
658 # checking HEAD file with -e is fragile if the repository was
659 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
660 # and then pruned.
661 sub check_head_link {
662 my ($dir) = @_;
663 my $headfile = "$dir/HEAD";
664 return ((-e $headfile) ||
665 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
666 }
667
668 sub check_export_ok {
669 my ($dir) = @_;
670 return (check_head_link($dir) &&
671 (!$export_ok || -e "$dir/$export_ok") &&
672 (!$export_auth_hook || $export_auth_hook->($dir)));
673 }
674
675 # process alternate names for backward compatibility
676 # filter out unsupported (unknown) snapshot formats
677 sub filter_snapshot_fmts {
678 my @fmts = @_;
679
680 @fmts = map {
681 exists $known_snapshot_format_aliases{$_} ?
682 $known_snapshot_format_aliases{$_} : $_} @fmts;
683 @fmts = grep {
684 exists $known_snapshot_formats{$_} &&
685 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
686 }
687
688 sub filter_and_validate_refs {
689 my @refs = @_;
690 my %unique_refs = ();
691
692 foreach my $ref (@refs) {
693 die_error(500, "Invalid ref '$ref' in 'extra-branch-refs' feature") unless (is_valid_ref_format($ref));
694 # 'heads' are added implicitly in get_branch_refs().
695 $unique_refs{$ref} = 1 if ($ref ne 'heads');
696 }
697 return sort keys %unique_refs;
698 }
699
700 # If it is set to code reference, it is code that it is to be run once per
701 # request, allowing updating configurations that change with each request,
702 # while running other code in config file only once.
703 #
704 # Otherwise, if it is false then gitweb would process config file only once;
705 # if it is true then gitweb config would be run for each request.
706 our $per_request_config = 1;
707
708 # read and parse gitweb config file given by its parameter.
709 # returns true on success, false on recoverable error, allowing
710 # to chain this subroutine, using first file that exists.
711 # dies on errors during parsing config file, as it is unrecoverable.
712 sub read_config_file {
713 my $filename = shift;
714 return unless defined $filename;
715 # die if there are errors parsing config file
716 if (-e $filename) {
717 do $filename;
718 die $@ if $@;
719 return 1;
720 }
721 return;
722 }
723
724 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM, $GITWEB_CONFIG_COMMON);
725 sub evaluate_gitweb_config {
726 our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
727 our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
728 our $GITWEB_CONFIG_COMMON = $ENV{'GITWEB_CONFIG_COMMON'} || "++GITWEB_CONFIG_COMMON++";
729
730 # Protect against duplications of file names, to not read config twice.
731 # Only one of $GITWEB_CONFIG and $GITWEB_CONFIG_SYSTEM is used, so
732 # there possibility of duplication of filename there doesn't matter.
733 $GITWEB_CONFIG = "" if ($GITWEB_CONFIG eq $GITWEB_CONFIG_COMMON);
734 $GITWEB_CONFIG_SYSTEM = "" if ($GITWEB_CONFIG_SYSTEM eq $GITWEB_CONFIG_COMMON);
735
736 # Common system-wide settings for convenience.
737 # Those settings can be ovverriden by GITWEB_CONFIG or GITWEB_CONFIG_SYSTEM.
738 read_config_file($GITWEB_CONFIG_COMMON);
739
740 # Use first config file that exists. This means use the per-instance
741 # GITWEB_CONFIG if exists, otherwise use GITWEB_SYSTEM_CONFIG.
742 read_config_file($GITWEB_CONFIG) and return;
743 read_config_file($GITWEB_CONFIG_SYSTEM);
744 }
745
746 # Get loadavg of system, to compare against $maxload.
747 # Currently it requires '/proc/loadavg' present to get loadavg;
748 # if it is not present it returns 0, which means no load checking.
749 sub get_loadavg {
750 if( -e '/proc/loadavg' ){
751 open my $fd, '<', '/proc/loadavg'
752 or return 0;
753 my @load = split(/\s+/, scalar <$fd>);
754 close $fd;
755
756 # The first three columns measure CPU and IO utilization of the last one,
757 # five, and 10 minute periods. The fourth column shows the number of
758 # currently running processes and the total number of processes in the m/n
759 # format. The last column displays the last process ID used.
760 return $load[0] || 0;
761 }
762 # additional checks for load average should go here for things that don't export
763 # /proc/loadavg
764
765 return 0;
766 }
767
768 # version of the core git binary
769 our $git_version;
770 sub evaluate_git_version {
771 our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
772 $number_of_git_cmds++;
773 }
774
775 sub check_loadavg {
776 if (defined $maxload && get_loadavg() > $maxload) {
777 die_error(503, "The load average on the server is too high");
778 }
779 }
780
781 # ======================================================================
782 # input validation and dispatch
783
784 # input parameters can be collected from a variety of sources (presently, CGI
785 # and PATH_INFO), so we define an %input_params hash that collects them all
786 # together during validation: this allows subsequent uses (e.g. href()) to be
787 # agnostic of the parameter origin
788
789 our %input_params = ();
790
791 # input parameters are stored with the long parameter name as key. This will
792 # also be used in the href subroutine to convert parameters to their CGI
793 # equivalent, and since the href() usage is the most frequent one, we store
794 # the name -> CGI key mapping here, instead of the reverse.
795 #
796 # XXX: Warning: If you touch this, check the search form for updating,
797 # too.
798
799 our @cgi_param_mapping = (
800 project => "p",
801 action => "a",
802 file_name => "f",
803 file_parent => "fp",
804 hash => "h",
805 hash_parent => "hp",
806 hash_base => "hb",
807 hash_parent_base => "hpb",
808 page => "pg",
809 order => "o",
810 searchtext => "s",
811 searchtype => "st",
812 snapshot_format => "sf",
813 extra_options => "opt",
814 search_use_regexp => "sr",
815 ctag => "by_tag",
816 diff_style => "ds",
817 project_filter => "pf",
818 # this must be last entry (for manipulation from JavaScript)
819 javascript => "js"
820 );
821 our %cgi_param_mapping = @cgi_param_mapping;
822
823 # we will also need to know the possible actions, for validation
824 our %actions = (
825 "blame" => \&git_blame,
826 "blame_incremental" => \&git_blame_incremental,
827 "blame_data" => \&git_blame_data,
828 "blobdiff" => \&git_blobdiff,
829 "blobdiff_plain" => \&git_blobdiff_plain,
830 "blob" => \&git_blob,
831 "blob_plain" => \&git_blob_plain,
832 "commitdiff" => \&git_commitdiff,
833 "commitdiff_plain" => \&git_commitdiff_plain,
834 "commit" => \&git_commit,
835 "forks" => \&git_forks,
836 "heads" => \&git_heads,
837 "history" => \&git_history,
838 "log" => \&git_log,
839 "patch" => \&git_patch,
840 "patches" => \&git_patches,
841 "remotes" => \&git_remotes,
842 "rss" => \&git_rss,
843 "atom" => \&git_atom,
844 "search" => \&git_search,
845 "search_help" => \&git_search_help,
846 "shortlog" => \&git_shortlog,
847 "summary" => \&git_summary,
848 "tag" => \&git_tag,
849 "tags" => \&git_tags,
850 "tree" => \&git_tree,
851 "snapshot" => \&git_snapshot,
852 "object" => \&git_object,
853 # those below don't need $project
854 "opml" => \&git_opml,
855 "project_list" => \&git_project_list,
856 "project_index" => \&git_project_index,
857 );
858
859 # finally, we have the hash of allowed extra_options for the commands that
860 # allow them
861 our %allowed_options = (
862 "--no-merges" => [ qw(rss atom log shortlog history) ],
863 );
864
865 # fill %input_params with the CGI parameters. All values except for 'opt'
866 # should be single values, but opt can be an array. We should probably
867 # build an array of parameters that can be multi-valued, but since for the time
868 # being it's only this one, we just single it out
869 sub evaluate_query_params {
870 our $cgi;
871
872 while (my ($name, $symbol) = each %cgi_param_mapping) {
873 if ($symbol eq 'opt') {
874 $input_params{$name} = [ map { decode_utf8($_) } $cgi->param($symbol) ];
875 } else {
876 $input_params{$name} = decode_utf8($cgi->param($symbol));
877 }
878 }
879 }
880
881 # now read PATH_INFO and update the parameter list for missing parameters
882 sub evaluate_path_info {
883 return if defined $input_params{'project'};
884 return if !$path_info;
885 $path_info =~ s,^/+,,;
886 return if !$path_info;
887
888 # find which part of PATH_INFO is project
889 my $project = $path_info;
890 $project =~ s,/+$,,;
891 while ($project && !check_head_link("$projectroot/$project")) {
892 $project =~ s,/*[^/]*$,,;
893 }
894 return unless $project;
895 $input_params{'project'} = $project;
896
897 # do not change any parameters if an action is given using the query string
898 return if $input_params{'action'};
899 $path_info =~ s,^\Q$project\E/*,,;
900
901 # next, check if we have an action
902 my $action = $path_info;
903 $action =~ s,/.*$,,;
904 if (exists $actions{$action}) {
905 $path_info =~ s,^$action/*,,;
906 $input_params{'action'} = $action;
907 }
908
909 # list of actions that want hash_base instead of hash, but can have no
910 # pathname (f) parameter
911 my @wants_base = (
912 'tree',
913 'history',
914 );
915
916 # we want to catch, among others
917 # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
918 my ($parentrefname, $parentpathname, $refname, $pathname) =
919 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
920
921 # first, analyze the 'current' part
922 if (defined $pathname) {
923 # we got "branch:filename" or "branch:dir/"
924 # we could use git_get_type(branch:pathname), but:
925 # - it needs $git_dir
926 # - it does a git() call
927 # - the convention of terminating directories with a slash
928 # makes it superfluous
929 # - embedding the action in the PATH_INFO would make it even
930 # more superfluous
931 $pathname =~ s,^/+,,;
932 if (!$pathname || substr($pathname, -1) eq "/") {
933 $input_params{'action'} ||= "tree";
934 $pathname =~ s,/$,,;
935 } else {
936 # the default action depends on whether we had parent info
937 # or not
938 if ($parentrefname) {
939 $input_params{'action'} ||= "blobdiff_plain";
940 } else {
941 $input_params{'action'} ||= "blob_plain";
942 }
943 }
944 $input_params{'hash_base'} ||= $refname;
945 $input_params{'file_name'} ||= $pathname;
946 } elsif (defined $refname) {
947 # we got "branch". In this case we have to choose if we have to
948 # set hash or hash_base.
949 #
950 # Most of the actions without a pathname only want hash to be
951 # set, except for the ones specified in @wants_base that want
952 # hash_base instead. It should also be noted that hand-crafted
953 # links having 'history' as an action and no pathname or hash
954 # set will fail, but that happens regardless of PATH_INFO.
955 if (defined $parentrefname) {
956 # if there is parent let the default be 'shortlog' action
957 # (for http://git.example.com/repo.git/A..B links); if there
958 # is no parent, dispatch will detect type of object and set
959 # action appropriately if required (if action is not set)
960 $input_params{'action'} ||= "shortlog";
961 }
962 if ($input_params{'action'} &&
963 grep { $_ eq $input_params{'action'} } @wants_base) {
964 $input_params{'hash_base'} ||= $refname;
965 } else {
966 $input_params{'hash'} ||= $refname;
967 }
968 }
969
970 # next, handle the 'parent' part, if present
971 if (defined $parentrefname) {
972 # a missing pathspec defaults to the 'current' filename, allowing e.g.
973 # someproject/blobdiff/oldrev..newrev:/filename
974 if ($parentpathname) {
975 $parentpathname =~ s,^/+,,;
976 $parentpathname =~ s,/$,,;
977 $input_params{'file_parent'} ||= $parentpathname;
978 } else {
979 $input_params{'file_parent'} ||= $input_params{'file_name'};
980 }
981 # we assume that hash_parent_base is wanted if a path was specified,
982 # or if the action wants hash_base instead of hash
983 if (defined $input_params{'file_parent'} ||
984 grep { $_ eq $input_params{'action'} } @wants_base) {
985 $input_params{'hash_parent_base'} ||= $parentrefname;
986 } else {
987 $input_params{'hash_parent'} ||= $parentrefname;
988 }
989 }
990
991 # for the snapshot action, we allow URLs in the form
992 # $project/snapshot/$hash.ext
993 # where .ext determines the snapshot and gets removed from the
994 # passed $refname to provide the $hash.
995 #
996 # To be able to tell that $refname includes the format extension, we
997 # require the following two conditions to be satisfied:
998 # - the hash input parameter MUST have been set from the $refname part
999 # of the URL (i.e. they must be equal)
1000 # - the snapshot format MUST NOT have been defined already (e.g. from
1001 # CGI parameter sf)
1002 # It's also useless to try any matching unless $refname has a dot,
1003 # so we check for that too
1004 if (defined $input_params{'action'} &&
1005 $input_params{'action'} eq 'snapshot' &&
1006 defined $refname && index($refname, '.') != -1 &&
1007 $refname eq $input_params{'hash'} &&
1008 !defined $input_params{'snapshot_format'}) {
1009 # We loop over the known snapshot formats, checking for
1010 # extensions. Allowed extensions are both the defined suffix
1011 # (which includes the initial dot already) and the snapshot
1012 # format key itself, with a prepended dot
1013 while (my ($fmt, $opt) = each %known_snapshot_formats) {
1014 my $hash = $refname;
1015 unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
1016 next;
1017 }
1018 my $sfx = $1;
1019 # a valid suffix was found, so set the snapshot format
1020 # and reset the hash parameter
1021 $input_params{'snapshot_format'} = $fmt;
1022 $input_params{'hash'} = $hash;
1023 # we also set the format suffix to the one requested
1024 # in the URL: this way a request for e.g. .tgz returns
1025 # a .tgz instead of a .tar.gz
1026 $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
1027 last;
1028 }
1029 }
1030 }
1031
1032 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
1033 $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
1034 $searchtext, $search_regexp, $project_filter);
1035 sub evaluate_and_validate_params {
1036 our $action = $input_params{'action'};
1037 if (defined $action) {
1038 if (!is_valid_action($action)) {
1039 die_error(400, "Invalid action parameter");
1040 }
1041 }
1042
1043 # parameters which are pathnames
1044 our $project = $input_params{'project'};
1045 if (defined $project) {
1046 if (!is_valid_project($project)) {
1047 undef $project;
1048 die_error(404, "No such project");
1049 }
1050 }
1051
1052 our $project_filter = $input_params{'project_filter'};
1053 if (defined $project_filter) {
1054 if (!is_valid_pathname($project_filter)) {
1055 die_error(404, "Invalid project_filter parameter");
1056 }
1057 }
1058
1059 our $file_name = $input_params{'file_name'};
1060 if (defined $file_name) {
1061 if (!is_valid_pathname($file_name)) {
1062 die_error(400, "Invalid file parameter");
1063 }
1064 }
1065
1066 our $file_parent = $input_params{'file_parent'};
1067 if (defined $file_parent) {
1068 if (!is_valid_pathname($file_parent)) {
1069 die_error(400, "Invalid file parent parameter");
1070 }
1071 }
1072
1073 # parameters which are refnames
1074 our $hash = $input_params{'hash'};
1075 if (defined $hash) {
1076 if (!is_valid_refname($hash)) {
1077 die_error(400, "Invalid hash parameter");
1078 }
1079 }
1080
1081 our $hash_parent = $input_params{'hash_parent'};
1082 if (defined $hash_parent) {
1083 if (!is_valid_refname($hash_parent)) {
1084 die_error(400, "Invalid hash parent parameter");
1085 }
1086 }
1087
1088 our $hash_base = $input_params{'hash_base'};
1089 if (defined $hash_base) {
1090 if (!is_valid_refname($hash_base)) {
1091 die_error(400, "Invalid hash base parameter");
1092 }
1093 }
1094
1095 our @extra_options = @{$input_params{'extra_options'}};
1096 # @extra_options is always defined, since it can only be (currently) set from
1097 # CGI, and $cgi->param() returns the empty array in array context if the param
1098 # is not set
1099 foreach my $opt (@extra_options) {
1100 if (not exists $allowed_options{$opt}) {
1101 die_error(400, "Invalid option parameter");
1102 }
1103 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1104 die_error(400, "Invalid option parameter for this action");
1105 }
1106 }
1107
1108 our $hash_parent_base = $input_params{'hash_parent_base'};
1109 if (defined $hash_parent_base) {
1110 if (!is_valid_refname($hash_parent_base)) {
1111 die_error(400, "Invalid hash parent base parameter");
1112 }
1113 }
1114
1115 # other parameters
1116 our $page = $input_params{'page'};
1117 if (defined $page) {
1118 if ($page =~ m/[^0-9]/) {
1119 die_error(400, "Invalid page parameter");
1120 }
1121 }
1122
1123 our $searchtype = $input_params{'searchtype'};
1124 if (defined $searchtype) {
1125 if ($searchtype =~ m/[^a-z]/) {
1126 die_error(400, "Invalid searchtype parameter");
1127 }
1128 }
1129
1130 our $search_use_regexp = $input_params{'search_use_regexp'};
1131
1132 our $searchtext = $input_params{'searchtext'};
1133 our $search_regexp = undef;
1134 if (defined $searchtext) {
1135 if (length($searchtext) < 2) {
1136 die_error(403, "At least two characters are required for search parameter");
1137 }
1138 if ($search_use_regexp) {
1139 $search_regexp = $searchtext;
1140 if (!eval { qr/$search_regexp/; 1; }) {
1141 (my $error = $@) =~ s/ at \S+ line \d+.*\n?//;
1142 die_error(400, "Invalid search regexp '$search_regexp'",
1143 esc_html($error));
1144 }
1145 } else {
1146 $search_regexp = quotemeta $searchtext;
1147 }
1148 }
1149 }
1150
1151 # path to the current git repository
1152 our $git_dir;
1153 sub evaluate_git_dir {
1154 our $git_dir = "$projectroot/$project" if $project;
1155 }
1156
1157 our (@snapshot_fmts, $git_avatar, @extra_branch_refs);
1158 sub configure_gitweb_features {
1159 # list of supported snapshot formats
1160 our @snapshot_fmts = gitweb_get_feature('snapshot');
1161 @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1162
1163 # check that the avatar feature is set to a known provider name,
1164 # and for each provider check if the dependencies are satisfied.
1165 # if the provider name is invalid or the dependencies are not met,
1166 # reset $git_avatar to the empty string.
1167 our ($git_avatar) = gitweb_get_feature('avatar');
1168 if ($git_avatar eq 'gravatar') {
1169 $git_avatar = '' unless (eval { require Digest::MD5; 1; });
1170 } elsif ($git_avatar eq 'picon') {
1171 # no dependencies
1172 } else {
1173 $git_avatar = '';
1174 }
1175
1176 our @extra_branch_refs = gitweb_get_feature('extra-branch-refs');
1177 @extra_branch_refs = filter_and_validate_refs (@extra_branch_refs);
1178 }
1179
1180 sub get_branch_refs {
1181 return ('heads', @extra_branch_refs);
1182 }
1183
1184 # custom error handler: 'die <message>' is Internal Server Error
1185 sub handle_errors_html {
1186 my $msg = shift; # it is already HTML escaped
1187
1188 # to avoid infinite loop where error occurs in die_error,
1189 # change handler to default handler, disabling handle_errors_html
1190 set_message("Error occurred when inside die_error:\n$msg");
1191
1192 # you cannot jump out of die_error when called as error handler;
1193 # the subroutine set via CGI::Carp::set_message is called _after_
1194 # HTTP headers are already written, so it cannot write them itself
1195 die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1196 }
1197 set_message(\&handle_errors_html);
1198
1199 # dispatch
1200 sub dispatch {
1201 if (!defined $action) {
1202 if (defined $hash) {
1203 $action = git_get_type($hash);
1204 $action or die_error(404, "Object does not exist");
1205 } elsif (defined $hash_base && defined $file_name) {
1206 $action = git_get_type("$hash_base:$file_name");
1207 $action or die_error(404, "File or directory does not exist");
1208 } elsif (defined $project) {
1209 $action = 'summary';
1210 } else {
1211 $action = 'project_list';
1212 }
1213 }
1214 if (!defined($actions{$action})) {
1215 die_error(400, "Unknown action");
1216 }
1217 if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1218 !$project) {
1219 die_error(400, "Project needed");
1220 }
1221 $actions{$action}->();
1222 }
1223
1224 sub reset_timer {
1225 our $t0 = [ gettimeofday() ]
1226 if defined $t0;
1227 our $number_of_git_cmds = 0;
1228 }
1229
1230 our $first_request = 1;
1231 sub run_request {
1232 reset_timer();
1233
1234 evaluate_uri();
1235 if ($first_request) {
1236 evaluate_gitweb_config();
1237 evaluate_git_version();
1238 }
1239 if ($per_request_config) {
1240 if (ref($per_request_config) eq 'CODE') {
1241 $per_request_config->();
1242 } elsif (!$first_request) {
1243 evaluate_gitweb_config();
1244 }
1245 }
1246 check_loadavg();
1247
1248 # $projectroot and $projects_list might be set in gitweb config file
1249 $projects_list ||= $projectroot;
1250
1251 evaluate_query_params();
1252 evaluate_path_info();
1253 evaluate_and_validate_params();
1254 evaluate_git_dir();
1255
1256 configure_gitweb_features();
1257
1258 dispatch();
1259 }
1260
1261 our $is_last_request = sub { 1 };
1262 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1263 our $CGI = 'CGI';
1264 our $cgi;
1265 sub configure_as_fcgi {
1266 require CGI::Fast;
1267 our $CGI = 'CGI::Fast';
1268
1269 my $request_number = 0;
1270 # let each child service 100 requests
1271 our $is_last_request = sub { ++$request_number > 100 };
1272 }
1273 sub evaluate_argv {
1274 my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1275 configure_as_fcgi()
1276 if $script_name =~ /\.fcgi$/;
1277
1278 return unless (@ARGV);
1279
1280 require Getopt::Long;
1281 Getopt::Long::GetOptions(
1282 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1283 'nproc|n=i' => sub {
1284 my ($arg, $val) = @_;
1285 return unless eval { require FCGI::ProcManager; 1; };
1286 my $proc_manager = FCGI::ProcManager->new({
1287 n_processes => $val,
1288 });
1289 our $pre_listen_hook = sub { $proc_manager->pm_manage() };
1290 our $pre_dispatch_hook = sub { $proc_manager->pm_pre_dispatch() };
1291 our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1292 },
1293 );
1294 }
1295
1296 sub run {
1297 evaluate_argv();
1298
1299 $first_request = 1;
1300 $pre_listen_hook->()
1301 if $pre_listen_hook;
1302
1303 REQUEST:
1304 while ($cgi = $CGI->new()) {
1305 $pre_dispatch_hook->()
1306 if $pre_dispatch_hook;
1307
1308 run_request();
1309
1310 $post_dispatch_hook->()
1311 if $post_dispatch_hook;
1312 $first_request = 0;
1313
1314 last REQUEST if ($is_last_request->());
1315 }
1316
1317 DONE_GITWEB:
1318 1;
1319 }
1320
1321 run();
1322
1323 if (defined caller) {
1324 # wrapped in a subroutine processing requests,
1325 # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1326 return;
1327 } else {
1328 # pure CGI script, serving single request
1329 exit;
1330 }
1331
1332 ## ======================================================================
1333 ## action links
1334
1335 # possible values of extra options
1336 # -full => 0|1 - use absolute/full URL ($my_uri/$my_url as base)
1337 # -replay => 1 - start from a current view (replay with modifications)
1338 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1339 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1340 sub href {
1341 my %params = @_;
1342 # default is to use -absolute url() i.e. $my_uri
1343 my $href = $params{-full} ? $my_url : $my_uri;
1344
1345 # implicit -replay, must be first of implicit params
1346 $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1347
1348 $params{'project'} = $project unless exists $params{'project'};
1349
1350 if ($params{-replay}) {
1351 while (my ($name, $symbol) = each %cgi_param_mapping) {
1352 if (!exists $params{$name}) {
1353 $params{$name} = $input_params{$name};
1354 }
1355 }
1356 }
1357
1358 my $use_pathinfo = gitweb_check_feature('pathinfo');
1359 if (defined $params{'project'} &&
1360 (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1361 # try to put as many parameters as possible in PATH_INFO:
1362 # - project name
1363 # - action
1364 # - hash_parent or hash_parent_base:/file_parent
1365 # - hash or hash_base:/filename
1366 # - the snapshot_format as an appropriate suffix
1367
1368 # When the script is the root DirectoryIndex for the domain,
1369 # $href here would be something like http://gitweb.example.com/
1370 # Thus, we strip any trailing / from $href, to spare us double
1371 # slashes in the final URL
1372 $href =~ s,/$,,;
1373
1374 # Then add the project name, if present
1375 $href .= "/".esc_path_info($params{'project'});
1376 delete $params{'project'};
1377
1378 # since we destructively absorb parameters, we keep this
1379 # boolean that remembers if we're handling a snapshot
1380 my $is_snapshot = $params{'action'} eq 'snapshot';
1381
1382 # Summary just uses the project path URL, any other action is
1383 # added to the URL
1384 if (defined $params{'action'}) {
1385 $href .= "/".esc_path_info($params{'action'})
1386 unless $params{'action'} eq 'summary';
1387 delete $params{'action'};
1388 }
1389
1390 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1391 # stripping nonexistent or useless pieces
1392 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1393 || $params{'hash_parent'} || $params{'hash'});
1394 if (defined $params{'hash_base'}) {
1395 if (defined $params{'hash_parent_base'}) {
1396 $href .= esc_path_info($params{'hash_parent_base'});
1397 # skip the file_parent if it's the same as the file_name
1398 if (defined $params{'file_parent'}) {
1399 if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1400 delete $params{'file_parent'};
1401 } elsif ($params{'file_parent'} !~ /\.\./) {
1402 $href .= ":/".esc_path_info($params{'file_parent'});
1403 delete $params{'file_parent'};
1404 }
1405 }
1406 $href .= "..";
1407 delete $params{'hash_parent'};
1408 delete $params{'hash_parent_base'};
1409 } elsif (defined $params{'hash_parent'}) {
1410 $href .= esc_path_info($params{'hash_parent'}). "..";
1411 delete $params{'hash_parent'};
1412 }
1413
1414 $href .= esc_path_info($params{'hash_base'});
1415 if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1416 $href .= ":/".esc_path_info($params{'file_name'});
1417 delete $params{'file_name'};
1418 }
1419 delete $params{'hash'};
1420 delete $params{'hash_base'};
1421 } elsif (defined $params{'hash'}) {
1422 $href .= esc_path_info($params{'hash'});
1423 delete $params{'hash'};
1424 }
1425
1426 # If the action was a snapshot, we can absorb the
1427 # snapshot_format parameter too
1428 if ($is_snapshot) {
1429 my $fmt = $params{'snapshot_format'};
1430 # snapshot_format should always be defined when href()
1431 # is called, but just in case some code forgets, we
1432 # fall back to the default
1433 $fmt ||= $snapshot_fmts[0];
1434 $href .= $known_snapshot_formats{$fmt}{'suffix'};
1435 delete $params{'snapshot_format'};
1436 }
1437 }
1438
1439 # now encode the parameters explicitly
1440 my @result = ();
1441 for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1442 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1443 if (defined $params{$name}) {
1444 if (ref($params{$name}) eq "ARRAY") {
1445 foreach my $par (@{$params{$name}}) {
1446 push @result, $symbol . "=" . esc_param($par);
1447 }
1448 } else {
1449 push @result, $symbol . "=" . esc_param($params{$name});
1450 }
1451 }
1452 }
1453 $href .= "?" . join(';', @result) if scalar @result;
1454
1455 # final transformation: trailing spaces must be escaped (URI-encoded)
1456 $href =~ s/(\s+)$/CGI::escape($1)/e;
1457
1458 if ($params{-anchor}) {
1459 $href .= "#".esc_param($params{-anchor});
1460 }
1461
1462 return $href;
1463 }
1464
1465
1466 ## ======================================================================
1467 ## validation, quoting/unquoting and escaping
1468
1469 sub is_valid_action {
1470 my $input = shift;
1471 return undef unless exists $actions{$input};
1472 return 1;
1473 }
1474
1475 sub is_valid_project {
1476 my $input = shift;
1477
1478 return unless defined $input;
1479 if (!is_valid_pathname($input) ||
1480 !(-d "$projectroot/$input") ||
1481 !check_export_ok("$projectroot/$input") ||
1482 ($strict_export && !project_in_list($input))) {
1483 return undef;
1484 } else {
1485 return 1;
1486 }
1487 }
1488
1489 sub is_valid_pathname {
1490 my $input = shift;
1491
1492 return undef unless defined $input;
1493 # no '.' or '..' as elements of path, i.e. no '.' or '..'
1494 # at the beginning, at the end, and between slashes.
1495 # also this catches doubled slashes
1496 if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1497 return undef;
1498 }
1499 # no null characters
1500 if ($input =~ m!\0!) {
1501 return undef;
1502 }
1503 return 1;
1504 }
1505
1506 sub is_valid_ref_format {
1507 my $input = shift;
1508
1509 return undef unless defined $input;
1510 # restrictions on ref name according to git-check-ref-format
1511 if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1512 return undef;
1513 }
1514 return 1;
1515 }
1516
1517 sub is_valid_refname {
1518 my $input = shift;
1519
1520 return undef unless defined $input;
1521 # textual hashes are O.K.
1522 if ($input =~ m/^[0-9a-fA-F]{40}$/) {
1523 return 1;
1524 }
1525 # it must be correct pathname
1526 is_valid_pathname($input) or return undef;
1527 # check git-check-ref-format restrictions
1528 is_valid_ref_format($input) or return undef;
1529 return 1;
1530 }
1531
1532 # decode sequences of octets in utf8 into Perl's internal form,
1533 # which is utf-8 with utf8 flag set if needed. gitweb writes out
1534 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1535 sub to_utf8 {
1536 my $str = shift;
1537 return undef unless defined $str;
1538
1539 if (utf8::is_utf8($str) || utf8::decode($str)) {
1540 return $str;
1541 } else {
1542 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1543 }
1544 }
1545
1546 # quote unsafe chars, but keep the slash, even when it's not
1547 # correct, but quoted slashes look too horrible in bookmarks
1548 sub esc_param {
1549 my $str = shift;
1550 return undef unless defined $str;
1551 $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1552 $str =~ s/ /\+/g;
1553 return $str;
1554 }
1555
1556 # the quoting rules for path_info fragment are slightly different
1557 sub esc_path_info {
1558 my $str = shift;
1559 return undef unless defined $str;
1560
1561 # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1562 $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1563
1564 return $str;
1565 }
1566
1567 # quote unsafe chars in whole URL, so some characters cannot be quoted
1568 sub esc_url {
1569 my $str = shift;
1570 return undef unless defined $str;
1571 $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1572 $str =~ s/ /\+/g;
1573 return $str;
1574 }
1575
1576 # quote unsafe characters in HTML attributes
1577 sub esc_attr {
1578
1579 # for XHTML conformance escaping '"' to '&quot;' is not enough
1580 return esc_html(@_);
1581 }
1582
1583 # replace invalid utf8 character with SUBSTITUTION sequence
1584 sub esc_html {
1585 my $str = shift;
1586 my %opts = @_;
1587
1588 return undef unless defined $str;
1589
1590 $str = to_utf8($str);
1591 $str = $cgi->escapeHTML($str);
1592 if ($opts{'-nbsp'}) {
1593 $str =~ s/ /&nbsp;/g;
1594 }
1595 $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1596 return $str;
1597 }
1598
1599 # quote control characters and escape filename to HTML
1600 sub esc_path {
1601 my $str = shift;
1602 my %opts = @_;
1603
1604 return undef unless defined $str;
1605
1606 $str = to_utf8($str);
1607 $str = $cgi->escapeHTML($str);
1608 if ($opts{'-nbsp'}) {
1609 $str =~ s/ /&nbsp;/g;
1610 }
1611 $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1612 return $str;
1613 }
1614
1615 # Sanitize for use in XHTML + application/xml+xhtm (valid XML 1.0)
1616 sub sanitize {
1617 my $str = shift;
1618
1619 return undef unless defined $str;
1620
1621 $str = to_utf8($str);
1622 $str =~ s|([[:cntrl:]])|(index("\t\n\r", $1) != -1 ? $1 : quot_cec($1))|eg;
1623 return $str;
1624 }
1625
1626 # Make control characters "printable", using character escape codes (CEC)
1627 sub quot_cec {
1628 my $cntrl = shift;
1629 my %opts = @_;
1630 my %es = ( # character escape codes, aka escape sequences
1631 "\t" => '\t', # tab (HT)
1632 "\n" => '\n', # line feed (LF)
1633 "\r" => '\r', # carrige return (CR)
1634 "\f" => '\f', # form feed (FF)
1635 "\b" => '\b', # backspace (BS)
1636 "\a" => '\a', # alarm (bell) (BEL)
1637 "\e" => '\e', # escape (ESC)
1638 "\013" => '\v', # vertical tab (VT)
1639 "\000" => '\0', # nul character (NUL)
1640 );
1641 my $chr = ( (exists $es{$cntrl})
1642 ? $es{$cntrl}
1643 : sprintf('\%2x', ord($cntrl)) );
1644 if ($opts{-nohtml}) {
1645 return $chr;
1646 } else {
1647 return "<span class=\"cntrl\">$chr</span>";
1648 }
1649 }
1650
1651 # Alternatively use unicode control pictures codepoints,
1652 # Unicode "printable representation" (PR)
1653 sub quot_upr {
1654 my $cntrl = shift;
1655 my %opts = @_;
1656
1657 my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1658 if ($opts{-nohtml}) {
1659 return $chr;
1660 } else {
1661 return "<span class=\"cntrl\">$chr</span>";
1662 }
1663 }
1664
1665 # git may return quoted and escaped filenames
1666 sub unquote {
1667 my $str = shift;
1668
1669 sub unq {
1670 my $seq = shift;
1671 my %es = ( # character escape codes, aka escape sequences
1672 't' => "\t", # tab (HT, TAB)
1673 'n' => "\n", # newline (NL)
1674 'r' => "\r", # return (CR)
1675 'f' => "\f", # form feed (FF)
1676 'b' => "\b", # backspace (BS)
1677 'a' => "\a", # alarm (bell) (BEL)
1678 'e' => "\e", # escape (ESC)
1679 'v' => "\013", # vertical tab (VT)
1680 );
1681
1682 if ($seq =~ m/^[0-7]{1,3}$/) {
1683 # octal char sequence
1684 return chr(oct($seq));
1685 } elsif (exists $es{$seq}) {
1686 # C escape sequence, aka character escape code
1687 return $es{$seq};
1688 }
1689 # quoted ordinary character
1690 return $seq;
1691 }
1692
1693 if ($str =~ m/^"(.*)"$/) {
1694 # needs unquoting
1695 $str = $1;
1696 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1697 }
1698 return $str;
1699 }
1700
1701 # escape tabs (convert tabs to spaces)
1702 sub untabify {
1703 my $line = shift;
1704
1705 while ((my $pos = index($line, "\t")) != -1) {
1706 if (my $count = (8 - ($pos % 8))) {
1707 my $spaces = ' ' x $count;
1708 $line =~ s/\t/$spaces/;
1709 }
1710 }
1711
1712 return $line;
1713 }
1714
1715 sub project_in_list {
1716 my $project = shift;
1717 my @list = git_get_projects_list();
1718 return @list && scalar(grep { $_->{'path'} eq $project } @list);
1719 }
1720
1721 ## ----------------------------------------------------------------------
1722 ## HTML aware string manipulation
1723
1724 # Try to chop given string on a word boundary between position
1725 # $len and $len+$add_len. If there is no word boundary there,
1726 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1727 # (marking chopped part) would be longer than given string.
1728 sub chop_str {
1729 my $str = shift;
1730 my $len = shift;
1731 my $add_len = shift || 10;
1732 my $where = shift || 'right'; # 'left' | 'center' | 'right'
1733
1734 # Make sure perl knows it is utf8 encoded so we don't
1735 # cut in the middle of a utf8 multibyte char.
1736 $str = to_utf8($str);
1737
1738 # allow only $len chars, but don't cut a word if it would fit in $add_len
1739 # if it doesn't fit, cut it if it's still longer than the dots we would add
1740 # remove chopped character entities entirely
1741
1742 # when chopping in the middle, distribute $len into left and right part
1743 # return early if chopping wouldn't make string shorter
1744 if ($where eq 'center') {
1745 return $str if ($len + 5 >= length($str)); # filler is length 5
1746 $len = int($len/2);
1747 } else {
1748 return $str if ($len + 4 >= length($str)); # filler is length 4
1749 }
1750
1751 # regexps: ending and beginning with word part up to $add_len
1752 my $endre = qr/.{$len}\w{0,$add_len}/;
1753 my $begre = qr/\w{0,$add_len}.{$len}/;
1754
1755 if ($where eq 'left') {
1756 $str =~ m/^(.*?)($begre)$/;
1757 my ($lead, $body) = ($1, $2);
1758 if (length($lead) > 4) {
1759 $lead = " ...";
1760 }
1761 return "$lead$body";
1762
1763 } elsif ($where eq 'center') {
1764 $str =~ m/^($endre)(.*)$/;
1765 my ($left, $str) = ($1, $2);
1766 $str =~ m/^(.*?)($begre)$/;
1767 my ($mid, $right) = ($1, $2);
1768 if (length($mid) > 5) {
1769 $mid = " ... ";
1770 }
1771 return "$left$mid$right";
1772
1773 } else {
1774 $str =~ m/^($endre)(.*)$/;
1775 my $body = $1;
1776 my $tail = $2;
1777 if (length($tail) > 4) {
1778 $tail = "... ";
1779 }
1780 return "$body$tail";
1781 }
1782 }
1783
1784 # takes the same arguments as chop_str, but also wraps a <span> around the
1785 # result with a title attribute if it does get chopped. Additionally, the
1786 # string is HTML-escaped.
1787 sub chop_and_escape_str {
1788 my ($str) = @_;
1789
1790 my $chopped = chop_str(@_);
1791 $str = to_utf8($str);
1792 if ($chopped eq $str) {
1793 return esc_html($chopped);
1794 } else {
1795 $str =~ s/[[:cntrl:]]/?/g;
1796 return $cgi->span({-title=>$str}, esc_html($chopped));
1797 }
1798 }
1799
1800 # Highlight selected fragments of string, using given CSS class,
1801 # and escape HTML. It is assumed that fragments do not overlap.
1802 # Regions are passed as list of pairs (array references).
1803 #
1804 # Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
1805 # '<span class="mark">foo</span>bar'
1806 sub esc_html_hl_regions {
1807 my ($str, $css_class, @sel) = @_;
1808 my %opts = grep { ref($_) ne 'ARRAY' } @sel;
1809 @sel = grep { ref($_) eq 'ARRAY' } @sel;
1810 return esc_html($str, %opts) unless @sel;
1811
1812 my $out = '';
1813 my $pos = 0;
1814
1815 for my $s (@sel) {
1816 my ($begin, $end) = @$s;
1817
1818 # Don't create empty <span> elements.
1819 next if $end <= $begin;
1820
1821 my $escaped = esc_html(substr($str, $begin, $end - $begin),
1822 %opts);
1823
1824 $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
1825 if ($begin - $pos > 0);
1826 $out .= $cgi->span({-class => $css_class}, $escaped);
1827
1828 $pos = $end;
1829 }
1830 $out .= esc_html(substr($str, $pos), %opts)
1831 if ($pos < length($str));
1832
1833 return $out;
1834 }
1835
1836 # return positions of beginning and end of each match
1837 sub matchpos_list {
1838 my ($str, $regexp) = @_;
1839 return unless (defined $str && defined $regexp);
1840
1841 my @matches;
1842 while ($str =~ /$regexp/g) {
1843 push @matches, [$-[0], $+[0]];
1844 }
1845 return @matches;
1846 }
1847
1848 # highlight match (if any), and escape HTML
1849 sub esc_html_match_hl {
1850 my ($str, $regexp) = @_;
1851 return esc_html($str) unless defined $regexp;
1852
1853 my @matches = matchpos_list($str, $regexp);
1854 return esc_html($str) unless @matches;
1855
1856 return esc_html_hl_regions($str, 'match', @matches);
1857 }
1858
1859
1860 # highlight match (if any) of shortened string, and escape HTML
1861 sub esc_html_match_hl_chopped {
1862 my ($str, $chopped, $regexp) = @_;
1863 return esc_html_match_hl($str, $regexp) unless defined $chopped;
1864
1865 my @matches = matchpos_list($str, $regexp);
1866 return esc_html($chopped) unless @matches;
1867
1868 # filter matches so that we mark chopped string
1869 my $tail = "... "; # see chop_str
1870 unless ($chopped =~ s/\Q$tail\E$//) {
1871 $tail = '';
1872 }
1873 my $chop_len = length($chopped);
1874 my $tail_len = length($tail);
1875 my @filtered;
1876
1877 for my $m (@matches) {
1878 if ($m->[0] > $chop_len) {
1879 push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
1880 last;
1881 } elsif ($m->[1] > $chop_len) {
1882 push @filtered, [ $m->[0], $chop_len + $tail_len ];
1883 last;
1884 }
1885 push @filtered, $m;
1886 }
1887
1888 return esc_html_hl_regions($chopped . $tail, 'match', @filtered);
1889 }
1890
1891 ## ----------------------------------------------------------------------
1892 ## functions returning short strings
1893
1894 # CSS class for given age value (in seconds)
1895 sub age_class {
1896 my $age = shift;
1897
1898 if (!defined $age) {
1899 return "noage";
1900 } elsif ($age < 60*60*2) {
1901 return "age0";
1902 } elsif ($age < 60*60*24*2) {
1903 return "age1";
1904 } else {
1905 return "age2";
1906 }
1907 }
1908
1909 # convert age in seconds to "nn units ago" string
1910 sub age_string {
1911 my $age = shift;
1912 my $age_str;
1913
1914 if ($age > 60*60*24*365*2) {
1915 $age_str = (int $age/60/60/24/365);
1916 $age_str .= " years ago";
1917 } elsif ($age > 60*60*24*(365/12)*2) {
1918 $age_str = int $age/60/60/24/(365/12);
1919 $age_str .= " months ago";
1920 } elsif ($age > 60*60*24*7*2) {
1921 $age_str = int $age/60/60/24/7;
1922 $age_str .= " weeks ago";
1923 } elsif ($age > 60*60*24*2) {
1924 $age_str = int $age/60/60/24;
1925 $age_str .= " days ago";
1926 } elsif ($age > 60*60*2) {
1927 $age_str = int $age/60/60;
1928 $age_str .= " hours ago";
1929 } elsif ($age > 60*2) {
1930 $age_str = int $age/60;
1931 $age_str .= " min ago";
1932 } elsif ($age > 2) {
1933 $age_str = int $age;
1934 $age_str .= " sec ago";
1935 } else {
1936 $age_str .= " right now";
1937 }
1938 return $age_str;
1939 }
1940
1941 use constant {
1942 S_IFINVALID => 0030000,
1943 S_IFGITLINK => 0160000,
1944 };
1945
1946 # submodule/subproject, a commit object reference
1947 sub S_ISGITLINK {
1948 my $mode = shift;
1949
1950 return (($mode & S_IFMT) == S_IFGITLINK)
1951 }
1952
1953 # convert file mode in octal to symbolic file mode string
1954 sub mode_str {
1955 my $mode = oct shift;
1956
1957 if (S_ISGITLINK($mode)) {
1958 return 'm---------';
1959 } elsif (S_ISDIR($mode & S_IFMT)) {
1960 return 'drwxr-xr-x';
1961 } elsif (S_ISLNK($mode)) {
1962 return 'lrwxrwxrwx';
1963 } elsif (S_ISREG($mode)) {
1964 # git cares only about the executable bit
1965 if ($mode & S_IXUSR) {
1966 return '-rwxr-xr-x';
1967 } else {
1968 return '-rw-r--r--';
1969 };
1970 } else {
1971 return '----------';
1972 }
1973 }
1974
1975 # convert file mode in octal to file type string
1976 sub file_type {
1977 my $mode = shift;
1978
1979 if ($mode !~ m/^[0-7]+$/) {
1980 return $mode;
1981 } else {
1982 $mode = oct $mode;
1983 }
1984
1985 if (S_ISGITLINK($mode)) {
1986 return "submodule";
1987 } elsif (S_ISDIR($mode & S_IFMT)) {
1988 return "directory";
1989 } elsif (S_ISLNK($mode)) {
1990 return "symlink";
1991 } elsif (S_ISREG($mode)) {
1992 return "file";
1993 } else {
1994 return "unknown";
1995 }
1996 }
1997
1998 # convert file mode in octal to file type description string
1999 sub file_type_long {
2000 my $mode = shift;
2001
2002 if ($mode !~ m/^[0-7]+$/) {
2003 return $mode;
2004 } else {
2005 $mode = oct $mode;
2006 }
2007
2008 if (S_ISGITLINK($mode)) {
2009 return "submodule";
2010 } elsif (S_ISDIR($mode & S_IFMT)) {
2011 return "directory";
2012 } elsif (S_ISLNK($mode)) {
2013 return "symlink";
2014 } elsif (S_ISREG($mode)) {
2015 if ($mode & S_IXUSR) {
2016 return "executable";
2017 } else {
2018 return "file";
2019 };
2020 } else {
2021 return "unknown";
2022 }
2023 }
2024
2025
2026 ## ----------------------------------------------------------------------
2027 ## functions returning short HTML fragments, or transforming HTML fragments
2028 ## which don't belong to other sections
2029
2030 # format line of commit message.
2031 sub format_log_line_html {
2032 my $line = shift;
2033
2034 $line = esc_html($line, -nbsp=>1);
2035 $line =~ s{\b([0-9a-fA-F]{8,40})\b}{
2036 $cgi->a({-href => href(action=>"object", hash=>$1),
2037 -class => "text"}, $1);
2038 }eg;
2039
2040 return $line;
2041 }
2042
2043 # format marker of refs pointing to given object
2044
2045 # the destination action is chosen based on object type and current context:
2046 # - for annotated tags, we choose the tag view unless it's the current view
2047 # already, in which case we go to shortlog view
2048 # - for other refs, we keep the current view if we're in history, shortlog or
2049 # log view, and select shortlog otherwise
2050 sub format_ref_marker {
2051 my ($refs, $id) = @_;
2052 my $markers = '';
2053
2054 if (defined $refs->{$id}) {
2055 foreach my $ref (@{$refs->{$id}}) {
2056 # this code exploits the fact that non-lightweight tags are the
2057 # only indirect objects, and that they are the only objects for which
2058 # we want to use tag instead of shortlog as action
2059 my ($type, $name) = qw();
2060 my $indirect = ($ref =~ s/\^\{\}$//);
2061 # e.g. tags/v2.6.11 or heads/next
2062 if ($ref =~ m!^(.*?)s?/(.*)$!) {
2063 $type = $1;
2064 $name = $2;
2065 } else {
2066 $type = "ref";
2067 $name = $ref;
2068 }
2069
2070 my $class = $type;
2071 $class .= " indirect" if $indirect;
2072
2073 my $dest_action = "shortlog";
2074
2075 if ($indirect) {
2076 $dest_action = "tag" unless $action eq "tag";
2077 } elsif ($action =~ /^(history|(short)?log)$/) {
2078 $dest_action = $action;
2079 }
2080
2081 my $dest = "";
2082 $dest .= "refs/" unless $ref =~ m!^refs/!;
2083 $dest .= $ref;
2084
2085 my $link = $cgi->a({
2086 -href => href(
2087 action=>$dest_action,
2088 hash=>$dest
2089 )}, $name);
2090
2091 $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
2092 $link . "</span>";
2093 }
2094 }
2095
2096 if ($markers) {
2097 return ' <span class="refs">'. $markers . '</span>';
2098 } else {
2099 return "";
2100 }
2101 }
2102
2103 # format, perhaps shortened and with markers, title line
2104 sub format_subject_html {
2105 my ($long, $short, $href, $extra) = @_;
2106 $extra = '' unless defined($extra);
2107
2108 if (length($short) < length($long)) {
2109 $long =~ s/[[:cntrl:]]/?/g;
2110 return $cgi->a({-href => $href, -class => "list subject",
2111 -title => to_utf8($long)},
2112 esc_html($short)) . $extra;
2113 } else {
2114 return $cgi->a({-href => $href, -class => "list subject"},
2115 esc_html($long)) . $extra;
2116 }
2117 }
2118
2119 # Rather than recomputing the url for an email multiple times, we cache it
2120 # after the first hit. This gives a visible benefit in views where the avatar
2121 # for the same email is used repeatedly (e.g. shortlog).
2122 # The cache is shared by all avatar engines (currently gravatar only), which
2123 # are free to use it as preferred. Since only one avatar engine is used for any
2124 # given page, there's no risk for cache conflicts.
2125 our %avatar_cache = ();
2126
2127 # Compute the picon url for a given email, by using the picon search service over at
2128 # http://www.cs.indiana.edu/picons/search.html
2129 sub picon_url {
2130 my $email = lc shift;
2131 if (!$avatar_cache{$email}) {
2132 my ($user, $domain) = split('@', $email);
2133 $avatar_cache{$email} =
2134 "//www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
2135 "$domain/$user/" .
2136 "users+domains+unknown/up/single";
2137 }
2138 return $avatar_cache{$email};
2139 }
2140
2141 # Compute the gravatar url for a given email, if it's not in the cache already.
2142 # Gravatar stores only the part of the URL before the size, since that's the
2143 # one computationally more expensive. This also allows reuse of the cache for
2144 # different sizes (for this particular engine).
2145 sub gravatar_url {
2146 my $email = lc shift;
2147 my $size = shift;
2148 $avatar_cache{$email} ||=
2149 "//www.gravatar.com/avatar/" .
2150 Digest::MD5::md5_hex($email) . "?s=";
2151 return $avatar_cache{$email} . $size;
2152 }
2153
2154 # Insert an avatar for the given $email at the given $size if the feature
2155 # is enabled.
2156 sub git_get_avatar {
2157 my ($email, %opts) = @_;
2158 my $pre_white = ($opts{-pad_before} ? "&nbsp;" : "");
2159 my $post_white = ($opts{-pad_after} ? "&nbsp;" : "");
2160 $opts{-size} ||= 'default';
2161 my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
2162 my $url = "";
2163 if ($git_avatar eq 'gravatar') {
2164 $url = gravatar_url($email, $size);
2165 } elsif ($git_avatar eq 'picon') {
2166 $url = picon_url($email);
2167 }
2168 # Other providers can be added by extending the if chain, defining $url
2169 # as needed. If no variant puts something in $url, we assume avatars
2170 # are completely disabled/unavailable.
2171 if ($url) {
2172 return $pre_white .
2173 "<img width=\"$size\" " .
2174 "class=\"avatar\" " .
2175 "src=\"".esc_url($url)."\" " .
2176 "alt=\"\" " .
2177 "/>" . $post_white;
2178 } else {
2179 return "";
2180 }
2181 }
2182
2183 sub format_search_author {
2184 my ($author, $searchtype, $displaytext) = @_;
2185 my $have_search = gitweb_check_feature('search');
2186
2187 if ($have_search) {
2188 my $performed = "";
2189 if ($searchtype eq 'author') {
2190 $performed = "authored";
2191 } elsif ($searchtype eq 'committer') {
2192 $performed = "committed";
2193 }
2194
2195 return $cgi->a({-href => href(action=>"search", hash=>$hash,
2196 searchtext=>$author,
2197 searchtype=>$searchtype), class=>"list",
2198 title=>"Search for commits $performed by $author"},
2199 $displaytext);
2200
2201 } else {
2202 return $displaytext;
2203 }
2204 }
2205
2206 # format the author name of the given commit with the given tag
2207 # the author name is chopped and escaped according to the other
2208 # optional parameters (see chop_str).
2209 sub format_author_html {
2210 my $tag = shift;
2211 my $co = shift;
2212 my $author = chop_and_escape_str($co->{'author_name'}, @_);
2213 return "<$tag class=\"author\">" .
2214 format_search_author($co->{'author_name'}, "author",
2215 git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2216 $author) .
2217 "</$tag>";
2218 }
2219
2220 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2221 sub format_git_diff_header_line {
2222 my $line = shift;
2223 my $diffinfo = shift;
2224 my ($from, $to) = @_;
2225
2226 if ($diffinfo->{'nparents'}) {
2227 # combined diff
2228 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2229 if ($to->{'href'}) {
2230 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2231 esc_path($to->{'file'}));
2232 } else { # file was deleted (no href)
2233 $line .= esc_path($to->{'file'});
2234 }
2235 } else {
2236 # "ordinary" diff
2237 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2238 if ($from->{'href'}) {
2239 $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2240 'a/' . esc_path($from->{'file'}));
2241 } else { # file was added (no href)
2242 $line .= 'a/' . esc_path($from->{'file'});
2243 }
2244 $line .= ' ';
2245 if ($to->{'href'}) {
2246 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2247 'b/' . esc_path($to->{'file'}));
2248 } else { # file was deleted
2249 $line .= 'b/' . esc_path($to->{'file'});
2250 }
2251 }
2252
2253 return "<div class=\"diff header\">$line</div>\n";
2254 }
2255
2256 # format extended diff header line, before patch itself
2257 sub format_extended_diff_header_line {
2258 my $line = shift;
2259 my $diffinfo = shift;
2260 my ($from, $to) = @_;
2261
2262 # match <path>
2263 if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2264 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2265 esc_path($from->{'file'}));
2266 }
2267 if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2268 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2269 esc_path($to->{'file'}));
2270 }
2271 # match single <mode>
2272 if ($line =~ m/\s(\d{6})$/) {
2273 $line .= '<span class="info"> (' .
2274 file_type_long($1) .
2275 ')</span>';
2276 }
2277 # match <hash>
2278 if ($line =~ m/^index [0-9a-fA-F]{40},[0-9a-fA-F]{40}/) {
2279 # can match only for combined diff
2280 $line = 'index ';
2281 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2282 if ($from->{'href'}[$i]) {
2283 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2284 -class=>"hash"},
2285 substr($diffinfo->{'from_id'}[$i],0,7));
2286 } else {
2287 $line .= '0' x 7;
2288 }
2289 # separator
2290 $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2291 }
2292 $line .= '..';
2293 if ($to->{'href'}) {
2294 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2295 substr($diffinfo->{'to_id'},0,7));
2296 } else {
2297 $line .= '0' x 7;
2298 }
2299
2300 } elsif ($line =~ m/^index [0-9a-fA-F]{40}..[0-9a-fA-F]{40}/) {
2301 # can match only for ordinary diff
2302 my ($from_link, $to_link);
2303 if ($from->{'href'}) {
2304 $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2305 substr($diffinfo->{'from_id'},0,7));
2306 } else {
2307 $from_link = '0' x 7;
2308 }
2309 if ($to->{'href'}) {
2310 $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2311 substr($diffinfo->{'to_id'},0,7));
2312 } else {
2313 $to_link = '0' x 7;
2314 }
2315 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2316 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2317 }
2318
2319 return $line . "<br/>\n";
2320 }
2321
2322 # format from-file/to-file diff header
2323 sub format_diff_from_to_header {
2324 my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2325 my $line;
2326 my $result = '';
2327
2328 $line = $from_line;
2329 #assert($line =~ m/^---/) if DEBUG;
2330 # no extra formatting for "^--- /dev/null"
2331 if (! $diffinfo->{'nparents'}) {
2332 # ordinary (single parent) diff
2333 if ($line =~ m!^--- "?a/!) {
2334 if ($from->{'href'}) {
2335 $line = '--- a/' .
2336 $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2337 esc_path($from->{'file'}));
2338 } else {
2339 $line = '--- a/' .
2340 esc_path($from->{'file'});
2341 }
2342 }
2343 $result .= qq!<div class="diff from_file">$line</div>\n!;
2344
2345 } else {
2346 # combined diff (merge commit)
2347 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2348 if ($from->{'href'}[$i]) {
2349 $line = '--- ' .
2350 $cgi->a({-href=>href(action=>"blobdiff",
2351 hash_parent=>$diffinfo->{'from_id'}[$i],
2352 hash_parent_base=>$parents[$i],
2353 file_parent=>$from->{'file'}[$i],
2354 hash=>$diffinfo->{'to_id'},
2355 hash_base=>$hash,
2356 file_name=>$to->{'file'}),
2357 -class=>"path",
2358 -title=>"diff" . ($i+1)},
2359 $i+1) .
2360 '/' .
2361 $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2362 esc_path($from->{'file'}[$i]));
2363 } else {
2364 $line = '--- /dev/null';
2365 }
2366 $result .= qq!<div class="diff from_file">$line</div>\n!;
2367 }
2368 }
2369
2370 $line = $to_line;
2371 #assert($line =~ m/^\+\+\+/) if DEBUG;
2372 # no extra formatting for "^+++ /dev/null"
2373 if ($line =~ m!^\+\+\+ "?b/!) {
2374 if ($to->{'href'}) {
2375 $line = '+++ b/' .
2376 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2377 esc_path($to->{'file'}));
2378 } else {
2379 $line = '+++ b/' .
2380 esc_path($to->{'file'});
2381 }
2382 }
2383 $result .= qq!<div class="diff to_file">$line</div>\n!;
2384
2385 return $result;
2386 }
2387
2388 # create note for patch simplified by combined diff
2389 sub format_diff_cc_simplified {
2390 my ($diffinfo, @parents) = @_;
2391 my $result = '';
2392
2393 $result .= "<div class=\"diff header\">" .
2394 "diff --cc ";
2395 if (!is_deleted($diffinfo)) {
2396 $result .= $cgi->a({-href => href(action=>"blob",
2397 hash_base=>$hash,
2398 hash=>$diffinfo->{'to_id'},
2399 file_name=>$diffinfo->{'to_file'}),
2400 -class => "path"},
2401 esc_path($diffinfo->{'to_file'}));
2402 } else {
2403 $result .= esc_path($diffinfo->{'to_file'});
2404 }
2405 $result .= "</div>\n" . # class="diff header"
2406 "<div class=\"diff nodifferences\">" .
2407 "Simple merge" .
2408 "</div>\n"; # class="diff nodifferences"
2409
2410 return $result;
2411 }
2412
2413 sub diff_line_class {
2414 my ($line, $from, $to) = @_;
2415
2416 # ordinary diff
2417 my $num_sign = 1;
2418 # combined diff
2419 if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2420 $num_sign = scalar @{$from->{'href'}};
2421 }
2422
2423 my @diff_line_classifier = (
2424 { regexp => qr/^\@\@{$num_sign} /, class => "chunk_header"},
2425 { regexp => qr/^\\/, class => "incomplete" },
2426 { regexp => qr/^ {$num_sign}/, class => "ctx" },
2427 # classifier for context must come before classifier add/rem,
2428 # or we would have to use more complicated regexp, for example
2429 # qr/(?= {0,$m}\+)[+ ]{$num_sign}/, where $m = $num_sign - 1;
2430 { regexp => qr/^[+ ]{$num_sign}/, class => "add" },
2431 { regexp => qr/^[- ]{$num_sign}/, class => "rem" },
2432 );
2433 for my $clsfy (@diff_line_classifier) {
2434 return $clsfy->{'class'}
2435 if ($line =~ $clsfy->{'regexp'});
2436 }
2437
2438 # fallback
2439 return "";
2440 }
2441
2442 # assumes that $from and $to are defined and correctly filled,
2443 # and that $line holds a line of chunk header for unified diff
2444 sub format_unidiff_chunk_header {
2445 my ($line, $from, $to) = @_;
2446
2447 my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2448 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2449
2450 $from_lines = 0 unless defined $from_lines;
2451 $to_lines = 0 unless defined $to_lines;
2452
2453 if ($from->{'href'}) {
2454 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2455 -class=>"list"}, $from_text);
2456 }
2457 if ($to->{'href'}) {
2458 $to_text = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2459 -class=>"list"}, $to_text);
2460 }
2461 $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2462 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2463 return $line;
2464 }
2465
2466 # assumes that $from and $to are defined and correctly filled,
2467 # and that $line holds a line of chunk header for combined diff
2468 sub format_cc_diff_chunk_header {
2469 my ($line, $from, $to) = @_;
2470
2471 my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2472 my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2473
2474 @from_text = split(' ', $ranges);
2475 for (my $i = 0; $i < @from_text; ++$i) {
2476 ($from_start[$i], $from_nlines[$i]) =
2477 (split(',', substr($from_text[$i], 1)), 0);
2478 }
2479
2480 $to_text = pop @from_text;
2481 $to_start = pop @from_start;
2482 $to_nlines = pop @from_nlines;
2483
2484 $line = "<span class=\"chunk_info\">$prefix ";
2485 for (my $i = 0; $i < @from_text; ++$i) {
2486 if ($from->{'href'}[$i]) {
2487 $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2488 -class=>"list"}, $from_text[$i]);
2489 } else {
2490 $line .= $from_text[$i];
2491 }
2492 $line .= " ";
2493 }
2494 if ($to->{'href'}) {
2495 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2496 -class=>"list"}, $to_text);
2497 } else {
2498 $line .= $to_text;
2499 }
2500 $line .= " $prefix</span>" .
2501 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2502 return $line;
2503 }
2504
2505 # process patch (diff) line (not to be used for diff headers),
2506 # returning HTML-formatted (but not wrapped) line.
2507 # If the line is passed as a reference, it is treated as HTML and not
2508 # esc_html()'ed.
2509 sub format_diff_line {
2510 my ($line, $diff_class, $from, $to) = @_;
2511
2512 if (ref($line)) {
2513 $line = $$line;
2514 } else {
2515 chomp $line;
2516 $line = untabify($line);
2517
2518 if ($from && $to && $line =~ m/^\@{2} /) {
2519 $line = format_unidiff_chunk_header($line, $from, $to);
2520 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2521 $line = format_cc_diff_chunk_header($line, $from, $to);
2522 } else {
2523 $line = esc_html($line, -nbsp=>1);
2524 }
2525 }
2526
2527 my $diff_classes = "diff";
2528 $diff_classes .= " $diff_class" if ($diff_class);
2529 $line = "<div class=\"$diff_classes\">$line</div>\n";
2530
2531 return $line;
2532 }
2533
2534 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2535 # linked. Pass the hash of the tree/commit to snapshot.
2536 sub format_snapshot_links {
2537 my ($hash) = @_;
2538 my $num_fmts = @snapshot_fmts;
2539 if ($num_fmts > 1) {
2540 # A parenthesized list of links bearing format names.
2541 # e.g. "snapshot (_tar.gz_ _zip_)"
2542 return "snapshot (" . join(' ', map
2543 $cgi->a({
2544 -href => href(
2545 action=>"snapshot",
2546 hash=>$hash,
2547 snapshot_format=>$_
2548 )
2549 }, $known_snapshot_formats{$_}{'display'})
2550 , @snapshot_fmts) . ")";
2551 } elsif ($num_fmts == 1) {
2552 # A single "snapshot" link whose tooltip bears the format name.
2553 # i.e. "_snapshot_"
2554 my ($fmt) = @snapshot_fmts;
2555 return
2556 $cgi->a({
2557 -href => href(
2558 action=>"snapshot",
2559 hash=>$hash,
2560 snapshot_format=>$fmt
2561 ),
2562 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2563 }, "snapshot");
2564 } else { # $num_fmts == 0
2565 return undef;
2566 }
2567 }
2568
2569 ## ......................................................................
2570 ## functions returning values to be passed, perhaps after some
2571 ## transformation, to other functions; e.g. returning arguments to href()
2572
2573 # returns hash to be passed to href to generate gitweb URL
2574 # in -title key it returns description of link
2575 sub get_feed_info {
2576 my $format = shift || 'Atom';
2577 my %res = (action => lc($format));
2578 my $matched_ref = 0;
2579
2580 # feed links are possible only for project views
2581 return unless (defined $project);
2582 # some views should link to OPML, or to generic project feed,
2583 # or don't have specific feed yet (so they should use generic)
2584 return if (!$action || $action =~ /^(?:tags|heads|forks|tag|search)$/x);
2585
2586 my $branch = undef;
2587 # branches refs uses 'refs/' + $get_branch_refs()[x] + '/' prefix
2588 # (fullname) to differentiate from tag links; this also makes
2589 # possible to detect branch links
2590 for my $ref (get_branch_refs()) {
2591 if ((defined $hash_base && $hash_base =~ m!^refs/\Q$ref\E/(.*)$!) ||
2592 (defined $hash && $hash =~ m!^refs/\Q$ref\E/(.*)$!)) {
2593 $branch = $1;
2594 $matched_ref = $ref;
2595 last;
2596 }
2597 }
2598 # find log type for feed description (title)
2599 my $type = 'log';
2600 if (defined $file_name) {
2601 $type = "history of $file_name";
2602 $type .= "/" if ($action eq 'tree');
2603 $type .= " on '$branch'" if (defined $branch);
2604 } else {
2605 $type = "log of $branch" if (defined $branch);
2606 }
2607
2608 $res{-title} = $type;
2609 $res{'hash'} = (defined $branch ? "refs/$matched_ref/$branch" : undef);
2610 $res{'file_name'} = $file_name;
2611
2612 return %res;
2613 }
2614
2615 ## ----------------------------------------------------------------------
2616 ## git utility subroutines, invoking git commands
2617
2618 # returns path to the core git executable and the --git-dir parameter as list
2619 sub git_cmd {
2620 $number_of_git_cmds++;
2621 return $GIT, '--git-dir='.$git_dir;
2622 }
2623
2624 # quote the given arguments for passing them to the shell
2625 # quote_command("command", "arg 1", "arg with ' and ! characters")
2626 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2627 # Try to avoid using this function wherever possible.
2628 sub quote_command {
2629 return join(' ',
2630 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2631 }
2632
2633 # get HEAD ref of given project as hash
2634 sub git_get_head_hash {
2635 return git_get_full_hash(shift, 'HEAD');
2636 }
2637
2638 sub git_get_full_hash {
2639 return git_get_hash(@_);
2640 }
2641
2642 sub git_get_short_hash {
2643 return git_get_hash(@_, '--short=7');
2644 }
2645
2646 sub git_get_hash {
2647 my ($project, $hash, @options) = @_;
2648 my $o_git_dir = $git_dir;
2649 my $retval = undef;
2650 $git_dir = "$projectroot/$project";
2651 if (open my $fd, '-|', git_cmd(), 'rev-parse',
2652 '--verify', '-q', @options, $hash) {
2653 $retval = <$fd>;
2654 chomp $retval if defined $retval;
2655 close $fd;
2656 }
2657 if (defined $o_git_dir) {
2658 $git_dir = $o_git_dir;
2659 }
2660 return $retval;
2661 }
2662
2663 # get type of given object
2664 sub git_get_type {
2665 my $hash = shift;
2666
2667 open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2668 my $type = <$fd>;
2669 close $fd or return;
2670 chomp $type;
2671 return $type;
2672 }
2673
2674 # repository configuration
2675 our $config_file = '';
2676 our %config;
2677
2678 # store multiple values for single key as anonymous array reference
2679 # single values stored directly in the hash, not as [ <value> ]
2680 sub hash_set_multi {
2681 my ($hash, $key, $value) = @_;
2682
2683 if (!exists $hash->{$key}) {
2684 $hash->{$key} = $value;
2685 } elsif (!ref $hash->{$key}) {
2686 $hash->{$key} = [ $hash->{$key}, $value ];
2687 } else {
2688 push @{$hash->{$key}}, $value;
2689 }
2690 }
2691
2692 # return hash of git project configuration
2693 # optionally limited to some section, e.g. 'gitweb'
2694 sub git_parse_project_config {
2695 my $section_regexp = shift;
2696 my %config;
2697
2698 local $/ = "\0";
2699
2700 open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2701 or return;
2702
2703 while (my $keyval = <$fh>) {
2704 chomp $keyval;
2705 my ($key, $value) = split(/\n/, $keyval, 2);
2706
2707 hash_set_multi(\%config, $key, $value)
2708 if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2709 }
2710 close $fh;
2711
2712 return %config;
2713 }
2714
2715 # convert config value to boolean: 'true' or 'false'
2716 # no value, number > 0, 'true' and 'yes' values are true
2717 # rest of values are treated as false (never as error)
2718 sub config_to_bool {
2719 my $val = shift;
2720
2721 return 1 if !defined $val; # section.key
2722
2723 # strip leading and trailing whitespace
2724 $val =~ s/^\s+//;
2725 $val =~ s/\s+$//;
2726
2727 return (($val =~ /^\d+$/ && $val) || # section.key = 1
2728 ($val =~ /^(?:true|yes)$/i)); # section.key = true
2729 }
2730
2731 # convert config value to simple decimal number
2732 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2733 # to be multiplied by 1024, 1048576, or 1073741824
2734 sub config_to_int {
2735 my $val = shift;
2736
2737 # strip leading and trailing whitespace
2738 $val =~ s/^\s+//;
2739 $val =~ s/\s+$//;
2740
2741 if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2742 $unit = lc($unit);
2743 # unknown unit is treated as 1
2744 return $num * ($unit eq 'g' ? 1073741824 :
2745 $unit eq 'm' ? 1048576 :
2746 $unit eq 'k' ? 1024 : 1);
2747 }
2748 return $val;
2749 }
2750
2751 # convert config value to array reference, if needed
2752 sub config_to_multi {
2753 my $val = shift;
2754
2755 return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2756 }
2757
2758 sub git_get_project_config {
2759 my ($key, $type) = @_;
2760
2761 return unless defined $git_dir;
2762
2763 # key sanity check
2764 return unless ($key);
2765 # only subsection, if exists, is case sensitive,
2766 # and not lowercased by 'git config -z -l'
2767 if (my ($hi, $mi, $lo) = ($key =~ /^([^.]*)\.(.*)\.([^.]*)$/)) {
2768 $lo =~ s/_//g;
2769 $key = join(".", lc($hi), $mi, lc($lo));
2770 return if ($lo =~ /\W/ || $hi =~ /\W/);
2771 } else {
2772 $key = lc($key);
2773 $key =~ s/_//g;
2774 return if ($key =~ /\W/);
2775 }
2776 $key =~ s/^gitweb\.//;
2777
2778 # type sanity check
2779 if (defined $type) {
2780 $type =~ s/^--//;
2781 $type = undef
2782 unless ($type eq 'bool' || $type eq 'int');
2783 }
2784
2785 # get config
2786 if (!defined $config_file ||
2787 $config_file ne "$git_dir/config") {
2788 %config = git_parse_project_config('gitweb');
2789 $config_file = "$git_dir/config";
2790 }
2791
2792 # check if config variable (key) exists
2793 return unless exists $config{"gitweb.$key"};
2794
2795 # ensure given type
2796 if (!defined $type) {
2797 return $config{"gitweb.$key"};
2798 } elsif ($type eq 'bool') {
2799 # backward compatibility: 'git config --bool' returns true/false
2800 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2801 } elsif ($type eq 'int') {
2802 return config_to_int($config{"gitweb.$key"});
2803 }
2804 return $config{"gitweb.$key"};
2805 }
2806
2807 # get hash of given path at given ref
2808 sub git_get_hash_by_path {
2809 my $base = shift;
2810 my $path = shift || return undef;
2811 my $type = shift;
2812
2813 $path =~ s,/+$,,;
2814
2815 open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2816 or die_error(500, "Open git-ls-tree failed");
2817 my $line = <$fd>;
2818 close $fd or return undef;
2819
2820 if (!defined $line) {
2821 # there is no tree or hash given by $path at $base
2822 return undef;
2823 }
2824
2825 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
2826 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t/;
2827 if (defined $type && $type ne $2) {
2828 # type doesn't match
2829 return undef;
2830 }
2831 return $3;
2832 }
2833
2834 # get path of entry with given hash at given tree-ish (ref)
2835 # used to get 'from' filename for combined diff (merge commit) for renames
2836 sub git_get_path_by_hash {
2837 my $base = shift || return;
2838 my $hash = shift || return;
2839
2840 local $/ = "\0";
2841
2842 open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2843 or return undef;
2844 while (my $line = <$fd>) {
2845 chomp $line;
2846
2847 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423 gitweb'
2848 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f gitweb/README'
2849 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2850 close $fd;
2851 return $1;
2852 }
2853 }
2854 close $fd;
2855 return undef;
2856 }
2857
2858 ## ......................................................................
2859 ## git utility functions, directly accessing git repository
2860
2861 # get the value of config variable either from file named as the variable
2862 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2863 # configuration variable in the repository config file.
2864 sub git_get_file_or_project_config {
2865 my ($path, $name) = @_;
2866
2867 $git_dir = "$projectroot/$path";
2868 open my $fd, '<', "$git_dir/$name"
2869 or return git_get_project_config($name);
2870 my $conf = <$fd>;
2871 close $fd;
2872 if (defined $conf) {
2873 chomp $conf;
2874 }
2875 return $conf;
2876 }
2877
2878 sub git_get_project_description {
2879 my $path = shift;
2880 return git_get_file_or_project_config($path, 'description');
2881 }
2882
2883 sub git_get_project_category {
2884 my $path = shift;
2885 return git_get_file_or_project_config($path, 'category');
2886 }
2887
2888
2889 # supported formats:
2890 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2891 # - if its contents is a number, use it as tag weight,
2892 # - otherwise add a tag with weight 1
2893 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2894 # the same value multiple times increases tag weight
2895 # * `gitweb.ctag' multi-valued repo config variable
2896 sub git_get_project_ctags {
2897 my $project = shift;
2898 my $ctags = {};
2899
2900 $git_dir = "$projectroot/$project";
2901 if (opendir my $dh, "$git_dir/ctags") {
2902 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2903 foreach my $tagfile (@files) {
2904 open my $ct, '<', $tagfile
2905 or next;
2906 my $val = <$ct>;
2907 chomp $val if $val;
2908 close $ct;
2909
2910 (my $ctag = $tagfile) =~ s#.*/##;
2911 if ($val =~ /^\d+$/) {
2912 $ctags->{$ctag} = $val;
2913 } else {
2914 $ctags->{$ctag} = 1;
2915 }
2916 }
2917 closedir $dh;
2918
2919 } elsif (open my $fh, '<', "$git_dir/ctags") {
2920 while (my $line = <$fh>) {
2921 chomp $line;
2922 $ctags->{$line}++ if $line;
2923 }
2924 close $fh;
2925
2926 } else {
2927 my $taglist = config_to_multi(git_get_project_config('ctag'));
2928 foreach my $tag (@$taglist) {
2929 $ctags->{$tag}++;
2930 }
2931 }
2932
2933 return $ctags;
2934 }
2935
2936 # return hash, where keys are content tags ('ctags'),
2937 # and values are sum of weights of given tag in every project
2938 sub git_gather_all_ctags {
2939 my $projects = shift;
2940 my $ctags = {};
2941
2942 foreach my $p (@$projects) {
2943 foreach my $ct (keys %{$p->{'ctags'}}) {
2944 $ctags->{$ct} += $p->{'ctags'}->{$ct};
2945 }
2946 }
2947
2948 return $ctags;
2949 }
2950
2951 sub git_populate_project_tagcloud {
2952 my $ctags = shift;
2953
2954 # First, merge different-cased tags; tags vote on casing
2955 my %ctags_lc;
2956 foreach (keys %$ctags) {
2957 $ctags_lc{lc $_}->{count} += $ctags->{$_};
2958 if (not $ctags_lc{lc $_}->{topcount}
2959 or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
2960 $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
2961 $ctags_lc{lc $_}->{topname} = $_;
2962 }
2963 }
2964
2965 my $cloud;
2966 my $matched = $input_params{'ctag'};
2967 if (eval { require HTML::TagCloud; 1; }) {
2968 $cloud = HTML::TagCloud->new;
2969 foreach my $ctag (sort keys %ctags_lc) {
2970 # Pad the title with spaces so that the cloud looks
2971 # less crammed.
2972 my $title = esc_html($ctags_lc{$ctag}->{topname});
2973 $title =~ s/ /&nbsp;/g;
2974 $title =~ s/^/&nbsp;/g;
2975 $title =~ s/$/&nbsp;/g;
2976 if (defined $matched && $matched eq $ctag) {
2977 $title = qq(<span class="match">$title</span>);
2978 }
2979 $cloud->add($title, href(project=>undef, ctag=>$ctag),
2980 $ctags_lc{$ctag}->{count});
2981 }
2982 } else {
2983 $cloud = {};
2984 foreach my $ctag (keys %ctags_lc) {
2985 my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
2986 if (defined $matched && $matched eq $ctag) {
2987 $title = qq(<span class="match">$title</span>);
2988 }
2989 $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
2990 $cloud->{$ctag}{ctag} =
2991 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
2992 }
2993 }
2994 return $cloud;
2995 }
2996
2997 sub git_show_project_tagcloud {
2998 my ($cloud, $count) = @_;
2999 if (ref $cloud eq 'HTML::TagCloud') {
3000 return $cloud->html_and_css($count);
3001 } else {
3002 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
3003 return
3004 '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
3005 join (', ', map {
3006 $cloud->{$_}->{'ctag'}
3007 } splice(@tags, 0, $count)) .
3008 '</div>';
3009 }
3010 }
3011
3012 sub git_get_project_url_list {
3013 my $path = shift;
3014
3015 $git_dir = "$projectroot/$path";
3016 open my $fd, '<', "$git_dir/cloneurl"
3017 or return wantarray ?
3018 @{ config_to_multi(git_get_project_config('url')) } :
3019 config_to_multi(git_get_project_config('url'));
3020 my @git_project_url_list = map { chomp; $_ } <$fd>;
3021 close $fd;
3022
3023 return wantarray ? @git_project_url_list : \@git_project_url_list;
3024 }
3025
3026 sub git_get_projects_list {
3027 my $filter = shift || '';
3028 my $paranoid = shift;
3029 my @list;
3030
3031 if (-d $projects_list) {
3032 # search in directory
3033 my $dir = $projects_list;
3034 # remove the trailing "/"
3035 $dir =~ s!/+$!!;
3036 my $pfxlen = length("$dir");
3037 my $pfxdepth = ($dir =~ tr!/!!);
3038 # when filtering, search only given subdirectory
3039 if ($filter && !$paranoid) {
3040 $dir .= "/$filter";
3041 $dir =~ s!/+$!!;
3042 }
3043
3044 File::Find::find({
3045 follow_fast => 1, # follow symbolic links
3046 follow_skip => 2, # ignore duplicates
3047 dangling_symlinks => 0, # ignore dangling symlinks, silently
3048 wanted => sub {
3049 # global variables
3050 our $project_maxdepth;
3051 our $projectroot;
3052 # skip project-list toplevel, if we get it.
3053 return if (m!^[/.]$!);
3054 # only directories can be git repositories
3055 return unless (-d $_);
3056 # don't traverse too deep (Find is super slow on os x)
3057 # $project_maxdepth excludes depth of $projectroot
3058 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
3059 $File::Find::prune = 1;
3060 return;
3061 }
3062
3063 my $path = substr($File::Find::name, $pfxlen + 1);
3064 # paranoidly only filter here
3065 if ($paranoid && $filter && $path !~ m!^\Q$filter\E/!) {
3066 next;
3067 }
3068 # we check related file in $projectroot
3069 if (check_export_ok("$projectroot/$path")) {
3070 push @list, { path => $path };
3071 $File::Find::prune = 1;
3072 }
3073 },
3074 }, "$dir");
3075
3076 } elsif (-f $projects_list) {
3077 # read from file(url-encoded):
3078 # 'git%2Fgit.git Linus+Torvalds'
3079 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3080 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3081 open my $fd, '<', $projects_list or return;
3082 PROJECT:
3083 while (my $line = <$fd>) {
3084 chomp $line;
3085 my ($path, $owner) = split ' ', $line;
3086 $path = unescape($path);
3087 $owner = unescape($owner);
3088 if (!defined $path) {
3089 next;
3090 }
3091 # if $filter is rpovided, check if $path begins with $filter
3092 if ($filter && $path !~ m!^\Q$filter\E/!) {
3093 next;
3094 }
3095 if (check_export_ok("$projectroot/$path")) {
3096 my $pr = {
3097 path => $path
3098 };
3099 if ($owner) {
3100 $pr->{'owner'} = to_utf8($owner);
3101 }
3102 push @list, $pr;
3103 }
3104 }
3105 close $fd;
3106 }
3107 return @list;
3108 }
3109
3110 # written with help of Tree::Trie module (Perl Artistic License, GPL compatibile)
3111 # as side effects it sets 'forks' field to list of forks for forked projects
3112 sub filter_forks_from_projects_list {
3113 my $projects = shift;
3114
3115 my %trie; # prefix tree of directories (path components)
3116 # generate trie out of those directories that might contain forks
3117 foreach my $pr (@$projects) {
3118 my $path = $pr->{'path'};
3119 $path =~ s/\.git$//; # forks of 'repo.git' are in 'repo/' directory
3120 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
3121 next unless ($path); # skip '.git' repository: tests, git-instaweb
3122 next unless (-d "$projectroot/$path"); # containing directory exists
3123 $pr->{'forks'} = []; # there can be 0 or more forks of project
3124
3125 # add to trie
3126 my @dirs = split('/', $path);
3127 # walk the trie, until either runs out of components or out of trie
3128 my $ref = \%trie;
3129 while (scalar @dirs &&
3130 exists($ref->{$dirs[0]})) {
3131 $ref = $ref->{shift @dirs};
3132 }
3133 # create rest of trie structure from rest of components
3134 foreach my $dir (@dirs) {
3135 $ref = $ref->{$dir} = {};
3136 }
3137 # create end marker, store $pr as a data
3138 $ref->{''} = $pr if (!exists $ref->{''});
3139 }
3140
3141 # filter out forks, by finding shortest prefix match for paths
3142 my @filtered;
3143 PROJECT:
3144 foreach my $pr (@$projects) {
3145 # trie lookup
3146 my $ref = \%trie;
3147 DIR:
3148 foreach my $dir (split('/', $pr->{'path'})) {
3149 if (exists $ref->{''}) {
3150 # found [shortest] prefix, is a fork - skip it
3151 push @{$ref->{''}{'forks'}}, $pr;
3152 next PROJECT;
3153 }
3154 if (!exists $ref->{$dir}) {
3155 # not in trie, cannot have prefix, not a fork
3156 push @filtered, $pr;
3157 next PROJECT;
3158 }
3159 # If the dir is there, we just walk one step down the trie.
3160 $ref = $ref->{$dir};
3161 }
3162 # we ran out of trie
3163 # (shouldn't happen: it's either no match, or end marker)
3164 push @filtered, $pr;
3165 }
3166
3167 return @filtered;
3168 }
3169
3170 # note: fill_project_list_info must be run first,
3171 # for 'descr_long' and 'ctags' to be filled
3172 sub search_projects_list {
3173 my ($projlist, %opts) = @_;
3174 my $tagfilter = $opts{'tagfilter'};
3175 my $search_re = $opts{'search_regexp'};
3176
3177 return @$projlist
3178 unless ($tagfilter || $search_re);
3179
3180 # searching projects require filling to be run before it;
3181 fill_project_list_info($projlist,
3182 $tagfilter ? 'ctags' : (),
3183 $search_re ? ('path', 'descr') : ());
3184 my @projects;
3185 PROJECT:
3186 foreach my $pr (@$projlist) {
3187
3188 if ($tagfilter) {
3189 next unless ref($pr->{'ctags'}) eq 'HASH';
3190 next unless
3191 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
3192 }
3193
3194 if ($search_re) {
3195 next unless
3196 $pr->{'path'} =~ /$search_re/ ||
3197 $pr->{'descr_long'} =~ /$search_re/;
3198 }
3199
3200 push @projects, $pr;
3201 }
3202
3203 return @projects;
3204 }
3205
3206 our $gitweb_project_owner = undef;
3207 sub git_get_project_list_from_file {
3208
3209 return if (defined $gitweb_project_owner);
3210
3211 $gitweb_project_owner = {};
3212 # read from file (url-encoded):
3213 # 'git%2Fgit.git Linus+Torvalds'
3214 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3215 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3216 if (-f $projects_list) {
3217 open(my $fd, '<', $projects_list);
3218 while (my $line = <$fd>) {
3219 chomp $line;
3220 my ($pr, $ow) = split ' ', $line;
3221 $pr = unescape($pr);
3222 $ow = unescape($ow);
3223 $gitweb_project_owner->{$pr} = to_utf8($ow);
3224 }
3225 close $fd;
3226 }
3227 }
3228
3229 sub git_get_project_owner {
3230 my $project = shift;
3231 my $owner;
3232
3233 return undef unless $project;
3234 $git_dir = "$projectroot/$project";
3235
3236 if (!defined $gitweb_project_owner) {
3237 git_get_project_list_from_file();
3238 }
3239
3240 if (exists $gitweb_project_owner->{$project}) {
3241 $owner = $gitweb_project_owner->{$project};
3242 }
3243 if (!defined $owner){
3244 $owner = git_get_project_config('owner');
3245 }
3246 if (!defined $owner) {
3247 $owner = get_file_owner("$git_dir");
3248 }
3249
3250 return $owner;
3251 }
3252
3253 sub git_get_last_activity {
3254 my ($path) = @_;
3255 my $fd;
3256
3257 $git_dir = "$projectroot/$path";
3258 open($fd, "-|", git_cmd(), 'for-each-ref',
3259 '--format=%(committer)',
3260 '--sort=-committerdate',
3261 '--count=1',
3262 map { "refs/$_" } get_branch_refs ()) or return;
3263 my $most_recent = <$fd>;
3264 close $fd or return;
3265 if (defined $most_recent &&
3266 $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
3267 my $timestamp = $1;
3268 my $age = time - $timestamp;
3269 return ($age, age_string($age));
3270 }
3271 return (undef, undef);
3272 }
3273
3274 # Implementation note: when a single remote is wanted, we cannot use 'git
3275 # remote show -n' because that command always work (assuming it's a remote URL
3276 # if it's not defined), and we cannot use 'git remote show' because that would
3277 # try to make a network roundtrip. So the only way to find if that particular
3278 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3279 # and when we find what we want.
3280 sub git_get_remotes_list {
3281 my $wanted = shift;
3282 my %remotes = ();
3283
3284 open my $fd, '-|' , git_cmd(), 'remote', '-v';
3285 return unless $fd;
3286 while (my $remote = <$fd>) {
3287 chomp $remote;
3288 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3289 next if $wanted and not $remote eq $wanted;
3290 my ($url, $key) = ($1, $2);
3291
3292 $remotes{$remote} ||= { 'heads' => () };
3293 $remotes{$remote}{$key} = $url;
3294 }
3295 close $fd or return;
3296 return wantarray ? %remotes : \%remotes;
3297 }
3298
3299 # Takes a hash of remotes as first parameter and fills it by adding the
3300 # available remote heads for each of the indicated remotes.
3301 sub fill_remote_heads {
3302 my $remotes = shift;
3303 my @heads = map { "remotes/$_" } keys %$remotes;
3304 my @remoteheads = git_get_heads_list(undef, @heads);
3305 foreach my $remote (keys %$remotes) {
3306 $remotes->{$remote}{'heads'} = [ grep {
3307 $_->{'name'} =~ s!^$remote/!!
3308 } @remoteheads ];
3309 }
3310 }
3311
3312 sub git_get_references {
3313 my $type = shift || "";
3314 my %refs;
3315 # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3316 # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3317 open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3318 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3319 or return;
3320
3321 while (my $line = <$fd>) {
3322 chomp $line;
3323 if ($line =~ m!^([0-9a-fA-F]{40})\srefs/($type.*)$!) {
3324 if (defined $refs{$1}) {
3325 push @{$refs{$1}}, $2;
3326 } else {
3327 $refs{$1} = [ $2 ];
3328 }
3329 }
3330 }
3331 close $fd or return;
3332 return \%refs;
3333 }
3334
3335 sub git_get_rev_name_tags {
3336 my $hash = shift || return undef;
3337
3338 open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3339 or return;
3340 my $name_rev = <$fd>;
3341 close $fd;
3342
3343 if ($name_rev =~ m|^$hash tags/(.*)$|) {
3344 return $1;
3345 } else {
3346 # catches also '$hash undefined' output
3347 return undef;
3348 }
3349 }
3350
3351 ## ----------------------------------------------------------------------
3352 ## parse to hash functions
3353
3354 sub parse_date {
3355 my $epoch = shift;
3356 my $tz = shift || "-0000";
3357
3358 my %date;
3359 my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3360 my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3361 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3362 $date{'hour'} = $hour;
3363 $date{'minute'} = $min;
3364 $date{'mday'} = $mday;
3365 $date{'day'} = $days[$wday];
3366 $date{'month'} = $months[$mon];
3367 $date{'rfc2822'} = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3368 $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3369 $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3370 $mday, $months[$mon], $hour ,$min;
3371 $date{'iso-8601'} = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3372 1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3373
3374 my ($tz_sign, $tz_hour, $tz_min) =
3375 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3376 $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3377 my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3378 ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3379 $date{'hour_local'} = $hour;
3380 $date{'minute_local'} = $min;
3381 $date{'tz_local'} = $tz;
3382 $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3383 1900+$year, $mon+1, $mday,
3384 $hour, $min, $sec, $tz);
3385 return %date;
3386 }
3387
3388 sub parse_tag {
3389 my $tag_id = shift;
3390 my %tag;
3391 my @comment;
3392
3393 open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3394 $tag{'id'} = $tag_id;
3395 while (my $line = <$fd>) {
3396 chomp $line;
3397 if ($line =~ m/^object ([0-9a-fA-F]{40})$/) {
3398 $tag{'object'} = $1;
3399 } elsif ($line =~ m/^type (.+)$/) {
3400 $tag{'type'} = $1;
3401 } elsif ($line =~ m/^tag (.+)$/) {
3402 $tag{'name'} = $1;
3403 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3404 $tag{'author'} = $1;
3405 $tag{'author_epoch'} = $2;
3406 $tag{'author_tz'} = $3;
3407 if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3408 $tag{'author_name'} = $1;
3409 $tag{'author_email'} = $2;
3410 } else {
3411 $tag{'author_name'} = $tag{'author'};
3412 }
3413 } elsif ($line =~ m/--BEGIN/) {
3414 push @comment, $line;
3415 last;
3416 } elsif ($line eq "") {
3417 last;
3418 }
3419 }
3420 push @comment, <$fd>;
3421 $tag{'comment'} = \@comment;
3422 close $fd or return;
3423 if (!defined $tag{'name'}) {
3424 return
3425 };
3426 return %tag
3427 }
3428
3429 sub parse_commit_text {
3430 my ($commit_text, $withparents) = @_;
3431 my @commit_lines = split '\n', $commit_text;
3432 my %co;
3433
3434 pop @commit_lines; # Remove '\0'
3435
3436 if (! @commit_lines) {
3437 return;
3438 }
3439
3440 my $header = shift @commit_lines;
3441 if ($header !~ m/^[0-9a-fA-F]{40}/) {
3442 return;
3443 }
3444 ($co{'id'}, my @parents) = split ' ', $header;
3445 while (my $line = shift @commit_lines) {
3446 last if $line eq "\n";
3447 if ($line =~ m/^tree ([0-9a-fA-F]{40})$/) {
3448 $co{'tree'} = $1;
3449 } elsif ((!defined $withparents) && ($line =~ m/^parent ([0-9a-fA-F]{40})$/)) {
3450 push @parents, $1;
3451 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3452 $co{'author'} = to_utf8($1);
3453 $co{'author_epoch'} = $2;
3454 $co{'author_tz'} = $3;
3455 if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3456 $co{'author_name'} = $1;
3457 $co{'author_email'} = $2;
3458 } else {
3459 $co{'author_name'} = $co{'author'};
3460 }
3461 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3462 $co{'committer'} = to_utf8($1);
3463 $co{'committer_epoch'} = $2;
3464 $co{'committer_tz'} = $3;
3465 if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3466 $co{'committer_name'} = $1;
3467 $co{'committer_email'} = $2;
3468 } else {
3469 $co{'committer_name'} = $co{'committer'};
3470 }
3471 }
3472 }
3473 if (!defined $co{'tree'}) {
3474 return;
3475 };
3476 $co{'parents'} = \@parents;
3477 $co{'parent'} = $parents[0];
3478
3479 foreach my $title (@commit_lines) {
3480 $title =~ s/^ //;
3481 if ($title ne "") {
3482 $co{'title'} = chop_str($title, 80, 5);
3483 # remove leading stuff of merges to make the interesting part visible
3484 if (length($title) > 50) {
3485 $title =~ s/^Automatic //;
3486 $title =~ s/^merge (of|with) /Merge ... /i;
3487 if (length($title) > 50) {
3488 $title =~ s/(http|rsync):\/\///;
3489 }
3490 if (length($title) > 50) {
3491 $title =~ s/(master|www|rsync)\.//;
3492 }
3493 if (length($title) > 50) {
3494 $title =~ s/kernel.org:?//;
3495 }
3496 if (length($title) > 50) {
3497 $title =~ s/\/pub\/scm//;
3498 }
3499 }
3500 $co{'title_short'} = chop_str($title, 50, 5);
3501 last;
3502 }
3503 }
3504 if (! defined $co{'title'} || $co{'title'} eq "") {
3505 $co{'title'} = $co{'title_short'} = '(no commit message)';
3506 }
3507 # remove added spaces
3508 foreach my $line (@commit_lines) {
3509 $line =~ s/^ //;
3510 }
3511 $co{'comment'} = \@commit_lines;
3512
3513 my $age = time - $co{'committer_epoch'};
3514 $co{'age'} = $age;
3515 $co{'age_string'} = age_string($age);
3516 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3517 if ($age > 60*60*24*7*2) {
3518 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3519 $co{'age_string_age'} = $co{'age_string'};
3520 } else {
3521 $co{'age_string_date'} = $co{'age_string'};
3522 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3523 }
3524 return %co;
3525 }
3526
3527 sub parse_commit {
3528 my ($commit_id) = @_;
3529 my %co;
3530
3531 local $/ = "\0";
3532
3533 open my $fd, "-|", git_cmd(), "rev-list",
3534 "--parents",
3535 "--header",
3536 "--max-count=1",
3537 $commit_id,
3538 "--",
3539 or die_error(500, "Open git-rev-list failed");
3540 %co = parse_commit_text(<$fd>, 1);
3541 close $fd;
3542
3543 return %co;
3544 }
3545
3546 sub parse_commits {
3547 my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3548 my @cos;
3549
3550 $maxcount ||= 1;
3551 $skip ||= 0;
3552
3553 local $/ = "\0";
3554
3555 open my $fd, "-|", git_cmd(), "rev-list",
3556 "--header",
3557 @args,
3558 ("--max-count=" . $maxcount),
3559 ("--skip=" . $skip),
3560 @extra_options,
3561 $commit_id,
3562 "--",
3563 ($filename ? ($filename) : ())
3564 or die_error(500, "Open git-rev-list failed");
3565 while (my $line = <$fd>) {
3566 my %co = parse_commit_text($line);
3567 push @cos, \%co;
3568 }
3569 close $fd;
3570
3571 return wantarray ? @cos : \@cos;
3572 }
3573
3574 # parse line of git-diff-tree "raw" output
3575 sub parse_difftree_raw_line {
3576 my $line = shift;
3577 my %res;
3578
3579 # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M ls-files.c'
3580 # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M rev-tree.c'
3581 if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ([0-9a-fA-F]{40}) ([0-9a-fA-F]{40}) (.)([0-9]{0,3})\t(.*)$/) {
3582 $res{'from_mode'} = $1;
3583 $res{'to_mode'} = $2;
3584 $res{'from_id'} = $3;
3585 $res{'to_id'} = $4;
3586 $res{'status'} = $5;
3587 $res{'similarity'} = $6;
3588 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3589 ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3590 } else {
3591 $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3592 }
3593 }
3594 # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3595 # combined diff (for merge commit)
3596 elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:[0-9a-fA-F]{40} )+)([a-zA-Z]+)\t(.*)$//) {
3597 $res{'nparents'} = length($1);
3598 $res{'from_mode'} = [ split(' ', $2) ];
3599 $res{'to_mode'} = pop @{$res{'from_mode'}};
3600 $res{'from_id'} = [ split(' ', $3) ];
3601 $res{'to_id'} = pop @{$res{'from_id'}};
3602 $res{'status'} = [ split('', $4) ];
3603 $res{'to_file'} = unquote($5);
3604 }
3605 # 'c512b523472485aef4fff9e57b229d9d243c967f'
3606 elsif ($line =~ m/^([0-9a-fA-F]{40})$/) {
3607 $res{'commit'} = $1;
3608 }
3609
3610 return wantarray ? %res : \%res;
3611 }
3612
3613 # wrapper: return parsed line of git-diff-tree "raw" output
3614 # (the argument might be raw line, or parsed info)
3615 sub parsed_difftree_line {
3616 my $line_or_ref = shift;
3617
3618 if (ref($line_or_ref) eq "HASH") {
3619 # pre-parsed (or generated by hand)
3620 return $line_or_ref;
3621 } else {
3622 return parse_difftree_raw_line($line_or_ref);
3623 }
3624 }
3625
3626 # parse line of git-ls-tree output
3627 sub parse_ls_tree_line {
3628 my $line = shift;
3629 my %opts = @_;
3630 my %res;
3631
3632 if ($opts{'-l'}) {
3633 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa 16717 panic.c'
3634 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40}) +(-|[0-9]+)\t(.+)$/s;
3635
3636 $res{'mode'} = $1;
3637 $res{'type'} = $2;
3638 $res{'hash'} = $3;
3639 $res{'size'} = $4;
3640 if ($opts{'-z'}) {
3641 $res{'name'} = $5;
3642 } else {
3643 $res{'name'} = unquote($5);
3644 }
3645 } else {
3646 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
3647 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t(.+)$/s;
3648
3649 $res{'mode'} = $1;
3650 $res{'type'} = $2;
3651 $res{'hash'} = $3;
3652 if ($opts{'-z'}) {
3653 $res{'name'} = $4;
3654 } else {
3655 $res{'name'} = unquote($4);
3656 }
3657 }
3658
3659 return wantarray ? %res : \%res;
3660 }
3661
3662 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3663 sub parse_from_to_diffinfo {
3664 my ($diffinfo, $from, $to, @parents) = @_;
3665
3666 if ($diffinfo->{'nparents'}) {
3667 # combined diff
3668 $from->{'file'} = [];
3669 $from->{'href'} = [];
3670 fill_from_file_info($diffinfo, @parents)
3671 unless exists $diffinfo->{'from_file'};
3672 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3673 $from->{'file'}[$i] =
3674 defined $diffinfo->{'from_file'}[$i] ?
3675 $diffinfo->{'from_file'}[$i] :
3676 $diffinfo->{'to_file'};
3677 if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3678 $from->{'href'}[$i] = href(action=>"blob",
3679 hash_base=>$parents[$i],
3680 hash=>$diffinfo->{'from_id'}[$i],
3681 file_name=>$from->{'file'}[$i]);
3682 } else {
3683 $from->{'href'}[$i] = undef;
3684 }
3685 }
3686 } else {
3687 # ordinary (not combined) diff
3688 $from->{'file'} = $diffinfo->{'from_file'};
3689 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3690 $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3691 hash=>$diffinfo->{'from_id'},
3692 file_name=>$from->{'file'});
3693 } else {
3694 delete $from->{'href'};
3695 }
3696 }
3697
3698 $to->{'file'} = $diffinfo->{'to_file'};
3699 if (!is_deleted($diffinfo)) { # file exists in result
3700 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3701 hash=>$diffinfo->{'to_id'},
3702 file_name=>$to->{'file'});
3703 } else {
3704 delete $to->{'href'};
3705 }
3706 }
3707
3708 ## ......................................................................
3709 ## parse to array of hashes functions
3710
3711 sub git_get_heads_list {
3712 my ($limit, @classes) = @_;
3713 @classes = get_branch_refs() unless @classes;
3714 my @patterns = map { "refs/$_" } @classes;
3715 my @headslist;
3716
3717 open my $fd, '-|', git_cmd(), 'for-each-ref',
3718 ($limit ? '--count='.($limit+1) : ()), '--sort=-committerdate',
3719 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3720 @patterns
3721 or return;
3722 while (my $line = <$fd>) {
3723 my %ref_item;
3724
3725 chomp $line;
3726 my ($refinfo, $committerinfo) = split(/\0/, $line);
3727 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3728 my ($committer, $epoch, $tz) =
3729 ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3730 $ref_item{'fullname'} = $name;
3731 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
3732 $name =~ s!^refs/($strip_refs|remotes)/!!;
3733 $ref_item{'name'} = $name;
3734 # for refs neither in 'heads' nor 'remotes' we want to
3735 # show their ref dir
3736 my $ref_dir = (defined $1) ? $1 : '';
3737 if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
3738 $ref_item{'name'} .= ' (' . $ref_dir . ')';
3739 }
3740
3741 $ref_item{'id'} = $hash;
3742 $ref_item{'title'} = $title || '(no commit message)';
3743 $ref_item{'epoch'} = $epoch;
3744 if ($epoch) {
3745 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3746 } else {
3747 $ref_item{'age'} = "unknown";
3748 }
3749
3750 push @headslist, \%ref_item;
3751 }
3752 close $fd;
3753
3754 return wantarray ? @headslist : \@headslist;
3755 }
3756
3757 sub git_get_tags_list {
3758 my $limit = shift;
3759 my @tagslist;
3760
3761 open my $fd, '-|', git_cmd(), 'for-each-ref',
3762 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3763 '--format=%(objectname) %(objecttype) %(refname) '.
3764 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3765 'refs/tags'
3766 or return;
3767 while (my $line = <$fd>) {
3768 my %ref_item;
3769
3770 chomp $line;
3771 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3772 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3773 my ($creator, $epoch, $tz) =
3774 ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3775 $ref_item{'fullname'} = $name;
3776 $name =~ s!^refs/tags/!!;
3777
3778 $ref_item{'type'} = $type;
3779 $ref_item{'id'} = $id;
3780 $ref_item{'name'} = $name;
3781 if ($type eq "tag") {
3782 $ref_item{'subject'} = $title;
3783 $ref_item{'reftype'} = $reftype;
3784 $ref_item{'refid'} = $refid;
3785 } else {
3786 $ref_item{'reftype'} = $type;
3787 $ref_item{'refid'} = $id;
3788 }
3789
3790 if ($type eq "tag" || $type eq "commit") {
3791 $ref_item{'epoch'} = $epoch;
3792 if ($epoch) {
3793 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3794 } else {
3795 $ref_item{'age'} = "unknown";
3796 }
3797 }
3798
3799 push @tagslist, \%ref_item;
3800 }
3801 close $fd;
3802
3803 return wantarray ? @tagslist : \@tagslist;
3804 }
3805
3806 ## ----------------------------------------------------------------------
3807 ## filesystem-related functions
3808
3809 sub get_file_owner {
3810 my $path = shift;
3811
3812 my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3813 my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3814 if (!defined $gcos) {
3815 return undef;
3816 }
3817 my $owner = $gcos;
3818 $owner =~ s/[,;].*$//;
3819 return to_utf8($owner);
3820 }
3821
3822 # assume that file exists
3823 sub insert_file {
3824 my $filename = shift;
3825
3826 open my $fd, '<', $filename;
3827 print map { to_utf8($_) } <$fd>;
3828 close $fd;
3829 }
3830
3831 ## ......................................................................
3832 ## mimetype related functions
3833
3834 sub mimetype_guess_file {
3835 my $filename = shift;
3836 my $mimemap = shift;
3837 -r $mimemap or return undef;
3838
3839 my %mimemap;
3840 open(my $mh, '<', $mimemap) or return undef;
3841 while (<$mh>) {
3842 next if m/^#/; # skip comments
3843 my ($mimetype, @exts) = split(/\s+/);
3844 foreach my $ext (@exts) {
3845 $mimemap{$ext} = $mimetype;
3846 }
3847 }
3848 close($mh);
3849
3850 $filename =~ /\.([^.]*)$/;
3851 return $mimemap{$1};
3852 }
3853
3854 sub mimetype_guess {
3855 my $filename = shift;
3856 my $mime;
3857 $filename =~ /\./ or return undef;
3858
3859 if ($mimetypes_file) {
3860 my $file = $mimetypes_file;
3861 if ($file !~ m!^/!) { # if it is relative path
3862 # it is relative to project
3863 $file = "$projectroot/$project/$file";
3864 }
3865 $mime = mimetype_guess_file($filename, $file);
3866 }
3867 $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3868 return $mime;
3869 }
3870
3871 sub blob_mimetype {
3872 my $fd = shift;
3873 my $filename = shift;
3874
3875 if ($filename) {
3876 my $mime = mimetype_guess($filename);
3877 $mime and return $mime;
3878 }
3879
3880 # just in case
3881 return $default_blob_plain_mimetype unless $fd;
3882
3883 if (-T $fd) {
3884 return 'text/plain';
3885 } elsif (! $filename) {
3886 return 'application/octet-stream';
3887 } elsif ($filename =~ m/\.png$/i) {
3888 return 'image/png';
3889 } elsif ($filename =~ m/\.gif$/i) {
3890 return 'image/gif';
3891 } elsif ($filename =~ m/\.jpe?g$/i) {
3892 return 'image/jpeg';
3893 } else {
3894 return 'application/octet-stream';
3895 }
3896 }
3897
3898 sub blob_contenttype {
3899 my ($fd, $file_name, $type) = @_;
3900
3901 $type ||= blob_mimetype($fd, $file_name);
3902 if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3903 $type .= "; charset=$default_text_plain_charset";
3904 }
3905
3906 return $type;
3907 }
3908
3909 # guess file syntax for syntax highlighting; return undef if no highlighting
3910 # the name of syntax can (in the future) depend on syntax highlighter used
3911 sub guess_file_syntax {
3912 my ($highlight, $mimetype, $file_name) = @_;
3913 return undef unless ($highlight && defined $file_name);
3914 my $basename = basename($file_name, '.in');
3915 return $highlight_basename{$basename}
3916 if exists $highlight_basename{$basename};
3917
3918 $basename =~ /\.([^.]*)$/;
3919 my $ext = $1 or return undef;
3920 return $highlight_ext{$ext}
3921 if exists $highlight_ext{$ext};
3922
3923 return undef;
3924 }
3925
3926 # run highlighter and return FD of its output,
3927 # or return original FD if no highlighting
3928 sub run_highlighter {
3929 my ($fd, $highlight, $syntax) = @_;
3930 return $fd unless ($highlight && defined $syntax);
3931
3932 close $fd;
3933 open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
3934 quote_command($highlight_bin).
3935 " --replace-tabs=8 --fragment --syntax $syntax |"
3936 or die_error(500, "Couldn't open file or run syntax highlighter");
3937 return $fd;
3938 }
3939
3940 ## ======================================================================
3941 ## functions printing HTML: header, footer, error page
3942
3943 sub get_page_title {
3944 my $title = to_utf8($site_name);
3945
3946 unless (defined $project) {
3947 if (defined $project_filter) {
3948 $title .= " - projects in '" . esc_path($project_filter) . "'";
3949 }
3950 return $title;
3951 }
3952 $title .= " - " . to_utf8($project);
3953
3954 return $title unless (defined $action);
3955 $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
3956
3957 return $title unless (defined $file_name);
3958 $title .= " - " . esc_path($file_name);
3959 if ($action eq "tree" && $file_name !~ m|/$|) {
3960 $title .= "/";
3961 }
3962
3963 return $title;
3964 }
3965
3966 sub get_content_type_html {
3967 # require explicit support from the UA if we are to send the page as
3968 # 'application/xhtml+xml', otherwise send it as plain old 'text/html'.
3969 # we have to do this because MSIE sometimes globs '*/*', pretending to
3970 # support xhtml+xml but choking when it gets what it asked for.
3971 if (defined $cgi->http('HTTP_ACCEPT') &&
3972 $cgi->http('HTTP_ACCEPT') =~ m/(,|;|\s|^)application\/xhtml\+xml(,|;|\s|$)/ &&
3973 $cgi->Accept('application/xhtml+xml') != 0) {
3974 return 'application/xhtml+xml';
3975 } else {
3976 return 'text/html';
3977 }
3978 }
3979
3980 sub print_feed_meta {
3981 if (defined $project) {
3982 my %href_params = get_feed_info();
3983 if (!exists $href_params{'-title'}) {
3984 $href_params{'-title'} = 'log';
3985 }
3986
3987 foreach my $format (qw(RSS Atom)) {
3988 my $type = lc($format);
3989 my %link_attr = (
3990 '-rel' => 'alternate',
3991 '-title' => esc_attr("$project - $href_params{'-title'} - $format feed"),
3992 '-type' => "application/$type+xml"
3993 );
3994
3995 $href_params{'extra_options'} = undef;
3996 $href_params{'action'} = $type;
3997 $link_attr{'-href'} = href(%href_params);
3998 print "<link ".
3999 "rel=\"$link_attr{'-rel'}\" ".
4000 "title=\"$link_attr{'-title'}\" ".
4001 "href=\"$link_attr{'-href'}\" ".
4002 "type=\"$link_attr{'-type'}\" ".
4003 "/>\n";
4004
4005 $href_params{'extra_options'} = '--no-merges';
4006 $link_attr{'-href'} = href(%href_params);
4007 $link_attr{'-title'} .= ' (no merges)';
4008 print "<link ".
4009 "rel=\"$link_attr{'-rel'}\" ".
4010 "title=\"$link_attr{'-title'}\" ".
4011 "href=\"$link_attr{'-href'}\" ".
4012 "type=\"$link_attr{'-type'}\" ".
4013 "/>\n";
4014 }
4015
4016 } else {
4017 printf('<link rel="alternate" title="%s projects list" '.
4018 'href="%s" type="text/plain; charset=utf-8" />'."\n",
4019 esc_attr($site_name), href(project=>undef, action=>"project_index"));
4020 printf('<link rel="alternate" title="%s projects feeds" '.
4021 'href="%s" type="text/x-opml" />'."\n",
4022 esc_attr($site_name), href(project=>undef, action=>"opml"));
4023 }
4024 }
4025
4026 sub print_header_links {
4027 my $status = shift;
4028
4029 # print out each stylesheet that exist, providing backwards capability
4030 # for those people who defined $stylesheet in a config file
4031 if (defined $stylesheet) {
4032 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4033 } else {
4034 foreach my $stylesheet (@stylesheets) {
4035 next unless $stylesheet;
4036 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4037 }
4038 }
4039 print_feed_meta()
4040 if ($status eq '200 OK');
4041 if (defined $favicon) {
4042 print qq(<link rel="shortcut icon" href=").esc_url($favicon).qq(" type="image/png" />\n);
4043 }
4044 }
4045
4046 sub print_nav_breadcrumbs_path {
4047 my $dirprefix = undef;
4048 while (my $part = shift) {
4049 $dirprefix .= "/" if defined $dirprefix;
4050 $dirprefix .= $part;
4051 print $cgi->a({-href => href(project => undef,
4052 project_filter => $dirprefix,
4053 action => "project_list")},
4054 esc_html($part)) . " / ";
4055 }
4056 }
4057
4058 sub print_nav_breadcrumbs {
4059 my %opts = @_;
4060
4061 for my $crumb (@extra_breadcrumbs, [ $home_link_str => $home_link ]) {
4062 print $cgi->a({-href => esc_url($crumb->[1])}, $crumb->[0]) . " / ";
4063 }
4064 if (defined $project) {
4065 my @dirname = split '/', $project;
4066 my $projectbasename = pop @dirname;
4067 print_nav_breadcrumbs_path(@dirname);
4068 print $cgi->a({-href => href(action=>"summary")}, esc_html($projectbasename));
4069 if (defined $action) {
4070 my $action_print = $action ;
4071 if (defined $opts{-action_extra}) {
4072 $action_print = $cgi->a({-href => href(action=>$action)},
4073 $action);
4074 }
4075 print " / $action_print";
4076 }
4077 if (defined $opts{-action_extra}) {
4078 print " / $opts{-action_extra}";
4079 }
4080 print "\n";
4081 } elsif (defined $project_filter) {
4082 print_nav_breadcrumbs_path(split '/', $project_filter);
4083 }
4084 }
4085
4086 sub print_search_form {
4087 if (!defined $searchtext) {
4088 $searchtext = "";
4089 }
4090 my $search_hash;
4091 if (defined $hash_base) {
4092 $search_hash = $hash_base;
4093 } elsif (defined $hash) {
4094 $search_hash = $hash;
4095 } else {
4096 $search_hash = "HEAD";
4097 }
4098 my $action = $my_uri;
4099 my $use_pathinfo = gitweb_check_feature('pathinfo');
4100 if ($use_pathinfo) {
4101 $action .= "/".esc_url($project);
4102 }
4103 print $cgi->startform(-method => "get", -action => $action) .
4104 "<div class=\"search\">\n" .
4105 (!$use_pathinfo &&
4106 $cgi->input({-name=>"p", -value=>$project, -type=>"hidden"}) . "\n") .
4107 $cgi->input({-name=>"a", -value=>"search", -type=>"hidden"}) . "\n" .
4108 $cgi->input({-name=>"h", -value=>$search_hash, -type=>"hidden"}) . "\n" .
4109 $cgi->popup_menu(-name => 'st', -default => 'commit',
4110 -values => ['commit', 'grep', 'author', 'committer', 'pickaxe']) .
4111 " " . $cgi->a({-href => href(action=>"search_help"),
4112 -title => "search help" }, "?") . " search:\n",
4113 $cgi->textfield(-name => "s", -value => $searchtext, -override => 1) . "\n" .
4114 "<span title=\"Extended regular expression\">" .
4115 $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
4116 -checked => $search_use_regexp) .
4117 "</span>" .
4118 "</div>" .
4119 $cgi->end_form() . "\n";
4120 }
4121
4122 sub git_header_html {
4123 my $status = shift || "200 OK";
4124 my $expires = shift;
4125 my %opts = @_;
4126
4127 my $title = get_page_title();
4128 my $content_type = get_content_type_html();
4129 print $cgi->header(-type=>$content_type, -charset => 'utf-8',
4130 -status=> $status, -expires => $expires)
4131 unless ($opts{'-no_http_header'});
4132 my $mod_perl_version = $ENV{'MOD_PERL'} ? " $ENV{'MOD_PERL'}" : '';
4133 print <<EOF;
4134 <?xml version="1.0" encoding="utf-8"?>
4135 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
4136 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en-US" lang="en-US">
4137 <!-- git web interface version $version, (C) 2005-2006, Kay Sievers <kay.sievers\@vrfy.org>, Christian Gierke -->
4138 <!-- git core binaries version $git_version -->
4139 <head>
4140 <meta http-equiv="content-type" content="$content_type; charset=utf-8"/>
4141 <meta name="generator" content="gitweb/$version git/$git_version$mod_perl_version"/>
4142 <meta name="robots" content="index, nofollow"/>
4143 <title>$title</title>
4144 EOF
4145 # the stylesheet, favicon etc urls won't work correctly with path_info
4146 # unless we set the appropriate base URL
4147 if ($ENV{'PATH_INFO'}) {
4148 print "<base href=\"".esc_url($base_url)."\" />\n";
4149 }
4150 print_header_links($status);
4151
4152 if (defined $site_html_head_string) {
4153 print to_utf8($site_html_head_string);
4154 }
4155
4156 print "</head>\n" .
4157 "<body>\n";
4158
4159 if (defined $site_header && -f $site_header) {
4160 insert_file($site_header);
4161 }
4162
4163 print "<div class=\"page_header\">\n";
4164 if (defined $logo) {
4165 print $cgi->a({-href => esc_url($logo_url),
4166 -title => $logo_label},
4167 $cgi->img({-src => esc_url($logo),
4168 -width => 72, -height => 27,
4169 -alt => "git",
4170 -class => "logo"}));
4171 }
4172 print_nav_breadcrumbs(%opts);
4173 print "</div>\n";
4174
4175 my $have_search = gitweb_check_feature('search');
4176 if (defined $project && $have_search) {
4177 print_search_form();
4178 }
4179 }
4180
4181 sub git_footer_html {
4182 my $feed_class = 'rss_logo';
4183
4184 print "<div class=\"page_footer\">\n";
4185 if (defined $project) {
4186 my $descr = git_get_project_description($project);
4187 if (defined $descr) {
4188 print "<div class=\"page_footer_text\">" . esc_html($descr) . "</div>\n";
4189 }
4190
4191 my %href_params = get_feed_info();
4192 if (!%href_params) {
4193 $feed_class .= ' generic';
4194 }
4195 $href_params{'-title'} ||= 'log';
4196
4197 foreach my $format (qw(RSS Atom)) {
4198 $href_params{'action'} = lc($format);
4199 print $cgi->a({-href => href(%href_params),
4200 -title => "$href_params{'-title'} $format feed",
4201 -class => $feed_class}, $format)."\n";
4202 }
4203
4204 } else {
4205 print $cgi->a({-href => href(project=>undef, action=>"opml",
4206 project_filter => $project_filter),
4207 -class => $feed_class}, "OPML") . " ";
4208 print $cgi->a({-href => href(project=>undef, action=>"project_index",
4209 project_filter => $project_filter),
4210 -class => $feed_class}, "TXT") . "\n";
4211 }
4212 print "</div>\n"; # class="page_footer"
4213
4214 if (defined $t0 && gitweb_check_feature('timed')) {
4215 print "<div id=\"generating_info\">\n";
4216 print 'This page took '.
4217 '<span id="generating_time" class="time_span">'.
4218 tv_interval($t0, [ gettimeofday() ]).
4219 ' seconds </span>'.
4220 ' and '.
4221 '<span id="generating_cmd">'.
4222 $number_of_git_cmds.
4223 '</span> git commands '.
4224 " to generate.\n";
4225 print "</div>\n"; # class="page_footer"
4226 }
4227
4228 if (defined $site_footer && -f $site_footer) {
4229 insert_file($site_footer);
4230 }
4231
4232 print qq!<script type="text/javascript" src="!.esc_url($javascript).qq!"></script>\n!;
4233 if (defined $action &&
4234 $action eq 'blame_incremental') {
4235 print qq!<script type="text/javascript">\n!.
4236 qq!startBlame("!. href(action=>"blame_data", -replay=>1) .qq!",\n!.
4237 qq! "!. href() .qq!");\n!.
4238 qq!</script>\n!;
4239 } else {
4240 my ($jstimezone, $tz_cookie, $datetime_class) =
4241 gitweb_get_feature('javascript-timezone');
4242
4243 print qq!<script type="text/javascript">\n!.
4244 qq!window.onload = function () {\n!;
4245 if (gitweb_check_feature('javascript-actions')) {
4246 print qq! fixLinks();\n!;
4247 }
4248 if ($jstimezone && $tz_cookie && $datetime_class) {
4249 print qq! var tz_cookie = { name: '$tz_cookie', expires: 14, path: '/' };\n!. # in days
4250 qq! onloadTZSetup('$jstimezone', tz_cookie, '$datetime_class');\n!;
4251 }
4252 print qq!};\n!.
4253 qq!</script>\n!;
4254 }
4255
4256 print "</body>\n" .
4257 "</html>";
4258 }
4259
4260 # die_error(<http_status_code>, <error_message>[, <detailed_html_description>])
4261 # Example: die_error(404, 'Hash not found')
4262 # By convention, use the following status codes (as defined in RFC 2616):
4263 # 400: Invalid or missing CGI parameters, or
4264 # requested object exists but has wrong type.
4265 # 403: Requested feature (like "pickaxe" or "snapshot") not enabled on
4266 # this server or project.
4267 # 404: Requested object/revision/project doesn't exist.
4268 # 500: The server isn't configured properly, or
4269 # an internal error occurred (e.g. failed assertions caused by bugs), or
4270 # an unknown error occurred (e.g. the git binary died unexpectedly).
4271 # 503: The server is currently unavailable (because it is overloaded,
4272 # or down for maintenance). Generally, this is a temporary state.
4273 sub die_error {
4274 my $status = shift || 500;
4275 my $error = esc_html(shift) || "Internal Server Error";
4276 my $extra = shift;
4277 my %opts = @_;
4278
4279 my %http_responses = (
4280 400 => '400 Bad Request',
4281 403 => '403 Forbidden',
4282 404 => '404 Not Found',
4283 500 => '500 Internal Server Error',
4284 503 => '503 Service Unavailable',
4285 );
4286 git_header_html($http_responses{$status}, undef, %opts);
4287 print <<EOF;
4288 <div class="page_body">
4289 <br /><br />
4290 $status - $error
4291 <br />
4292 EOF
4293 if (defined $extra) {
4294 print "<hr />\n" .
4295 "$extra\n";
4296 }
4297 print "</div>\n";
4298
4299 git_footer_html();
4300 goto DONE_GITWEB
4301 unless ($opts{'-error_handler'});
4302 }
4303
4304 ## ----------------------------------------------------------------------
4305 ## functions printing or outputting HTML: navigation
4306
4307 sub git_print_page_nav {
4308 my ($current, $suppress, $head, $treehead, $treebase, $extra) = @_;
4309 $extra = '' if !defined $extra; # pager or formats
4310
4311 my @navs =