Cope with non-ASCII $site_name
[gitweb] / gitweb.cgi
1 #!/usr/bin/perl
2
3 # gitweb - simple web interface to track changes in git repositories
4 #
5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
6 # (C) 2005, Christian Gierke
7 #
8 # This program is licensed under the GPLv2
9
10 use 5.008;
11 use strict;
12 use warnings;
13 # handle ACL in file access tests
14 use filetest 'access';
15 use CGI qw(:standard :escapeHTML -nosticky);
16 use CGI::Util qw(unescape);
17 use CGI::Carp qw(fatalsToBrowser set_message);
18 use Encode;
19 use Fcntl ':mode';
20 use File::Find qw();
21 use File::Basename qw(basename);
22 use Time::HiRes qw(gettimeofday tv_interval);
23 use Digest::MD5 qw(md5_hex);
24
25 binmode STDOUT, ':utf8';
26
27 if (!defined($CGI::VERSION) || $CGI::VERSION < 4.08) {
28 eval 'sub CGI::multi_param { CGI::param(@_) }'
29 }
30
31 our $t0 = [ gettimeofday() ];
32 our $number_of_git_cmds = 0;
33
34 BEGIN {
35 CGI->compile() if $ENV{'MOD_PERL'};
36 }
37
38 our $version = "2.34.1";
39
40 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
41 sub evaluate_uri {
42 our $cgi;
43
44 our $my_url = $cgi->url();
45 our $my_uri = $cgi->url(-absolute => 1);
46
47 # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
48 # needed and used only for URLs with nonempty PATH_INFO
49 our $base_url = $my_url;
50
51 # When the script is used as DirectoryIndex, the URL does not contain the name
52 # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
53 # have to do it ourselves. We make $path_info global because it's also used
54 # later on.
55 #
56 # Another issue with the script being the DirectoryIndex is that the resulting
57 # $my_url data is not the full script URL: this is good, because we want
58 # generated links to keep implying the script name if it wasn't explicitly
59 # indicated in the URL we're handling, but it means that $my_url cannot be used
60 # as base URL.
61 # Therefore, if we needed to strip PATH_INFO, then we know that we have
62 # to build the base URL ourselves:
63 our $path_info = decode_utf8($ENV{"PATH_INFO"});
64 if ($path_info) {
65 # $path_info has already been URL-decoded by the web server, but
66 # $my_url and $my_uri have not. URL-decode them so we can properly
67 # strip $path_info.
68 $my_url = unescape($my_url);
69 $my_uri = unescape($my_uri);
70 if ($my_url =~ s,\Q$path_info\E$,, &&
71 $my_uri =~ s,\Q$path_info\E$,, &&
72 defined $ENV{'SCRIPT_NAME'}) {
73 $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
74 }
75 }
76
77 # target of the home link on top of all pages
78 our $home_link = $my_uri || "/";
79 }
80
81 # core git executable to use
82 # this can just be "git" if your webserver has a sensible PATH
83 our $GIT = "/usr/bin/git";
84
85 # absolute fs-path which will be prepended to the project path
86 #our $projectroot = "/pub/scm";
87 our $projectroot = "/pub/git";
88
89 # fs traversing limit for getting project list
90 # the number is relative to the projectroot
91 our $project_maxdepth = 2007;
92
93 # string of the home link on top of all pages
94 our $home_link_str = "projects";
95
96 # extra breadcrumbs preceding the home link
97 our @extra_breadcrumbs = ();
98
99 # name of your site or organization to appear in page titles
100 # replace this with something more descriptive for clearer bookmarks
101 our $site_name = ""
102 || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
103
104 # html snippet to include in the <head> section of each page
105 our $site_html_head_string = "";
106 # filename of html text to include at top of each page
107 our $site_header = "";
108 # html text to include at home page
109 our $home_text = "indextext.html";
110 # filename of html text to include at bottom of each page
111 our $site_footer = "";
112
113 # URI of stylesheets
114 our @stylesheets = ("static/gitweb.css");
115 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
116 our $stylesheet = undef;
117 # URI of GIT logo (72x27 size)
118 our $logo = "static/git-logo.png";
119 # URI of GIT favicon, assumed to be image/png type
120 our $favicon = "static/git-favicon.png";
121 # URI of gitweb.js (JavaScript code for gitweb)
122 our $javascript = "static/gitweb.js";
123
124 # URI and label (title) of GIT logo link
125 #our $logo_url = "http://www.kernel.org/pub/software/scm/git/docs/";
126 #our $logo_label = "git documentation";
127 our $logo_url = "http://git-scm.com/";
128 our $logo_label = "git homepage";
129
130 # source of projects list
131 our $projects_list = "";
132
133 # the width (in characters) of the projects list "Description" column
134 our $projects_list_description_width = 25;
135
136 # group projects by category on the projects list
137 # (enabled if this variable evaluates to true)
138 our $projects_list_group_categories = 0;
139
140 # default category if none specified
141 # (leave the empty string for no category)
142 our $project_list_default_category = "";
143
144 # default order of projects list
145 # valid values are none, project, descr, owner, and age
146 our $default_projects_order = "project";
147
148 # show repository only if this file exists
149 # (only effective if this variable evaluates to true)
150 our $export_ok = "";
151
152 # don't generate age column on the projects list page
153 our $omit_age_column = 0;
154
155 # don't generate information about owners of repositories
156 our $omit_owner=0;
157
158 # show repository only if this subroutine returns true
159 # when given the path to the project, for example:
160 # sub { return -e "$_[0]/git-daemon-export-ok"; }
161 our $export_auth_hook = undef;
162
163 # only allow viewing of repositories also shown on the overview page
164 our $strict_export = "";
165
166 # list of git base URLs used for URL to where fetch project from,
167 # i.e. full URL is "$git_base_url/$project"
168 our @git_base_url_list = grep { $_ ne '' } ("");
169
170 # default blob_plain mimetype and default charset for text/plain blob
171 our $default_blob_plain_mimetype = 'text/plain';
172 our $default_text_plain_charset = undef;
173
174 # file to use for guessing MIME types before trying /etc/mime.types
175 # (relative to the current git repository)
176 our $mimetypes_file = undef;
177
178 # assume this charset if line contains non-UTF-8 characters;
179 # it should be valid encoding (see Encoding::Supported(3pm) for list),
180 # for which encoding all byte sequences are valid, for example
181 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
182 # could be even 'utf-8' for the old behavior)
183 our $fallback_encoding = 'latin1';
184
185 # rename detection options for git-diff and git-diff-tree
186 # - default is '-M', with the cost proportional to
187 # (number of removed files) * (number of new files).
188 # - more costly is '-C' (which implies '-M'), with the cost proportional to
189 # (number of changed files + number of removed files) * (number of new files)
190 # - even more costly is '-C', '--find-copies-harder' with cost
191 # (number of files in the original tree) * (number of new files)
192 # - one might want to include '-B' option, e.g. '-B', '-M'
193 our @diff_opts = ('-M'); # taken from git_commit
194
195 # Disables features that would allow repository owners to inject script into
196 # the gitweb domain.
197 our $prevent_xss = 0;
198
199 # Path to the highlight executable to use (must be the one from
200 # http://www.andre-simon.de due to assumptions about parameters and output).
201 # Useful if highlight is not installed on your webserver's PATH.
202 # [Default: highlight]
203 our $highlight_bin = "highlight";
204
205 # information about snapshot formats that gitweb is capable of serving
206 our %known_snapshot_formats = (
207 # name => {
208 # 'display' => display name,
209 # 'type' => mime type,
210 # 'suffix' => filename suffix,
211 # 'format' => --format for git-archive,
212 # 'compressor' => [compressor command and arguments]
213 # (array reference, optional)
214 # 'disabled' => boolean (optional)}
215 #
216 'tgz' => {
217 'display' => 'tar.gz',
218 'type' => 'application/x-gzip',
219 'suffix' => '.tar.gz',
220 'format' => 'tar',
221 'compressor' => ['gzip', '-n']},
222
223 'tbz2' => {
224 'display' => 'tar.bz2',
225 'type' => 'application/x-bzip2',
226 'suffix' => '.tar.bz2',
227 'format' => 'tar',
228 'compressor' => ['bzip2']},
229
230 'txz' => {
231 'display' => 'tar.xz',
232 'type' => 'application/x-xz',
233 'suffix' => '.tar.xz',
234 'format' => 'tar',
235 'compressor' => ['xz'],
236 'disabled' => 1},
237
238 'zip' => {
239 'display' => 'zip',
240 'type' => 'application/x-zip',
241 'suffix' => '.zip',
242 'format' => 'zip'},
243 );
244
245 # Aliases so we understand old gitweb.snapshot values in repository
246 # configuration.
247 our %known_snapshot_format_aliases = (
248 'gzip' => 'tgz',
249 'bzip2' => 'tbz2',
250 'xz' => 'txz',
251
252 # backward compatibility: legacy gitweb config support
253 'x-gzip' => undef, 'gz' => undef,
254 'x-bzip2' => undef, 'bz2' => undef,
255 'x-zip' => undef, '' => undef,
256 );
257
258 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
259 # are changed, it may be appropriate to change these values too via
260 # $GITWEB_CONFIG.
261 our %avatar_size = (
262 'default' => 16,
263 'double' => 32
264 );
265
266 # Used to set the maximum load that we will still respond to gitweb queries.
267 # If server load exceed this value then return "503 server busy" error.
268 # If gitweb cannot determined server load, it is taken to be 0.
269 # Leave it undefined (or set to 'undef') to turn off load checking.
270 our $maxload = 300;
271
272 # configuration for 'highlight' (http://www.andre-simon.de/)
273 # match by basename
274 our %highlight_basename = (
275 #'Program' => 'py',
276 #'Library' => 'py',
277 'SConstruct' => 'py', # SCons equivalent of Makefile
278 'Makefile' => 'make',
279 );
280 # match by extension
281 our %highlight_ext = (
282 # main extensions, defining name of syntax;
283 # see files in /usr/share/highlight/langDefs/ directory
284 (map { $_ => $_ } qw(py rb java css js tex bib xml awk bat ini spec tcl sql)),
285 # alternate extensions, see /etc/highlight/filetypes.conf
286 (map { $_ => 'c' } qw(c h)),
287 (map { $_ => 'sh' } qw(sh bash zsh ksh)),
288 (map { $_ => 'cpp' } qw(cpp cxx c++ cc)),
289 (map { $_ => 'php' } qw(php php3 php4 php5 phps)),
290 (map { $_ => 'pl' } qw(pl perl pm)), # perhaps also 'cgi'
291 (map { $_ => 'make'} qw(make mak mk)),
292 (map { $_ => 'xml' } qw(xml xhtml html htm)),
293 );
294
295 # You define site-wide feature defaults here; override them with
296 # $GITWEB_CONFIG as necessary.
297 our %feature = (
298 # feature => {
299 # 'sub' => feature-sub (subroutine),
300 # 'override' => allow-override (boolean),
301 # 'default' => [ default options...] (array reference)}
302 #
303 # if feature is overridable (it means that allow-override has true value),
304 # then feature-sub will be called with default options as parameters;
305 # return value of feature-sub indicates if to enable specified feature
306 #
307 # if there is no 'sub' key (no feature-sub), then feature cannot be
308 # overridden
309 #
310 # use gitweb_get_feature(<feature>) to retrieve the <feature> value
311 # (an array) or gitweb_check_feature(<feature>) to check if <feature>
312 # is enabled
313
314 # Enable the 'blame' blob view, showing the last commit that modified
315 # each line in the file. This can be very CPU-intensive.
316
317 # To enable system wide have in $GITWEB_CONFIG
318 # $feature{'blame'}{'default'} = [1];
319 # To have project specific config enable override in $GITWEB_CONFIG
320 # $feature{'blame'}{'override'} = 1;
321 # and in project config gitweb.blame = 0|1;
322 'blame' => {
323 'sub' => sub { feature_bool('blame', @_) },
324 'override' => 0,
325 'default' => [0]},
326
327 # Enable the 'snapshot' link, providing a compressed archive of any
328 # tree. This can potentially generate high traffic if you have large
329 # project.
330
331 # Value is a list of formats defined in %known_snapshot_formats that
332 # you wish to offer.
333 # To disable system wide have in $GITWEB_CONFIG
334 # $feature{'snapshot'}{'default'} = [];
335 # To have project specific config enable override in $GITWEB_CONFIG
336 # $feature{'snapshot'}{'override'} = 1;
337 # and in project config, a comma-separated list of formats or "none"
338 # to disable. Example: gitweb.snapshot = tbz2,zip;
339 'snapshot' => {
340 'sub' => \&feature_snapshot,
341 'override' => 0,
342 'default' => ['tgz']},
343
344 # Enable text search, which will list the commits which match author,
345 # committer or commit text to a given string. Enabled by default.
346 # Project specific override is not supported.
347 #
348 # Note that this controls all search features, which means that if
349 # it is disabled, then 'grep' and 'pickaxe' search would also be
350 # disabled.
351 'search' => {
352 'override' => 0,
353 'default' => [1]},
354
355 # Enable grep search, which will list the files in currently selected
356 # tree containing the given string. Enabled by default. This can be
357 # potentially CPU-intensive, of course.
358 # Note that you need to have 'search' feature enabled too.
359
360 # To enable system wide have in $GITWEB_CONFIG
361 # $feature{'grep'}{'default'} = [1];
362 # To have project specific config enable override in $GITWEB_CONFIG
363 # $feature{'grep'}{'override'} = 1;
364 # and in project config gitweb.grep = 0|1;
365 'grep' => {
366 'sub' => sub { feature_bool('grep', @_) },
367 'override' => 0,
368 'default' => [1]},
369
370 # Enable the pickaxe search, which will list the commits that modified
371 # a given string in a file. This can be practical and quite faster
372 # alternative to 'blame', but still potentially CPU-intensive.
373 # Note that you need to have 'search' feature enabled too.
374
375 # To enable system wide have in $GITWEB_CONFIG
376 # $feature{'pickaxe'}{'default'} = [1];
377 # To have project specific config enable override in $GITWEB_CONFIG
378 # $feature{'pickaxe'}{'override'} = 1;
379 # and in project config gitweb.pickaxe = 0|1;
380 'pickaxe' => {
381 'sub' => sub { feature_bool('pickaxe', @_) },
382 'override' => 0,
383 'default' => [1]},
384
385 # Enable showing size of blobs in a 'tree' view, in a separate
386 # column, similar to what 'ls -l' does. This cost a bit of IO.
387
388 # To disable system wide have in $GITWEB_CONFIG
389 # $feature{'show-sizes'}{'default'} = [0];
390 # To have project specific config enable override in $GITWEB_CONFIG
391 # $feature{'show-sizes'}{'override'} = 1;
392 # and in project config gitweb.showsizes = 0|1;
393 'show-sizes' => {
394 'sub' => sub { feature_bool('showsizes', @_) },
395 'override' => 0,
396 'default' => [1]},
397
398 # Make gitweb use an alternative format of the URLs which can be
399 # more readable and natural-looking: project name is embedded
400 # directly in the path and the query string contains other
401 # auxiliary information. All gitweb installations recognize
402 # URL in either format; this configures in which formats gitweb
403 # generates links.
404
405 # To enable system wide have in $GITWEB_CONFIG
406 # $feature{'pathinfo'}{'default'} = [1];
407 # Project specific override is not supported.
408
409 # Note that you will need to change the default location of CSS,
410 # favicon, logo and possibly other files to an absolute URL. Also,
411 # if gitweb.cgi serves as your indexfile, you will need to force
412 # $my_uri to contain the script name in your $GITWEB_CONFIG.
413 'pathinfo' => {
414 'override' => 0,
415 'default' => [0]},
416
417 # Make gitweb consider projects in project root subdirectories
418 # to be forks of existing projects. Given project $projname.git,
419 # projects matching $projname/*.git will not be shown in the main
420 # projects list, instead a '+' mark will be added to $projname
421 # there and a 'forks' view will be enabled for the project, listing
422 # all the forks. If project list is taken from a file, forks have
423 # to be listed after the main project.
424
425 # To enable system wide have in $GITWEB_CONFIG
426 # $feature{'forks'}{'default'} = [1];
427 # Project specific override is not supported.
428 'forks' => {
429 'override' => 0,
430 'default' => [0]},
431
432 # Insert custom links to the action bar of all project pages.
433 # This enables you mainly to link to third-party scripts integrating
434 # into gitweb; e.g. git-browser for graphical history representation
435 # or custom web-based repository administration interface.
436
437 # The 'default' value consists of a list of triplets in the form
438 # (label, link, position) where position is the label after which
439 # to insert the link and link is a format string where %n expands
440 # to the project name, %f to the project path within the filesystem,
441 # %h to the current hash (h gitweb parameter) and %b to the current
442 # hash base (hb gitweb parameter); %% expands to %.
443
444 # To enable system wide have in $GITWEB_CONFIG e.g.
445 # $feature{'actions'}{'default'} = [('graphiclog',
446 # '/git-browser/by-commit.html?r=%n', 'summary')];
447 # Project specific override is not supported.
448 'actions' => {
449 'override' => 0,
450 'default' => []},
451
452 # Allow gitweb scan project content tags of project repository,
453 # and display the popular Web 2.0-ish "tag cloud" near the projects
454 # list. Note that this is something COMPLETELY different from the
455 # normal Git tags.
456
457 # gitweb by itself can show existing tags, but it does not handle
458 # tagging itself; you need to do it externally, outside gitweb.
459 # The format is described in git_get_project_ctags() subroutine.
460 # You may want to install the HTML::TagCloud Perl module to get
461 # a pretty tag cloud instead of just a list of tags.
462
463 # To enable system wide have in $GITWEB_CONFIG
464 # $feature{'ctags'}{'default'} = [1];
465 # Project specific override is not supported.
466
467 # In the future whether ctags editing is enabled might depend
468 # on the value, but using 1 should always mean no editing of ctags.
469 'ctags' => {
470 'override' => 0,
471 'default' => [0]},
472
473 # The maximum number of patches in a patchset generated in patch
474 # view. Set this to 0 or undef to disable patch view, or to a
475 # negative number to remove any limit.
476
477 # To disable system wide have in $GITWEB_CONFIG
478 # $feature{'patches'}{'default'} = [0];
479 # To have project specific config enable override in $GITWEB_CONFIG
480 # $feature{'patches'}{'override'} = 1;
481 # and in project config gitweb.patches = 0|n;
482 # where n is the maximum number of patches allowed in a patchset.
483 'patches' => {
484 'sub' => \&feature_patches,
485 'override' => 0,
486 'default' => [16]},
487
488 # Avatar support. When this feature is enabled, views such as
489 # shortlog or commit will display an avatar associated with
490 # the email of the committer(s) and/or author(s).
491
492 # Currently available providers are gravatar and picon.
493 # If an unknown provider is specified, the feature is disabled.
494
495 # Picon currently relies on the indiana.edu database.
496
497 # To enable system wide have in $GITWEB_CONFIG
498 # $feature{'avatar'}{'default'} = ['<provider>'];
499 # where <provider> is either gravatar or picon.
500 # To have project specific config enable override in $GITWEB_CONFIG
501 # $feature{'avatar'}{'override'} = 1;
502 # and in project config gitweb.avatar = <provider>;
503 'avatar' => {
504 'sub' => \&feature_avatar,
505 'override' => 0,
506 'default' => ['']},
507
508 # Enable displaying how much time and how many git commands
509 # it took to generate and display page. Disabled by default.
510 # Project specific override is not supported.
511 'timed' => {
512 'override' => 0,
513 'default' => [0]},
514
515 # Enable turning some links into links to actions which require
516 # JavaScript to run (like 'blame_incremental'). Not enabled by
517 # default. Project specific override is currently not supported.
518 'javascript-actions' => {
519 'override' => 0,
520 'default' => [0]},
521
522 # Syntax highlighting support. This is based on Daniel Svensson's
523 # and Sham Chukoury's work in gitweb-xmms2.git.
524 # It requires the 'highlight' program present in $PATH,
525 # and therefore is disabled by default.
526
527 # To enable system wide have in $GITWEB_CONFIG
528 # $feature{'highlight'}{'default'} = [1];
529
530 'highlight' => {
531 'sub' => sub { feature_bool('highlight', @_) },
532 'override' => 0,
533 'default' => [0]},
534
535 # Enable displaying of remote heads in the heads list
536
537 # To enable system wide have in $GITWEB_CONFIG
538 # $feature{'remote_heads'}{'default'} = [1];
539 # To have project specific config enable override in $GITWEB_CONFIG
540 # $feature{'remote_heads'}{'override'} = 1;
541 # and in project config gitweb.remoteheads = 0|1;
542 'remote_heads' => {
543 'sub' => sub { feature_bool('remote_heads', @_) },
544 'override' => 0,
545 'default' => [0]},
546
547 # Enable showing branches under other refs in addition to heads
548
549 # To set system wide extra branch refs have in $GITWEB_CONFIG
550 # $feature{'extra-branch-refs'}{'default'} = ['dirs', 'of', 'choice'];
551 # To have project specific config enable override in $GITWEB_CONFIG
552 # $feature{'extra-branch-refs'}{'override'} = 1;
553 # and in project config gitweb.extrabranchrefs = dirs of choice
554 # Every directory is separated with whitespace.
555
556 'extra-branch-refs' => {
557 'sub' => \&feature_extra_branch_refs,
558 'override' => 0,
559 'default' => []},
560
561 # Redact e-mail addresses.
562
563 # To enable system wide have in $GITWEB_CONFIG
564 # $feature{'email-privacy'}{'default'} = [1];
565 'email-privacy' => {
566 'sub' => sub { feature_bool('email-privacy', @_) },
567 'override' => 1,
568 'default' => [0]},
569 );
570
571 sub gitweb_get_feature {
572 my ($name) = @_;
573 return unless exists $feature{$name};
574 my ($sub, $override, @defaults) = (
575 $feature{$name}{'sub'},
576 $feature{$name}{'override'},
577 @{$feature{$name}{'default'}});
578 # project specific override is possible only if we have project
579 our $git_dir; # global variable, declared later
580 if (!$override || !defined $git_dir) {
581 return @defaults;
582 }
583 if (!defined $sub) {
584 warn "feature $name is not overridable";
585 return @defaults;
586 }
587 return $sub->(@defaults);
588 }
589
590 # A wrapper to check if a given feature is enabled.
591 # With this, you can say
592 #
593 # my $bool_feat = gitweb_check_feature('bool_feat');
594 # gitweb_check_feature('bool_feat') or somecode;
595 #
596 # instead of
597 #
598 # my ($bool_feat) = gitweb_get_feature('bool_feat');
599 # (gitweb_get_feature('bool_feat'))[0] or somecode;
600 #
601 sub gitweb_check_feature {
602 return (gitweb_get_feature(@_))[0];
603 }
604
605
606 sub feature_bool {
607 my $key = shift;
608 my ($val) = git_get_project_config($key, '--bool');
609
610 if (!defined $val) {
611 return ($_[0]);
612 } elsif ($val eq 'true') {
613 return (1);
614 } elsif ($val eq 'false') {
615 return (0);
616 }
617 }
618
619 sub feature_snapshot {
620 my (@fmts) = @_;
621
622 my ($val) = git_get_project_config('snapshot');
623
624 if ($val) {
625 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
626 }
627
628 return @fmts;
629 }
630
631 sub feature_patches {
632 my @val = (git_get_project_config('patches', '--int'));
633
634 if (@val) {
635 return @val;
636 }
637
638 return ($_[0]);
639 }
640
641 sub feature_avatar {
642 my @val = (git_get_project_config('avatar'));
643
644 return @val ? @val : @_;
645 }
646
647 sub feature_extra_branch_refs {
648 my (@branch_refs) = @_;
649 my $values = git_get_project_config('extrabranchrefs');
650
651 if ($values) {
652 $values = config_to_multi ($values);
653 @branch_refs = ();
654 foreach my $value (@{$values}) {
655 push @branch_refs, split /\s+/, $value;
656 }
657 }
658
659 return @branch_refs;
660 }
661
662 # checking HEAD file with -e is fragile if the repository was
663 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
664 # and then pruned.
665 sub check_head_link {
666 my ($dir) = @_;
667 my $headfile = "$dir/HEAD";
668 return ((-e $headfile) ||
669 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
670 }
671
672 sub check_export_ok {
673 my ($dir) = @_;
674 return (check_head_link($dir) &&
675 (!$export_ok || -e "$dir/$export_ok") &&
676 (!$export_auth_hook || $export_auth_hook->($dir)));
677 }
678
679 # process alternate names for backward compatibility
680 # filter out unsupported (unknown) snapshot formats
681 sub filter_snapshot_fmts {
682 my @fmts = @_;
683
684 @fmts = map {
685 exists $known_snapshot_format_aliases{$_} ?
686 $known_snapshot_format_aliases{$_} : $_} @fmts;
687 @fmts = grep {
688 exists $known_snapshot_formats{$_} &&
689 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
690 }
691
692 sub filter_and_validate_refs {
693 my @refs = @_;
694 my %unique_refs = ();
695
696 foreach my $ref (@refs) {
697 die_error(500, "Invalid ref '$ref' in 'extra-branch-refs' feature") unless (is_valid_ref_format($ref));
698 # 'heads' are added implicitly in get_branch_refs().
699 $unique_refs{$ref} = 1 if ($ref ne 'heads');
700 }
701 return sort keys %unique_refs;
702 }
703
704 # If it is set to code reference, it is code that it is to be run once per
705 # request, allowing updating configurations that change with each request,
706 # while running other code in config file only once.
707 #
708 # Otherwise, if it is false then gitweb would process config file only once;
709 # if it is true then gitweb config would be run for each request.
710 our $per_request_config = 1;
711
712 # read and parse gitweb config file given by its parameter.
713 # returns true on success, false on recoverable error, allowing
714 # to chain this subroutine, using first file that exists.
715 # dies on errors during parsing config file, as it is unrecoverable.
716 sub read_config_file {
717 my $filename = shift;
718 return unless defined $filename;
719 # die if there are errors parsing config file
720 if (-e $filename) {
721 do $filename;
722 die $@ if $@;
723 return 1;
724 }
725 return;
726 }
727
728 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM, $GITWEB_CONFIG_COMMON);
729 sub evaluate_gitweb_config {
730 our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "gitweb_config.perl";
731 our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "/etc/gitweb.conf";
732 our $GITWEB_CONFIG_COMMON = $ENV{'GITWEB_CONFIG_COMMON'} || "/etc/gitweb-common.conf";
733
734 # Protect against duplications of file names, to not read config twice.
735 # Only one of $GITWEB_CONFIG and $GITWEB_CONFIG_SYSTEM is used, so
736 # there possibility of duplication of filename there doesn't matter.
737 $GITWEB_CONFIG = "" if ($GITWEB_CONFIG eq $GITWEB_CONFIG_COMMON);
738 $GITWEB_CONFIG_SYSTEM = "" if ($GITWEB_CONFIG_SYSTEM eq $GITWEB_CONFIG_COMMON);
739
740 # Common system-wide settings for convenience.
741 # Those settings can be overridden by GITWEB_CONFIG or GITWEB_CONFIG_SYSTEM.
742 read_config_file($GITWEB_CONFIG_COMMON);
743
744 # Use first config file that exists. This means use the per-instance
745 # GITWEB_CONFIG if exists, otherwise use GITWEB_SYSTEM_CONFIG.
746 read_config_file($GITWEB_CONFIG) and return;
747 read_config_file($GITWEB_CONFIG_SYSTEM);
748 }
749
750 # Get loadavg of system, to compare against $maxload.
751 # Currently it requires '/proc/loadavg' present to get loadavg;
752 # if it is not present it returns 0, which means no load checking.
753 sub get_loadavg {
754 if( -e '/proc/loadavg' ){
755 open my $fd, '<', '/proc/loadavg'
756 or return 0;
757 my @load = split(/\s+/, scalar <$fd>);
758 close $fd;
759
760 # The first three columns measure CPU and IO utilization of the last one,
761 # five, and 10 minute periods. The fourth column shows the number of
762 # currently running processes and the total number of processes in the m/n
763 # format. The last column displays the last process ID used.
764 return $load[0] || 0;
765 }
766 # additional checks for load average should go here for things that don't export
767 # /proc/loadavg
768
769 return 0;
770 }
771
772 # version of the core git binary
773 our $git_version;
774 sub evaluate_git_version {
775 our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
776 $number_of_git_cmds++;
777 }
778
779 sub check_loadavg {
780 if (defined $maxload && get_loadavg() > $maxload) {
781 die_error(503, "The load average on the server is too high");
782 }
783 }
784
785 # ======================================================================
786 # input validation and dispatch
787
788 # Various hash size-related values.
789 my $sha1_len = 40;
790 my $sha256_extra_len = 24;
791 my $sha256_len = $sha1_len + $sha256_extra_len;
792
793 # A regex matching $len hex characters. $len may be a range (e.g. 7,64).
794 sub oid_nlen_regex {
795 my $len = shift;
796 my $hchr = qr/[0-9a-fA-F]/;
797 return qr/(?:(?:$hchr){$len})/;
798 }
799
800 # A regex matching two sets of $nlen hex characters, prefixed by the literal
801 # string $prefix and with the literal string $infix between them.
802 sub oid_nlen_prefix_infix_regex {
803 my $nlen = shift;
804 my $prefix = shift;
805 my $infix = shift;
806
807 my $rx = oid_nlen_regex($nlen);
808
809 return qr/^\Q$prefix\E$rx\Q$infix\E$rx$/;
810 }
811
812 # A regex matching a valid object ID.
813 our $oid_regex;
814 {
815 my $x = oid_nlen_regex($sha1_len);
816 my $y = oid_nlen_regex($sha256_extra_len);
817 $oid_regex = qr/(?:$x(?:$y)?)/;
818 }
819
820 # input parameters can be collected from a variety of sources (presently, CGI
821 # and PATH_INFO), so we define an %input_params hash that collects them all
822 # together during validation: this allows subsequent uses (e.g. href()) to be
823 # agnostic of the parameter origin
824
825 our %input_params = ();
826
827 # input parameters are stored with the long parameter name as key. This will
828 # also be used in the href subroutine to convert parameters to their CGI
829 # equivalent, and since the href() usage is the most frequent one, we store
830 # the name -> CGI key mapping here, instead of the reverse.
831 #
832 # XXX: Warning: If you touch this, check the search form for updating,
833 # too.
834
835 our @cgi_param_mapping = (
836 project => "p",
837 action => "a",
838 file_name => "f",
839 file_parent => "fp",
840 hash => "h",
841 hash_parent => "hp",
842 hash_base => "hb",
843 hash_parent_base => "hpb",
844 page => "pg",
845 order => "o",
846 searchtext => "s",
847 searchtype => "st",
848 snapshot_format => "sf",
849 extra_options => "opt",
850 search_use_regexp => "sr",
851 ctag => "by_tag",
852 diff_style => "ds",
853 project_filter => "pf",
854 # this must be last entry (for manipulation from JavaScript)
855 javascript => "js"
856 );
857 our %cgi_param_mapping = @cgi_param_mapping;
858
859 # we will also need to know the possible actions, for validation
860 our %actions = (
861 "blame" => \&git_blame,
862 "blame_incremental" => \&git_blame_incremental,
863 "blame_data" => \&git_blame_data,
864 "blobdiff" => \&git_blobdiff,
865 "blobdiff_plain" => \&git_blobdiff_plain,
866 "blob" => \&git_blob,
867 "blob_plain" => \&git_blob_plain,
868 "commitdiff" => \&git_commitdiff,
869 "commitdiff_plain" => \&git_commitdiff_plain,
870 "commit" => \&git_commit,
871 "forks" => \&git_forks,
872 "heads" => \&git_heads,
873 "history" => \&git_history,
874 "log" => \&git_log,
875 "patch" => \&git_patch,
876 "patches" => \&git_patches,
877 "remotes" => \&git_remotes,
878 "atom" => \&git_atom,
879 "search" => \&git_search,
880 "search_help" => \&git_search_help,
881 "shortlog" => \&git_shortlog,
882 "summary" => \&git_summary,
883 "tag" => \&git_tag,
884 "tags" => \&git_tags,
885 "tree" => \&git_tree,
886 "snapshot" => \&git_snapshot,
887 "object" => \&git_object,
888 # those below don't need $project
889 "opml" => \&git_opml,
890 "project_list" => \&git_project_list,
891 "project_index" => \&git_project_index,
892 );
893
894 # finally, we have the hash of allowed extra_options for the commands that
895 # allow them
896 our %allowed_options = (
897 "--no-merges" => [ qw(atom log shortlog history) ],
898 );
899
900 # fill %input_params with the CGI parameters. All values except for 'opt'
901 # should be single values, but opt can be an array. We should probably
902 # build an array of parameters that can be multi-valued, but since for the time
903 # being it's only this one, we just single it out
904 sub evaluate_query_params {
905 our $cgi;
906
907 while (my ($name, $symbol) = each %cgi_param_mapping) {
908 if ($symbol eq 'opt') {
909 $input_params{$name} = [ map { decode_utf8($_) } $cgi->multi_param($symbol) ];
910 } else {
911 $input_params{$name} = decode_utf8($cgi->param($symbol));
912 }
913 }
914 }
915
916 # now read PATH_INFO and update the parameter list for missing parameters
917 sub evaluate_path_info {
918 return if defined $input_params{'project'};
919 return if !$path_info;
920 $path_info =~ s,^/+,,;
921 return if !$path_info;
922
923 # find which part of PATH_INFO is project
924 my $project = $path_info;
925 $project =~ s,/+$,,;
926 while ($project && !check_head_link("$projectroot/$project")) {
927 $project =~ s,/*[^/]*$,,;
928 }
929 return unless $project;
930 $input_params{'project'} = $project;
931
932 # do not change any parameters if an action is given using the query string
933 return if $input_params{'action'};
934 $path_info =~ s,^\Q$project\E/*,,;
935
936 # next, check if we have an action
937 my $action = $path_info;
938 $action =~ s,/.*$,,;
939 if (exists $actions{$action}) {
940 $path_info =~ s,^$action/*,,;
941 $input_params{'action'} = $action;
942 }
943
944 # list of actions that want hash_base instead of hash, but can have no
945 # pathname (f) parameter
946 my @wants_base = (
947 'tree',
948 'history',
949 );
950
951 # we want to catch, among others
952 # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
953 my ($parentrefname, $parentpathname, $refname, $pathname) =
954 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
955
956 # first, analyze the 'current' part
957 if (defined $pathname) {
958 # we got "branch:filename" or "branch:dir/"
959 # we could use git_get_type(branch:pathname), but:
960 # - it needs $git_dir
961 # - it does a git() call
962 # - the convention of terminating directories with a slash
963 # makes it superfluous
964 # - embedding the action in the PATH_INFO would make it even
965 # more superfluous
966 $pathname =~ s,^/+,,;
967 if (!$pathname || substr($pathname, -1) eq "/") {
968 $input_params{'action'} ||= "tree";
969 $pathname =~ s,/$,,;
970 } else {
971 # the default action depends on whether we had parent info
972 # or not
973 if ($parentrefname) {
974 $input_params{'action'} ||= "blobdiff_plain";
975 } else {
976 $input_params{'action'} ||= "blob_plain";
977 }
978 }
979 $input_params{'hash_base'} ||= $refname;
980 $input_params{'file_name'} ||= $pathname;
981 } elsif (defined $refname) {
982 # we got "branch". In this case we have to choose if we have to
983 # set hash or hash_base.
984 #
985 # Most of the actions without a pathname only want hash to be
986 # set, except for the ones specified in @wants_base that want
987 # hash_base instead. It should also be noted that hand-crafted
988 # links having 'history' as an action and no pathname or hash
989 # set will fail, but that happens regardless of PATH_INFO.
990 if (defined $parentrefname) {
991 # if there is parent let the default be 'shortlog' action
992 # (for http://git.example.com/repo.git/A..B links); if there
993 # is no parent, dispatch will detect type of object and set
994 # action appropriately if required (if action is not set)
995 $input_params{'action'} ||= "shortlog";
996 }
997 if ($input_params{'action'} &&
998 grep { $_ eq $input_params{'action'} } @wants_base) {
999 $input_params{'hash_base'} ||= $refname;
1000 } else {
1001 $input_params{'hash'} ||= $refname;
1002 }
1003 }
1004
1005 # next, handle the 'parent' part, if present
1006 if (defined $parentrefname) {
1007 # a missing pathspec defaults to the 'current' filename, allowing e.g.
1008 # someproject/blobdiff/oldrev..newrev:/filename
1009 if ($parentpathname) {
1010 $parentpathname =~ s,^/+,,;
1011 $parentpathname =~ s,/$,,;
1012 $input_params{'file_parent'} ||= $parentpathname;
1013 } else {
1014 $input_params{'file_parent'} ||= $input_params{'file_name'};
1015 }
1016 # we assume that hash_parent_base is wanted if a path was specified,
1017 # or if the action wants hash_base instead of hash
1018 if (defined $input_params{'file_parent'} ||
1019 grep { $_ eq $input_params{'action'} } @wants_base) {
1020 $input_params{'hash_parent_base'} ||= $parentrefname;
1021 } else {
1022 $input_params{'hash_parent'} ||= $parentrefname;
1023 }
1024 }
1025
1026 # for the snapshot action, we allow URLs in the form
1027 # $project/snapshot/$hash.ext
1028 # where .ext determines the snapshot and gets removed from the
1029 # passed $refname to provide the $hash.
1030 #
1031 # To be able to tell that $refname includes the format extension, we
1032 # require the following two conditions to be satisfied:
1033 # - the hash input parameter MUST have been set from the $refname part
1034 # of the URL (i.e. they must be equal)
1035 # - the snapshot format MUST NOT have been defined already (e.g. from
1036 # CGI parameter sf)
1037 # It's also useless to try any matching unless $refname has a dot,
1038 # so we check for that too
1039 if (defined $input_params{'action'} &&
1040 $input_params{'action'} eq 'snapshot' &&
1041 defined $refname && index($refname, '.') != -1 &&
1042 $refname eq $input_params{'hash'} &&
1043 !defined $input_params{'snapshot_format'}) {
1044 # We loop over the known snapshot formats, checking for
1045 # extensions. Allowed extensions are both the defined suffix
1046 # (which includes the initial dot already) and the snapshot
1047 # format key itself, with a prepended dot
1048 while (my ($fmt, $opt) = each %known_snapshot_formats) {
1049 my $hash = $refname;
1050 unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
1051 next;
1052 }
1053 my $sfx = $1;
1054 # a valid suffix was found, so set the snapshot format
1055 # and reset the hash parameter
1056 $input_params{'snapshot_format'} = $fmt;
1057 $input_params{'hash'} = $hash;
1058 # we also set the format suffix to the one requested
1059 # in the URL: this way a request for e.g. .tgz returns
1060 # a .tgz instead of a .tar.gz
1061 $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
1062 last;
1063 }
1064 }
1065 }
1066
1067 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
1068 $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
1069 $searchtext, $search_regexp, $project_filter);
1070 sub evaluate_and_validate_params {
1071 our $action = $input_params{'action'};
1072 if (defined $action) {
1073 if (!is_valid_action($action)) {
1074 die_error(400, "Invalid action parameter");
1075 }
1076 }
1077
1078 # parameters which are pathnames
1079 our $project = $input_params{'project'};
1080 if (defined $project) {
1081 if (!is_valid_project($project)) {
1082 undef $project;
1083 die_error(404, "No such project");
1084 }
1085 }
1086
1087 our $project_filter = $input_params{'project_filter'};
1088 if (defined $project_filter) {
1089 if (!is_valid_pathname($project_filter)) {
1090 die_error(404, "Invalid project_filter parameter");
1091 }
1092 }
1093
1094 our $file_name = $input_params{'file_name'};
1095 if (defined $file_name) {
1096 if (!is_valid_pathname($file_name)) {
1097 die_error(400, "Invalid file parameter");
1098 }
1099 }
1100
1101 our $file_parent = $input_params{'file_parent'};
1102 if (defined $file_parent) {
1103 if (!is_valid_pathname($file_parent)) {
1104 die_error(400, "Invalid file parent parameter");
1105 }
1106 }
1107
1108 # parameters which are refnames
1109 our $hash = $input_params{'hash'};
1110 if (defined $hash) {
1111 if (!is_valid_refname($hash)) {
1112 die_error(400, "Invalid hash parameter");
1113 }
1114 }
1115
1116 our $hash_parent = $input_params{'hash_parent'};
1117 if (defined $hash_parent) {
1118 if (!is_valid_refname($hash_parent)) {
1119 die_error(400, "Invalid hash parent parameter");
1120 }
1121 }
1122
1123 our $hash_base = $input_params{'hash_base'};
1124 if (defined $hash_base) {
1125 if (!is_valid_refname($hash_base)) {
1126 die_error(400, "Invalid hash base parameter");
1127 }
1128 }
1129
1130 our @extra_options = @{$input_params{'extra_options'}};
1131 # @extra_options is always defined, since it can only be (currently) set from
1132 # CGI, and $cgi->param() returns the empty array in array context if the param
1133 # is not set
1134 foreach my $opt (@extra_options) {
1135 if (not exists $allowed_options{$opt}) {
1136 die_error(400, "Invalid option parameter");
1137 }
1138 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1139 die_error(400, "Invalid option parameter for this action");
1140 }
1141 }
1142
1143 our $hash_parent_base = $input_params{'hash_parent_base'};
1144 if (defined $hash_parent_base) {
1145 if (!is_valid_refname($hash_parent_base)) {
1146 die_error(400, "Invalid hash parent base parameter");
1147 }
1148 }
1149
1150 # other parameters
1151 our $page = $input_params{'page'};
1152 if (defined $page) {
1153 if ($page =~ m/[^0-9]/) {
1154 die_error(400, "Invalid page parameter");
1155 }
1156 }
1157
1158 our $searchtype = $input_params{'searchtype'};
1159 if (defined $searchtype) {
1160 if ($searchtype =~ m/[^a-z]/) {
1161 die_error(400, "Invalid searchtype parameter");
1162 }
1163 }
1164
1165 our $search_use_regexp = $input_params{'search_use_regexp'};
1166
1167 our $searchtext = $input_params{'searchtext'};
1168 our $search_regexp = undef;
1169 if (defined $searchtext) {
1170 if (length($searchtext) < 2) {
1171 die_error(403, "At least two characters are required for search parameter");
1172 }
1173 if ($search_use_regexp) {
1174 $search_regexp = $searchtext;
1175 if (!eval { qr/$search_regexp/; 1; }) {
1176 (my $error = $@) =~ s/ at \S+ line \d+.*\n?//;
1177 die_error(400, "Invalid search regexp '$search_regexp'",
1178 esc_html($error));
1179 }
1180 } else {
1181 $search_regexp = quotemeta $searchtext;
1182 }
1183 }
1184 }
1185
1186 # path to the current git repository
1187 our $git_dir;
1188 sub evaluate_git_dir {
1189 our $git_dir = "$projectroot/$project" if $project;
1190 }
1191
1192 our (@snapshot_fmts, $git_avatar, @extra_branch_refs);
1193 sub configure_gitweb_features {
1194 # list of supported snapshot formats
1195 our @snapshot_fmts = gitweb_get_feature('snapshot');
1196 @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1197
1198 our ($git_avatar) = gitweb_get_feature('avatar');
1199 $git_avatar = '' unless $git_avatar =~ /^(?:gravatar|picon)$/s;
1200
1201 our @extra_branch_refs = gitweb_get_feature('extra-branch-refs');
1202 @extra_branch_refs = filter_and_validate_refs (@extra_branch_refs);
1203 }
1204
1205 sub get_branch_refs {
1206 return ('heads', @extra_branch_refs);
1207 }
1208
1209 # custom error handler: 'die <message>' is Internal Server Error
1210 sub handle_errors_html {
1211 my $msg = shift; # it is already HTML escaped
1212
1213 # to avoid infinite loop where error occurs in die_error,
1214 # change handler to default handler, disabling handle_errors_html
1215 set_message("Error occurred when inside die_error:\n$msg");
1216
1217 # you cannot jump out of die_error when called as error handler;
1218 # the subroutine set via CGI::Carp::set_message is called _after_
1219 # HTTP headers are already written, so it cannot write them itself
1220 die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1221 }
1222 set_message(\&handle_errors_html);
1223
1224 # dispatch
1225 sub dispatch {
1226 if (!defined $action) {
1227 if (defined $hash) {
1228 $action = git_get_type($hash);
1229 $action or die_error(404, "Object does not exist");
1230 } elsif (defined $hash_base && defined $file_name) {
1231 $action = git_get_type("$hash_base:$file_name");
1232 $action or die_error(404, "File or directory does not exist");
1233 } elsif (defined $project) {
1234 $action = 'summary';
1235 } else {
1236 $action = 'project_list';
1237 }
1238 }
1239 if (!defined($actions{$action})) {
1240 die_error(400, "Unknown action");
1241 }
1242 if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1243 !$project) {
1244 die_error(400, "Project needed");
1245 }
1246 $actions{$action}->();
1247 }
1248
1249 sub reset_timer {
1250 our $t0 = [ gettimeofday() ]
1251 if defined $t0;
1252 our $number_of_git_cmds = 0;
1253 }
1254
1255 our $first_request = 1;
1256 sub run_request {
1257 reset_timer();
1258
1259 evaluate_uri();
1260 if ($first_request) {
1261 evaluate_gitweb_config();
1262 evaluate_git_version();
1263 }
1264 if ($per_request_config) {
1265 if (ref($per_request_config) eq 'CODE') {
1266 $per_request_config->();
1267 } elsif (!$first_request) {
1268 evaluate_gitweb_config();
1269 }
1270 }
1271 check_loadavg();
1272
1273 # $projectroot and $projects_list might be set in gitweb config file
1274 $projects_list ||= $projectroot;
1275
1276 evaluate_query_params();
1277 evaluate_path_info();
1278 evaluate_and_validate_params();
1279 evaluate_git_dir();
1280
1281 configure_gitweb_features();
1282
1283 dispatch();
1284 }
1285
1286 our $is_last_request = sub { 1 };
1287 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1288 our $CGI = 'CGI';
1289 our $cgi;
1290 our $FCGI_Stream_PRINT_raw = \&FCGI::Stream::PRINT;
1291 sub configure_as_fcgi {
1292 require CGI::Fast;
1293 our $CGI = 'CGI::Fast';
1294 # FCGI is not Unicode aware hence the UTF-8 encoding must be done manually.
1295 # However no encoding must be done within git_blob_plain() and git_snapshot()
1296 # which must still output in raw binary mode.
1297 no warnings 'redefine';
1298 my $enc = Encode::find_encoding('UTF-8');
1299 *FCGI::Stream::PRINT = sub {
1300 my @OUTPUT = @_;
1301 for (my $i = 1; $i < @_; $i++) {
1302 $OUTPUT[$i] = $enc->encode($_[$i], Encode::FB_CROAK|Encode::LEAVE_SRC);
1303 }
1304 @_ = @OUTPUT;
1305 goto $FCGI_Stream_PRINT_raw;
1306 };
1307
1308 my $request_number = 0;
1309 # let each child service 100 requests
1310 our $is_last_request = sub { ++$request_number > 100 };
1311 }
1312 sub evaluate_argv {
1313 my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1314 configure_as_fcgi()
1315 if $script_name =~ /\.fcgi$/;
1316
1317 return unless (@ARGV);
1318
1319 require Getopt::Long;
1320 Getopt::Long::GetOptions(
1321 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1322 'nproc|n=i' => sub {
1323 my ($arg, $val) = @_;
1324 return unless eval { require FCGI::ProcManager; 1; };
1325 my $proc_manager = FCGI::ProcManager->new({
1326 n_processes => $val,
1327 });
1328 our $pre_listen_hook = sub { $proc_manager->pm_manage() };
1329 our $pre_dispatch_hook = sub { $proc_manager->pm_pre_dispatch() };
1330 our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1331 },
1332 );
1333 }
1334
1335 sub run {
1336 evaluate_argv();
1337
1338 $first_request = 1;
1339 $pre_listen_hook->()
1340 if $pre_listen_hook;
1341
1342 REQUEST:
1343 while ($cgi = $CGI->new()) {
1344 $pre_dispatch_hook->()
1345 if $pre_dispatch_hook;
1346
1347 run_request();
1348
1349 $post_dispatch_hook->()
1350 if $post_dispatch_hook;
1351 $first_request = 0;
1352
1353 last REQUEST if ($is_last_request->());
1354 }
1355
1356 DONE_GITWEB:
1357 1;
1358 }
1359
1360 run();
1361
1362 if (defined caller) {
1363 # wrapped in a subroutine processing requests,
1364 # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1365 return;
1366 } else {
1367 # pure CGI script, serving single request
1368 exit;
1369 }
1370
1371 ## ======================================================================
1372 ## action links
1373
1374 # possible values of extra options
1375 # -full => 0|1 - use absolute/full URL ($my_uri/$my_url as base)
1376 # -replay => 1 - start from a current view (replay with modifications)
1377 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1378 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1379 sub href {
1380 my %params = @_;
1381 # default is to use -absolute url() i.e. $my_uri
1382 my $href = $params{-full} ? $my_url : $my_uri;
1383
1384 # implicit -replay, must be first of implicit params
1385 $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1386
1387 $params{'project'} = $project unless exists $params{'project'};
1388
1389 if ($params{-replay}) {
1390 while (my ($name, $symbol) = each %cgi_param_mapping) {
1391 if (!exists $params{$name}) {
1392 $params{$name} = $input_params{$name};
1393 }
1394 }
1395 }
1396
1397 my $use_pathinfo = gitweb_check_feature('pathinfo');
1398 if (defined $params{'project'} &&
1399 (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1400 # try to put as many parameters as possible in PATH_INFO:
1401 # - project name
1402 # - action
1403 # - hash_parent or hash_parent_base:/file_parent
1404 # - hash or hash_base:/filename
1405 # - the snapshot_format as an appropriate suffix
1406
1407 # When the script is the root DirectoryIndex for the domain,
1408 # $href here would be something like http://gitweb.example.com/
1409 # Thus, we strip any trailing / from $href, to spare us double
1410 # slashes in the final URL
1411 $href =~ s,/$,,;
1412
1413 # Then add the project name, if present
1414 $href .= "/".esc_path_info($params{'project'});
1415 delete $params{'project'};
1416
1417 # since we destructively absorb parameters, we keep this
1418 # boolean that remembers if we're handling a snapshot
1419 my $is_snapshot = $params{'action'} eq 'snapshot';
1420
1421 # Summary just uses the project path URL, any other action is
1422 # added to the URL
1423 if (defined $params{'action'}) {
1424 $href .= "/".esc_path_info($params{'action'})
1425 unless $params{'action'} eq 'summary';
1426 delete $params{'action'};
1427 }
1428
1429 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1430 # stripping nonexistent or useless pieces
1431 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1432 || $params{'hash_parent'} || $params{'hash'});
1433 if (defined $params{'hash_base'}) {
1434 if (defined $params{'hash_parent_base'}) {
1435 $href .= esc_path_info($params{'hash_parent_base'});
1436 # skip the file_parent if it's the same as the file_name
1437 if (defined $params{'file_parent'}) {
1438 if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1439 delete $params{'file_parent'};
1440 } elsif ($params{'file_parent'} !~ /\.\./) {
1441 $href .= ":/".esc_path_info($params{'file_parent'});
1442 delete $params{'file_parent'};
1443 }
1444 }
1445 $href .= "..";
1446 delete $params{'hash_parent'};
1447 delete $params{'hash_parent_base'};
1448 } elsif (defined $params{'hash_parent'}) {
1449 $href .= esc_path_info($params{'hash_parent'}). "..";
1450 delete $params{'hash_parent'};
1451 }
1452
1453 $href .= esc_path_info($params{'hash_base'});
1454 if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1455 $href .= ":/".esc_path_info($params{'file_name'});
1456 delete $params{'file_name'};
1457 }
1458 delete $params{'hash'};
1459 delete $params{'hash_base'};
1460 } elsif (defined $params{'hash'}) {
1461 $href .= esc_path_info($params{'hash'});
1462 delete $params{'hash'};
1463 }
1464
1465 # If the action was a snapshot, we can absorb the
1466 # snapshot_format parameter too
1467 if ($is_snapshot) {
1468 my $fmt = $params{'snapshot_format'};
1469 # snapshot_format should always be defined when href()
1470 # is called, but just in case some code forgets, we
1471 # fall back to the default
1472 $fmt ||= $snapshot_fmts[0];
1473 $href .= $known_snapshot_formats{$fmt}{'suffix'};
1474 delete $params{'snapshot_format'};
1475 }
1476 }
1477
1478 # now encode the parameters explicitly
1479 my @result = ();
1480 for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1481 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1482 if (defined $params{$name}) {
1483 if (ref($params{$name}) eq "ARRAY") {
1484 foreach my $par (@{$params{$name}}) {
1485 push @result, $symbol . "=" . esc_param($par);
1486 }
1487 } else {
1488 push @result, $symbol . "=" . esc_param($params{$name});
1489 }
1490 }
1491 }
1492 $href .= "?" . join(';', @result) if scalar @result;
1493
1494 # final transformation: trailing spaces must be escaped (URI-encoded)
1495 $href =~ s/(\s+)$/CGI::escape($1)/e;
1496
1497 if ($params{-anchor}) {
1498 $href .= "#".esc_param($params{-anchor});
1499 }
1500
1501 return $href;
1502 }
1503
1504
1505 ## ======================================================================
1506 ## validation, quoting/unquoting and escaping
1507
1508 sub is_valid_action {
1509 my $input = shift;
1510 return undef unless exists $actions{$input};
1511 return 1;
1512 }
1513
1514 sub is_valid_project {
1515 my $input = shift;
1516
1517 return unless defined $input;
1518 if (!is_valid_pathname($input) ||
1519 !(-d "$projectroot/$input") ||
1520 !check_export_ok("$projectroot/$input") ||
1521 ($strict_export && !project_in_list($input))) {
1522 return undef;
1523 } else {
1524 return 1;
1525 }
1526 }
1527
1528 sub is_valid_pathname {
1529 my $input = shift;
1530
1531 return undef unless defined $input;
1532 # no '.' or '..' as elements of path, i.e. no '.' or '..'
1533 # at the beginning, at the end, and between slashes.
1534 # also this catches doubled slashes
1535 if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1536 return undef;
1537 }
1538 # no null characters
1539 if ($input =~ m!\0!) {
1540 return undef;
1541 }
1542 return 1;
1543 }
1544
1545 sub is_valid_ref_format {
1546 my $input = shift;
1547
1548 return undef unless defined $input;
1549 # restrictions on ref name according to git-check-ref-format
1550 if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1551 return undef;
1552 }
1553 return 1;
1554 }
1555
1556 sub is_valid_refname {
1557 my $input = shift;
1558
1559 return undef unless defined $input;
1560 # textual hashes are O.K.
1561 if ($input =~ m/^$oid_regex$/) {
1562 return 1;
1563 }
1564 # it must be correct pathname
1565 is_valid_pathname($input) or return undef;
1566 # check git-check-ref-format restrictions
1567 is_valid_ref_format($input) or return undef;
1568 return 1;
1569 }
1570
1571 # decode sequences of octets in utf8 into Perl's internal form,
1572 # which is utf-8 with utf8 flag set if needed. gitweb writes out
1573 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1574 sub to_utf8 {
1575 my $str = shift;
1576 return undef unless defined $str;
1577
1578 if (utf8::is_utf8($str) || utf8::decode($str)) {
1579 return $str;
1580 } else {
1581 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1582 }
1583 }
1584
1585 # quote unsafe chars, but keep the slash, even when it's not
1586 # correct, but quoted slashes look too horrible in bookmarks
1587 sub esc_param {
1588 my $str = shift;
1589 return undef unless defined $str;
1590 $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1591 $str =~ s/ /\+/g;
1592 return $str;
1593 }
1594
1595 # the quoting rules for path_info fragment are slightly different
1596 sub esc_path_info {
1597 my $str = shift;
1598 return undef unless defined $str;
1599
1600 # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1601 $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1602
1603 return $str;
1604 }
1605
1606 # quote unsafe chars in whole URL, so some characters cannot be quoted
1607 sub esc_url {
1608 my $str = shift;
1609 return undef unless defined $str;
1610 $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1611 $str =~ s/ /\+/g;
1612 return $str;
1613 }
1614
1615 # quote unsafe characters in HTML attributes
1616 sub esc_attr {
1617
1618 # for XHTML conformance escaping '"' to '&quot;' is not enough
1619 return esc_html(@_);
1620 }
1621
1622 # replace invalid utf8 character with SUBSTITUTION sequence
1623 sub esc_html {
1624 my $str = shift;
1625 my %opts = @_;
1626
1627 return undef unless defined $str;
1628
1629 $str = to_utf8($str);
1630 $str = $cgi->escapeHTML($str);
1631 if ($opts{'-nbsp'}) {
1632 $str =~ s/ /&nbsp;/g;
1633 }
1634 $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1635 return $str;
1636 }
1637
1638 # quote control characters and escape filename to HTML
1639 sub esc_path {
1640 my $str = shift;
1641 my %opts = @_;
1642
1643 return undef unless defined $str;
1644
1645 $str = to_utf8($str);
1646 $str = $cgi->escapeHTML($str);
1647 if ($opts{'-nbsp'}) {
1648 $str =~ s/ /&nbsp;/g;
1649 }
1650 $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1651 return $str;
1652 }
1653
1654 # Sanitize for use in XHTML + application/xml+xhtml (valid XML 1.0)
1655 sub sanitize {
1656 my $str = shift;
1657
1658 return undef unless defined $str;
1659
1660 $str = to_utf8($str);
1661 $str =~ s|([[:cntrl:]])|(index("\t\n\r", $1) != -1 ? $1 : quot_cec($1))|eg;
1662 return $str;
1663 }
1664
1665 # Make control characters "printable", using character escape codes (CEC)
1666 sub quot_cec {
1667 my $cntrl = shift;
1668 my %opts = @_;
1669 my %es = ( # character escape codes, aka escape sequences
1670 "\t" => '\t', # tab (HT)
1671 "\n" => '\n', # line feed (LF)
1672 "\r" => '\r', # carriage return (CR)
1673 "\f" => '\f', # form feed (FF)
1674 "\b" => '\b', # backspace (BS)
1675 "\a" => '\a', # alarm (bell) (BEL)
1676 "\e" => '\e', # escape (ESC)
1677 "\013" => '\v', # vertical tab (VT)
1678 "\000" => '\0', # nul character (NUL)
1679 );
1680 my $chr = ( (exists $es{$cntrl})
1681 ? $es{$cntrl}
1682 : sprintf('\%2x', ord($cntrl)) );
1683 if ($opts{-nohtml}) {
1684 return $chr;
1685 } else {
1686 return "<span class=\"cntrl\">$chr</span>";
1687 }
1688 }
1689
1690 # Alternatively use unicode control pictures codepoints,
1691 # Unicode "printable representation" (PR)
1692 sub quot_upr {
1693 my $cntrl = shift;
1694 my %opts = @_;
1695
1696 my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1697 if ($opts{-nohtml}) {
1698 return $chr;
1699 } else {
1700 return "<span class=\"cntrl\">$chr</span>";
1701 }
1702 }
1703
1704 # git may return quoted and escaped filenames
1705 sub unquote {
1706 my $str = shift;
1707
1708 sub unq {
1709 my $seq = shift;
1710 my %es = ( # character escape codes, aka escape sequences
1711 't' => "\t", # tab (HT, TAB)
1712 'n' => "\n", # newline (NL)
1713 'r' => "\r", # return (CR)
1714 'f' => "\f", # form feed (FF)
1715 'b' => "\b", # backspace (BS)
1716 'a' => "\a", # alarm (bell) (BEL)
1717 'e' => "\e", # escape (ESC)
1718 'v' => "\013", # vertical tab (VT)
1719 );
1720
1721 if ($seq =~ m/^[0-7]{1,3}$/) {
1722 # octal char sequence
1723 return chr(oct($seq));
1724 } elsif (exists $es{$seq}) {
1725 # C escape sequence, aka character escape code
1726 return $es{$seq};
1727 }
1728 # quoted ordinary character
1729 return $seq;
1730 }
1731
1732 if ($str =~ m/^"(.*)"$/) {
1733 # needs unquoting
1734 $str = $1;
1735 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1736 }
1737 return $str;
1738 }
1739
1740 # escape tabs (convert tabs to spaces)
1741 sub untabify {
1742 my $line = shift;
1743
1744 while ((my $pos = index($line, "\t")) != -1) {
1745 if (my $count = (8 - ($pos % 8))) {
1746 my $spaces = ' ' x $count;
1747 $line =~ s/\t/$spaces/;
1748 }
1749 }
1750
1751 return $line;
1752 }
1753
1754 sub project_in_list {
1755 my $project = shift;
1756 my @list = git_get_projects_list();
1757 return @list && scalar(grep { $_->{'path'} eq $project } @list);
1758 }
1759
1760 ## ----------------------------------------------------------------------
1761 ## HTML aware string manipulation
1762
1763 # Try to chop given string on a word boundary between position
1764 # $len and $len+$add_len. If there is no word boundary there,
1765 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1766 # (marking chopped part) would be longer than given string.
1767 sub chop_str {
1768 my $str = shift;
1769 my $len = shift;
1770 my $add_len = shift || 10;
1771 my $where = shift || 'right'; # 'left' | 'center' | 'right'
1772
1773 # Make sure perl knows it is utf8 encoded so we don't
1774 # cut in the middle of a utf8 multibyte char.
1775 $str = to_utf8($str);
1776
1777 # allow only $len chars, but don't cut a word if it would fit in $add_len
1778 # if it doesn't fit, cut it if it's still longer than the dots we would add
1779 # remove chopped character entities entirely
1780
1781 # when chopping in the middle, distribute $len into left and right part
1782 # return early if chopping wouldn't make string shorter
1783 if ($where eq 'center') {
1784 return $str if ($len + 5 >= length($str)); # filler is length 5
1785 $len = int($len/2);
1786 } else {
1787 return $str if ($len + 4 >= length($str)); # filler is length 4
1788 }
1789
1790 # regexps: ending and beginning with word part up to $add_len
1791 my $endre = qr/.{$len}\w{0,$add_len}/;
1792 my $begre = qr/\w{0,$add_len}.{$len}/;
1793
1794 if ($where eq 'left') {
1795 $str =~ m/^(.*?)($begre)$/;
1796 my ($lead, $body) = ($1, $2);
1797 if (length($lead) > 4) {
1798 $lead = " ...";
1799 }
1800 return "$lead$body";
1801
1802 } elsif ($where eq 'center') {
1803 $str =~ m/^($endre)(.*)$/;
1804 my ($left, $str) = ($1, $2);
1805 $str =~ m/^(.*?)($begre)$/;
1806 my ($mid, $right) = ($1, $2);
1807 if (length($mid) > 5) {
1808 $mid = " ... ";
1809 }
1810 return "$left$mid$right";
1811
1812 } else {
1813 $str =~ m/^($endre)(.*)$/;
1814 my $body = $1;
1815 my $tail = $2;
1816 if (length($tail) > 4) {
1817 $tail = "... ";
1818 }
1819 return "$body$tail";
1820 }
1821 }
1822
1823 # takes the same arguments as chop_str, but also wraps a <span> around the
1824 # result with a title attribute if it does get chopped. Additionally, the
1825 # string is HTML-escaped.
1826 sub chop_and_escape_str {
1827 my ($str) = @_;
1828
1829 my $chopped = chop_str(@_);
1830 $str = to_utf8($str);
1831 if ($chopped eq $str) {
1832 return esc_html($chopped);
1833 } else {
1834 $str =~ s/[[:cntrl:]]/?/g;
1835 return $cgi->span({-title=>$str}, esc_html($chopped));
1836 }
1837 }
1838
1839 # Highlight selected fragments of string, using given CSS class,
1840 # and escape HTML. It is assumed that fragments do not overlap.
1841 # Regions are passed as list of pairs (array references).
1842 #
1843 # Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
1844 # '<span class="mark">foo</span>bar'
1845 sub esc_html_hl_regions {
1846 my ($str, $css_class, @sel) = @_;
1847 my %opts = grep { ref($_) ne 'ARRAY' } @sel;
1848 @sel = grep { ref($_) eq 'ARRAY' } @sel;
1849 return esc_html($str, %opts) unless @sel;
1850
1851 my $out = '';
1852 my $pos = 0;
1853
1854 for my $s (@sel) {
1855 my ($begin, $end) = @$s;
1856
1857 # Don't create empty <span> elements.
1858 next if $end <= $begin;
1859
1860 my $escaped = esc_html(substr($str, $begin, $end - $begin),
1861 %opts);
1862
1863 $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
1864 if ($begin - $pos > 0);
1865 $out .= "<mark class=\"$css_class\">$escaped</mark>";
1866
1867 $pos = $end;
1868 }
1869 $out .= esc_html(substr($str, $pos), %opts)
1870 if ($pos < length($str));
1871
1872 return $out;
1873 }
1874
1875 # return positions of beginning and end of each match
1876 sub matchpos_list {
1877 my ($str, $regexp) = @_;
1878 return unless (defined $str && defined $regexp);
1879
1880 my @matches;
1881 while ($str =~ /$regexp/g) {
1882 push @matches, [$-[0], $+[0]];
1883 }
1884 return @matches;
1885 }
1886
1887 # highlight match (if any), and escape HTML
1888 sub esc_html_match_hl {
1889 my ($str, $regexp) = @_;
1890 return esc_html($str) unless defined $regexp;
1891
1892 my @matches = matchpos_list($str, $regexp);
1893 return esc_html($str) unless @matches;
1894
1895 return esc_html_hl_regions($str, undef, @matches);
1896 }
1897
1898
1899 # highlight match (if any) of shortened string, and escape HTML
1900 sub esc_html_match_hl_chopped {
1901 my ($str, $chopped, $regexp) = @_;
1902 return esc_html_match_hl($str, $regexp) unless defined $chopped;
1903
1904 my @matches = matchpos_list($str, $regexp);
1905 return esc_html($chopped) unless @matches;
1906
1907 # filter matches so that we mark chopped string
1908 my $tail = "... "; # see chop_str
1909 unless ($chopped =~ s/\Q$tail\E$//) {
1910 $tail = '';
1911 }
1912 my $chop_len = length($chopped);
1913 my $tail_len = length($tail);
1914 my @filtered;
1915
1916 for my $m (@matches) {
1917 if ($m->[0] > $chop_len) {
1918 push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
1919 last;
1920 } elsif ($m->[1] > $chop_len) {
1921 push @filtered, [ $m->[0], $chop_len + $tail_len ];
1922 last;
1923 }
1924 push @filtered, $m;
1925 }
1926
1927 return esc_html_hl_regions($chopped . $tail, undef, @filtered);
1928 }
1929
1930 ## ----------------------------------------------------------------------
1931 ## functions returning short strings
1932
1933 # CSS class for given age value (in seconds)
1934 sub age_class {
1935 my $age = shift;
1936
1937 if (!defined $age) {
1938 return "noage";
1939 } elsif ($age < 60*60*2) {
1940 return "age0";
1941 } elsif ($age < 60*60*24*2) {
1942 return "age1";
1943 } else {
1944 return "age2";
1945 }
1946 }
1947
1948 # convert age in seconds to "nn units ago" string
1949 sub age_string {
1950 my $age = shift;
1951 my $age_str;
1952
1953 if ($age > 60*60*24*365*2) {
1954 $age_str = (int $age/60/60/24/365);
1955 $age_str .= " years ago";
1956 } elsif ($age > 60*60*24*(365/12)*2) {
1957 $age_str = int $age/60/60/24/(365/12);
1958 $age_str .= " months ago";
1959 } elsif ($age > 60*60*24*7*2) {
1960 $age_str = int $age/60/60/24/7;
1961 $age_str .= " weeks ago";
1962 } elsif ($age > 60*60*24*2) {
1963 $age_str = int $age/60/60/24;
1964 $age_str .= " days ago";
1965 } elsif ($age > 60*60*2) {
1966 $age_str = int $age/60/60;
1967 $age_str .= " hours ago";
1968 } elsif ($age > 60*2) {
1969 $age_str = int $age/60;
1970 $age_str .= " min ago";
1971 } elsif ($age > 2) {
1972 $age_str = int $age;
1973 $age_str .= " sec ago";
1974 } else {
1975 $age_str .= " right now";
1976 }
1977 return $age_str;
1978 }
1979
1980 use constant {
1981 S_IFINVALID => 0030000,
1982 S_IFGITLINK => 0160000,
1983 };
1984
1985 # submodule/subproject, a commit object reference
1986 sub S_ISGITLINK {
1987 my $mode = shift;
1988
1989 return (($mode & S_IFMT) == S_IFGITLINK)
1990 }
1991
1992 # convert file mode in octal to symbolic file mode string
1993 sub mode_str {
1994 my $mode = oct shift;
1995
1996 if (S_ISGITLINK($mode)) {
1997 return 'm---------';
1998 } elsif (S_ISDIR($mode & S_IFMT)) {
1999 return 'drwxr-xr-x';
2000 } elsif (S_ISLNK($mode)) {
2001 return 'lrwxrwxrwx';
2002 } elsif (S_ISREG($mode)) {
2003 # git cares only about the executable bit
2004 if ($mode & S_IXUSR) {
2005 return '-rwxr-xr-x';
2006 } else {
2007 return '-rw-r--r--';
2008 };
2009 } else {
2010 return '----------';
2011 }
2012 }
2013
2014 # convert file mode in octal to file type string
2015 sub file_type {
2016 my $mode = shift;
2017
2018 if ($mode !~ m/^[0-7]+$/) {
2019 return $mode;
2020 } else {
2021 $mode = oct $mode;
2022 }
2023
2024 if (S_ISGITLINK($mode)) {
2025 return "submodule";
2026 } elsif (S_ISDIR($mode & S_IFMT)) {
2027 return "directory";
2028 } elsif (S_ISLNK($mode)) {
2029 return "symlink";
2030 } elsif (S_ISREG($mode)) {
2031 return "file";
2032 } else {
2033 return "unknown";
2034 }
2035 }
2036
2037 # convert file mode in octal to file type description string
2038 sub file_type_long {
2039 my $mode = shift;
2040
2041 if ($mode !~ m/^[0-7]+$/) {
2042 return $mode;
2043 } else {
2044 $mode = oct $mode;
2045 }
2046
2047 if (S_ISGITLINK($mode)) {
2048 return "submodule";
2049 } elsif (S_ISDIR($mode & S_IFMT)) {
2050 return "directory";
2051 } elsif (S_ISLNK($mode)) {
2052 return "symlink";
2053 } elsif (S_ISREG($mode)) {
2054 if ($mode & S_IXUSR) {
2055 return "executable";
2056 } else {
2057 return "file";
2058 };
2059 } else {
2060 return "unknown";
2061 }
2062 }
2063
2064
2065 ## ----------------------------------------------------------------------
2066 ## functions returning short HTML fragments, or transforming HTML fragments
2067 ## which don't belong to other sections
2068
2069 # format line of commit message.
2070 sub format_log_line_html {
2071 my $line = shift;
2072
2073 # Potentially abbreviated OID.
2074 my $regex = oid_nlen_regex("7,64");
2075
2076 $line = esc_html($line, -nbsp=>1);
2077 $line =~ s{
2078 \b
2079 (
2080 # The output of "git describe", e.g. v2.10.0-297-gf6727b0
2081 # or hadoop-20160921-113441-20-g094fb7d
2082 (?<!-) # see strbuf_check_tag_ref(). Tags can't start with -
2083 [A-Za-z0-9.-]+
2084 (?!\.) # refs can't end with ".", see check_refname_format()
2085 -g$regex
2086 |
2087 # Just a normal looking Git SHA1
2088 $regex
2089 )
2090 \b
2091 }{
2092 $cgi->a({-href => href(action=>"object", hash=>$1),
2093 -class => "text"}, $1);
2094 }egx;
2095
2096 return $line;
2097 }
2098
2099 # format marker of refs pointing to given object
2100
2101 # the destination action is chosen based on object type and current context:
2102 # - for annotated tags, we choose the tag view unless it's the current view
2103 # already, in which case we go to shortlog view
2104 # - for other refs, we keep the current view if we're in history, shortlog or
2105 # log view, and select shortlog otherwise
2106 sub format_ref_marker {
2107 my ($refs, $id) = @_;
2108 my $markers = '';
2109
2110 if (defined $refs->{$id}) {
2111 foreach my $ref (@{$refs->{$id}}) {
2112 # this code exploits the fact that non-lightweight tags are the
2113 # only indirect objects, and that they are the only objects for which
2114 # we want to use tag instead of shortlog as action
2115 my ($type, $name) = qw();
2116 my $indirect = ($ref =~ s/\^\{\}$//);
2117 # e.g. tags/v2.6.11 or heads/next
2118 if ($ref =~ m!^(.*?)s?/(.*)$!) {
2119 $type = $1;
2120 $name = $2;
2121 } else {
2122 $type = "ref";
2123 $name = $ref;
2124 }
2125
2126 my $class = $type;
2127 $class .= " indirect" if $indirect;
2128
2129 my $dest_action = "shortlog";
2130
2131 if ($indirect) {
2132 $dest_action = "tag" unless $action eq "tag";
2133 } elsif ($action =~ /^(history|(short)?log)$/) {
2134 $dest_action = $action;
2135 }
2136
2137 my $dest = "";
2138 $dest .= "refs/" unless $ref =~ m!^refs/!;
2139 $dest .= $ref;
2140
2141 my $link = $cgi->a({
2142 -href => href(
2143 action=>$dest_action,
2144 hash=>$dest
2145 )}, esc_html($name));
2146
2147 $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
2148 $link . "</span>";
2149 }
2150 }
2151
2152 if ($markers) {
2153 return ' <span class="refs">'. $markers . '</span>';
2154 } else {
2155 return "";
2156 }
2157 }
2158
2159 # format, perhaps shortened and with markers, title line
2160 sub format_subject_html {
2161 my ($long, $short, $href, $extra) = @_;
2162 $extra = '' unless defined($extra);
2163
2164 if (length($short) < length($long)) {
2165 $long =~ s/[[:cntrl:]]/?/g;
2166 return $cgi->a({-href => $href, -class => "list subject",
2167 -title => to_utf8($long)},
2168 esc_html($short)) . $extra;
2169 } else {
2170 return $cgi->a({-href => $href, -class => "list subject"},
2171 esc_html($long)) . $extra;
2172 }
2173 }
2174
2175 # Rather than recomputing the url for an email multiple times, we cache it
2176 # after the first hit. This gives a visible benefit in views where the avatar
2177 # for the same email is used repeatedly (e.g. shortlog).
2178 # The cache is shared by all avatar engines (currently gravatar only), which
2179 # are free to use it as preferred. Since only one avatar engine is used for any
2180 # given page, there's no risk for cache conflicts.
2181 our %avatar_cache = ();
2182
2183 # Compute the picon url for a given email, by using the picon search service over at
2184 # http://www.cs.indiana.edu/picons/search.html
2185 sub picon_url {
2186 my $email = lc shift;
2187 if (!$avatar_cache{$email}) {
2188 my ($user, $domain) = split('@', $email);
2189 $avatar_cache{$email} =
2190 "//www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
2191 "$domain/$user/" .
2192 "users+domains+unknown/up/single";
2193 }
2194 return $avatar_cache{$email};
2195 }
2196
2197 # Compute the gravatar url for a given email, if it's not in the cache already.
2198 # Gravatar stores only the part of the URL before the size, since that's the
2199 # one computationally more expensive. This also allows reuse of the cache for
2200 # different sizes (for this particular engine).
2201 sub gravatar_url {
2202 my $email = lc shift;
2203 my $size = shift;
2204 $avatar_cache{$email} ||=
2205 "//www.gravatar.com/avatar/" .
2206 md5_hex($email) . "?s=";
2207 return $avatar_cache{$email} . $size;
2208 }
2209
2210 # Insert an avatar for the given $email at the given $size if the feature
2211 # is enabled.
2212 sub git_get_avatar {
2213 my ($email, %opts) = @_;
2214 my $pre_white = ($opts{-pad_before} ? "&nbsp;" : "");
2215 my $post_white = ($opts{-pad_after} ? "&nbsp;" : "");
2216 $opts{-size} ||= 'default';
2217 my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
2218 my $url = "";
2219 if ($git_avatar eq 'gravatar') {
2220 $url = gravatar_url($email, $size);
2221 } elsif ($git_avatar eq 'picon') {
2222 $url = picon_url($email);
2223 }
2224 # Other providers can be added by extending the if chain, defining $url
2225 # as needed. If no variant puts something in $url, we assume avatars
2226 # are completely disabled/unavailable.
2227 if ($url) {
2228 return $pre_white .
2229 "<img width=\"$size\" " .
2230 "class=\"avatar\" " .
2231 "src=\"".esc_url($url)."\" " .
2232 "alt=\"\" " .
2233 "/>" . $post_white;
2234 } else {
2235 return "";
2236 }
2237 }
2238
2239 sub format_search_author {
2240 my ($author, $searchtype, $displaytext) = @_;
2241 my $have_search = gitweb_check_feature('search');
2242
2243 if ($have_search) {
2244 my $performed = "";
2245 if ($searchtype eq 'author') {
2246 $performed = "authored";
2247 } elsif ($searchtype eq 'committer') {
2248 $performed = "committed";
2249 }
2250
2251 return $cgi->a({-href => href(action=>"search", hash=>$hash,
2252 searchtext=>$author,
2253 searchtype=>$searchtype), class=>"list",
2254 title=>"Search for commits $performed by $author"},
2255 $displaytext);
2256
2257 } else {
2258 return $displaytext;
2259 }
2260 }
2261
2262 # format the author name of the given commit with the given tag
2263 # the author name is chopped and escaped according to the other
2264 # optional parameters (see chop_str).
2265 sub format_author_html {
2266 my $tag = shift;
2267 my $co = shift;
2268 my $author = chop_and_escape_str($co->{'author_name'}, @_);
2269 return "<$tag class=\"author\">" .
2270 format_search_author($co->{'author_name'}, "author",
2271 git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2272 $author) .
2273 "</$tag>";
2274 }
2275
2276 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2277 sub format_git_diff_header_line {
2278 my $line = shift;
2279 my $diffinfo = shift;
2280 my ($from, $to) = @_;
2281
2282 if ($diffinfo->{'nparents'}) {
2283 # combined diff
2284 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2285 if ($to->{'href'}) {
2286 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2287 esc_path($to->{'file'}));
2288 } else { # file was deleted (no href)
2289 $line .= esc_path($to->{'file'});
2290 }
2291 } else {
2292 # "ordinary" diff
2293 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2294 if ($from->{'href'}) {
2295 $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2296 'a/' . esc_path($from->{'file'}));
2297 } else { # file was added (no href)
2298 $line .= 'a/' . esc_path($from->{'file'});
2299 }
2300 $line .= ' ';
2301 if ($to->{'href'}) {
2302 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2303 'b/' . esc_path($to->{'file'}));
2304 } else { # file was deleted
2305 $line .= 'b/' . esc_path($to->{'file'});
2306 }
2307 }
2308
2309 return "<div class=\"diff header\">$line</div>\n";
2310 }
2311
2312 # format extended diff header line, before patch itself
2313 sub format_extended_diff_header_line {
2314 my $line = shift;
2315 my $diffinfo = shift;
2316 my ($from, $to) = @_;
2317
2318 # match <path>
2319 if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2320 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2321 esc_path($from->{'file'}));
2322 }
2323 if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2324 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2325 esc_path($to->{'file'}));
2326 }
2327 # match single <mode>
2328 if ($line =~ m/\s(\d{6})$/) {
2329 $line .= '<span class="info"> (' .
2330 file_type_long($1) .
2331 ')</span>';
2332 }
2333 # match <hash>
2334 if ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", ",") |
2335 $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", ",")) {
2336 # can match only for combined diff
2337 $line = 'index ';
2338 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2339 if ($from->{'href'}[$i]) {
2340 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2341 -class=>"hash"},
2342 substr($diffinfo->{'from_id'}[$i],0,7));
2343 } else {
2344 $line .= '0' x 7;
2345 }
2346 # separator
2347 $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2348 }
2349 $line .= '..';
2350 if ($to->{'href'}) {
2351 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2352 substr($diffinfo->{'to_id'},0,7));
2353 } else {
2354 $line .= '0' x 7;
2355 }
2356
2357 } elsif ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", "..") |
2358 $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", "..")) {
2359 # can match only for ordinary diff
2360 my ($from_link, $to_link);
2361 if ($from->{'href'}) {
2362 $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2363 substr($diffinfo->{'from_id'},0,7));
2364 } else {
2365 $from_link = '0' x 7;
2366 }
2367 if ($to->{'href'}) {
2368 $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2369 substr($diffinfo->{'to_id'},0,7));
2370 } else {
2371 $to_link = '0' x 7;
2372 }
2373 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2374 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2375 }
2376
2377 return $line . "<br/>\n";
2378 }
2379
2380 # format from-file/to-file diff header
2381 sub format_diff_from_to_header {
2382 my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2383 my $line;
2384 my $result = '';
2385
2386 $line = $from_line;
2387 #assert($line =~ m/^---/) if DEBUG;
2388 # no extra formatting for "^--- /dev/null"
2389 if (! $diffinfo->{'nparents'}) {
2390 # ordinary (single parent) diff
2391 if ($line =~ m!^--- "?a/!) {
2392 if ($from->{'href'}) {
2393 $line = '--- a/' .
2394 $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2395 esc_path($from->{'file'}));
2396 } else {
2397 $line = '--- a/' .
2398 esc_path($from->{'file'});
2399 }
2400 }
2401 $result .= qq!<div class="diff from_file">$line</div>\n!;
2402
2403 } else {
2404 # combined diff (merge commit)
2405 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2406 if ($from->{'href'}[$i]) {
2407 $line = '--- ' .
2408 $cgi->a({-href=>href(action=>"blobdiff",
2409 hash_parent=>$diffinfo->{'from_id'}[$i],
2410 hash_parent_base=>$parents[$i],
2411 file_parent=>$from->{'file'}[$i],
2412 hash=>$diffinfo->{'to_id'},
2413 hash_base=>$hash,
2414 file_name=>$to->{'file'}),
2415 -class=>"path",
2416 -title=>"diff" . ($i+1)},
2417 $i+1) .
2418 '/' .
2419 $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2420 esc_path($from->{'file'}[$i]));
2421 } else {
2422 $line = '--- /dev/null';
2423 }
2424 $result .= qq!<div class="diff from_file">$line</div>\n!;
2425 }
2426 }
2427
2428 $line = $to_line;
2429 #assert($line =~ m/^\+\+\+/) if DEBUG;
2430 # no extra formatting for "^+++ /dev/null"
2431 if ($line =~ m!^\+\+\+ "?b/!) {
2432 if ($to->{'href'}) {
2433 $line = '+++ b/' .
2434 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2435 esc_path($to->{'file'}));
2436 } else {
2437 $line = '+++ b/' .
2438 esc_path($to->{'file'});
2439 }
2440 }
2441 $result .= qq!<div class="diff to_file">$line</div>\n!;
2442
2443 return $result;
2444 }
2445
2446 # create note for patch simplified by combined diff
2447 sub format_diff_cc_simplified {
2448 my ($diffinfo, @parents) = @_;
2449 my $result = '';
2450
2451 $result .= "<div class=\"diff header\">" .
2452 "diff --cc ";
2453 if (!is_deleted($diffinfo)) {
2454 $result .= $cgi->a({-href => href(action=>"blob",
2455 hash_base=>$hash,
2456 hash=>$diffinfo->{'to_id'},
2457 file_name=>$diffinfo->{'to_file'}),
2458 -class => "path"},
2459 esc_path($diffinfo->{'to_file'}));
2460 } else {
2461 $result .= esc_path($diffinfo->{'to_file'});
2462 }
2463 $result .= "</div>\n" . # class="diff header"
2464 "<div class=\"diff nodifferences\">" .
2465 "Simple merge" .
2466 "</div>\n"; # class="diff nodifferences"
2467
2468 return $result;
2469 }
2470
2471 sub diff_line_class {
2472 my ($line, $from, $to) = @_;
2473
2474 # ordinary diff
2475 my $num_sign = 1;
2476 # combined diff
2477 if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2478 $num_sign = scalar @{$from->{'href'}};
2479 }
2480
2481 my @diff_line_classifier = (
2482 { regexp => qr/^\@\@{$num_sign} /, class => "chunk_header"},
2483 { regexp => qr/^\\/, class => "incomplete" },
2484 { regexp => qr/^ {$num_sign}/, class => "ctx" },
2485 # classifier for context must come before classifier add/rem,
2486 # or we would have to use more complicated regexp, for example
2487 # qr/(?= {0,$m}\+)[+ ]{$num_sign}/, where $m = $num_sign - 1;
2488 { regexp => qr/^[+ ]{$num_sign}/, class => "add" },
2489 { regexp => qr/^[- ]{$num_sign}/, class => "rem" },
2490 );
2491 for my $clsfy (@diff_line_classifier) {
2492 return $clsfy->{'class'}
2493 if ($line =~ $clsfy->{'regexp'});
2494 }
2495
2496 # fallback
2497 return "";
2498 }
2499
2500 # assumes that $from and $to are defined and correctly filled,
2501 # and that $line holds a line of chunk header for unified diff
2502 sub format_unidiff_chunk_header {
2503 my ($line, $from, $to) = @_;
2504
2505 my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2506 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2507
2508 $from_lines = 0 unless defined $from_lines;
2509 $to_lines = 0 unless defined $to_lines;
2510
2511 if ($from->{'href'}) {
2512 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2513 -class=>"list"}, $from_text);
2514 }
2515 if ($to->{'href'}) {
2516 $to_text = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2517 -class=>"list"}, $to_text);
2518 }
2519 $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2520 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2521 return $line;
2522 }
2523
2524 # assumes that $from and $to are defined and correctly filled,
2525 # and that $line holds a line of chunk header for combined diff
2526 sub format_cc_diff_chunk_header {
2527 my ($line, $from, $to) = @_;
2528
2529 my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2530 my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2531
2532 @from_text = split(' ', $ranges);
2533 for (my $i = 0; $i < @from_text; ++$i) {
2534 ($from_start[$i], $from_nlines[$i]) =
2535 (split(',', substr($from_text[$i], 1)), 0);
2536 }
2537
2538 $to_text = pop @from_text;
2539 $to_start = pop @from_start;
2540 $to_nlines = pop @from_nlines;
2541
2542 $line = "<span class=\"chunk_info\">$prefix ";
2543 for (my $i = 0; $i < @from_text; ++$i) {
2544 if ($from->{'href'}[$i]) {
2545 $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2546 -class=>"list"}, $from_text[$i]);
2547 } else {
2548 $line .= $from_text[$i];
2549 }
2550 $line .= " ";
2551 }
2552 if ($to->{'href'}) {
2553 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2554 -class=>"list"}, $to_text);
2555 } else {
2556 $line .= $to_text;
2557 }
2558 $line .= " $prefix</span>" .
2559 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2560 return $line;
2561 }
2562
2563 # process patch (diff) line (not to be used for diff headers),
2564 # returning HTML-formatted (but not wrapped) line.
2565 # If the line is passed as a reference, it is treated as HTML and not
2566 # esc_html()'ed.
2567 sub format_diff_line {
2568 my ($line, $diff_class, $from, $to) = @_;
2569
2570 if (ref($line)) {
2571 $line = $$line;
2572 } else {
2573 chomp $line;
2574 $line = untabify($line);
2575
2576 if ($from && $to && $line =~ m/^\@{2} /) {
2577 $line = format_unidiff_chunk_header($line, $from, $to);
2578 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2579 $line = format_cc_diff_chunk_header($line, $from, $to);
2580 } else {
2581 $line = esc_html($line, -nbsp=>1);
2582 }
2583 }
2584
2585 my $diff_classes = "diff";
2586 $diff_classes .= " $diff_class" if ($diff_class);
2587 $line = "<div class=\"$diff_classes\">$line</div>\n";
2588
2589 return $line;
2590 }
2591
2592 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2593 # linked. Pass the hash of the tree/commit to snapshot.
2594 sub format_snapshot_links {
2595 my ($hash) = @_;
2596 my $num_fmts = @snapshot_fmts;
2597 if ($num_fmts > 1) {
2598 # A parenthesized list of links bearing format names.
2599 # e.g. "snapshot (_tar.gz_ _zip_)"
2600 return "snapshot (" . join(' ', map
2601 $cgi->a({
2602 -href => href(
2603 action=>"snapshot",
2604 hash=>$hash,
2605 snapshot_format=>$_
2606 )
2607 }, $known_snapshot_formats{$_}{'display'})
2608 , @snapshot_fmts) . ")";
2609 } elsif ($num_fmts == 1) {
2610 # A single "snapshot" link whose tooltip bears the format name.
2611 # i.e. "_snapshot_"
2612 my ($fmt) = @snapshot_fmts;
2613 return
2614 $cgi->a({
2615 -href => href(
2616 action=>"snapshot",
2617 hash=>$hash,
2618 snapshot_format=>$fmt
2619 ),
2620 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2621 }, "snapshot");
2622 } else { # $num_fmts == 0
2623 return undef;
2624 }
2625 }
2626
2627 ## ......................................................................
2628 ## functions returning values to be passed, perhaps after some
2629 ## transformation, to other functions; e.g. returning arguments to href()
2630
2631 # returns hash to be passed to href to generate gitweb URL
2632 # in -title key it returns description of link
2633 sub get_feed_info {
2634 my $format = shift || 'Atom';
2635 my %res = (action => lc($format));
2636 my $matched_ref = 0;
2637
2638 # feed links are possible only for project views
2639 return unless (defined $project);
2640 # some views should link to OPML, or to generic project feed,
2641 # or don't have specific feed yet (so they should use generic)
2642 return if (!$action || $action =~ /^(?:tags|heads|forks|tag|search)$/x);
2643
2644 my $branch = undef;
2645 # branches refs uses 'refs/' + $get_branch_refs()[x] + '/' prefix
2646 # (fullname) to differentiate from tag links; this also makes
2647 # possible to detect branch links
2648 for my $ref (get_branch_refs()) {
2649 if ((defined $hash_base && $hash_base =~ m!^refs/\Q$ref\E/(.*)$!) ||
2650 (defined $hash && $hash =~ m!^refs/\Q$ref\E/(.*)$!)) {
2651 $branch = $1;
2652 $matched_ref = $ref;
2653 last;
2654 }
2655 }
2656 # find log type for feed description (title)
2657 my $type = 'log';
2658 if (defined $file_name) {
2659 $type = "history of $file_name";
2660 $type .= "/" if ($action eq 'tree');
2661 $type .= " on '$branch'" if (defined $branch);
2662 } else {
2663 $type = "log of $branch" if (defined $branch);
2664 }
2665
2666 $res{-title} = $type;
2667 $res{'hash'} = (defined $branch ? "refs/$matched_ref/$branch" : undef);
2668 $res{'file_name'} = $file_name;
2669
2670 return %res;
2671 }
2672
2673 ## ----------------------------------------------------------------------
2674 ## git utility subroutines, invoking git commands
2675
2676 # returns path to the core git executable and the --git-dir parameter as list
2677 sub git_cmd {
2678 $number_of_git_cmds++;
2679 return $GIT, '--git-dir='.$git_dir;
2680 }
2681
2682 # quote the given arguments for passing them to the shell
2683 # quote_command("command", "arg 1", "arg with ' and ! characters")
2684 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2685 # Try to avoid using this function wherever possible.
2686 sub quote_command {
2687 return join(' ',
2688 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2689 }
2690
2691 # get HEAD ref of given project as hash
2692 sub git_get_head_hash {
2693 return git_get_full_hash(shift, 'HEAD');
2694 }
2695
2696 sub git_get_full_hash {
2697 return git_get_hash(@_);
2698 }
2699
2700 sub git_get_short_hash {
2701 return git_get_hash(@_, '--short=7');
2702 }
2703
2704 sub git_get_hash {
2705 my ($project, $hash, @options) = @_;
2706 my $o_git_dir = $git_dir;
2707 my $retval = undef;
2708 $git_dir = "$projectroot/$project";
2709 if (open my $fd, '-|', git_cmd(), 'rev-parse',
2710 '--verify', '-q', @options, $hash) {
2711 $retval = <$fd>;
2712 chomp $retval if defined $retval;
2713 close $fd;
2714 }
2715 if (defined $o_git_dir) {
2716 $git_dir = $o_git_dir;
2717 }
2718 return $retval;
2719 }
2720
2721 # get type of given object
2722 sub git_get_type {
2723 my $hash = shift;
2724
2725 open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2726 my $type = <$fd>;
2727 close $fd or return;
2728 chomp $type;
2729 return $type;
2730 }
2731
2732 # repository configuration
2733 our $config_file = '';
2734 our %config;
2735
2736 # store multiple values for single key as anonymous array reference
2737 # single values stored directly in the hash, not as [ <value> ]
2738 sub hash_set_multi {
2739 my ($hash, $key, $value) = @_;
2740
2741 if (!exists $hash->{$key}) {
2742 $hash->{$key} = $value;
2743 } elsif (!ref $hash->{$key}) {
2744 $hash->{$key} = [ $hash->{$key}, $value ];
2745 } else {
2746 push @{$hash->{$key}}, $value;
2747 }
2748 }
2749
2750 # return hash of git project configuration
2751 # optionally limited to some section, e.g. 'gitweb'
2752 sub git_parse_project_config {
2753 my $section_regexp = shift;
2754 my %config;
2755
2756 local $/ = "\0";
2757
2758 open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2759 or return;
2760
2761 while (my $keyval = <$fh>) {
2762 chomp $keyval;
2763 my ($key, $value) = split(/\n/, $keyval, 2);
2764
2765 hash_set_multi(\%config, $key, $value)
2766 if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2767 }
2768 close $fh;
2769
2770 return %config;
2771 }
2772
2773 # convert config value to boolean: 'true' or 'false'
2774 # no value, number > 0, 'true' and 'yes' values are true
2775 # rest of values are treated as false (never as error)
2776 sub config_to_bool {
2777 my $val = shift;
2778
2779 return 1 if !defined $val; # section.key
2780
2781 # strip leading and trailing whitespace
2782 $val =~ s/^\s+//;
2783 $val =~ s/\s+$//;
2784
2785 return (($val =~ /^\d+$/ && $val) || # section.key = 1
2786 ($val =~ /^(?:true|yes)$/i)); # section.key = true
2787 }
2788
2789 # convert config value to simple decimal number
2790 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2791 # to be multiplied by 1024, 1048576, or 1073741824
2792 sub config_to_int {
2793 my $val = shift;
2794
2795 # strip leading and trailing whitespace
2796 $val =~ s/^\s+//;
2797 $val =~ s/\s+$//;
2798
2799 if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2800 $unit = lc($unit);
2801 # unknown unit is treated as 1
2802 return $num * ($unit eq 'g' ? 1073741824 :
2803 $unit eq 'm' ? 1048576 :
2804 $unit eq 'k' ? 1024 : 1);
2805 }
2806 return $val;
2807 }
2808
2809 # convert config value to array reference, if needed
2810 sub config_to_multi {
2811 my $val = shift;
2812
2813 return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2814 }
2815
2816 sub git_get_project_config {
2817 my ($key, $type) = @_;
2818
2819 return unless defined $git_dir;
2820
2821 # key sanity check
2822 return unless ($key);
2823 # only subsection, if exists, is case sensitive,
2824 # and not lowercased by 'git config -z -l'
2825 if (my ($hi, $mi, $lo) = ($key =~ /^([^.]*)\.(.*)\.([^.]*)$/)) {
2826 $lo =~ s/_//g;
2827 $key = join(".", lc($hi), $mi, lc($lo));
2828 return if ($lo =~ /\W/ || $hi =~ /\W/);
2829 } else {
2830 $key = lc($key);
2831 $key =~ s/_//g;
2832 return if ($key =~ /\W/);
2833 }
2834 $key =~ s/^gitweb\.//;
2835
2836 # type sanity check
2837 if (defined $type) {
2838 $type =~ s/^--//;
2839 $type = undef
2840 unless ($type eq 'bool' || $type eq 'int');
2841 }
2842
2843 # get config
2844 if (!defined $config_file ||
2845 $config_file ne "$git_dir/config") {
2846 %config = git_parse_project_config('gitweb');
2847 $config_file = "$git_dir/config";
2848 }
2849
2850 # check if config variable (key) exists
2851 return unless exists $config{"gitweb.$key"};
2852
2853 # ensure given type
2854 if (!defined $type) {
2855 return $config{"gitweb.$key"};
2856 } elsif ($type eq 'bool') {
2857 # backward compatibility: 'git config --bool' returns true/false
2858 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2859 } elsif ($type eq 'int') {
2860 return config_to_int($config{"gitweb.$key"});
2861 }
2862 return $config{"gitweb.$key"};
2863 }
2864
2865 # get hash of given path at given ref
2866 sub git_get_hash_by_path {
2867 my $base = shift;
2868 my $path = shift || return undef;
2869 my $type = shift;
2870
2871 $path =~ s,/+$,,;
2872
2873 open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2874 or die_error(500, "Open git-ls-tree failed");
2875 my $line = <$fd>;
2876 close $fd or return undef;
2877
2878 if (!defined $line) {
2879 # there is no tree or hash given by $path at $base
2880 return undef;
2881 }
2882
2883 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
2884 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/;
2885 if (defined $type && $type ne $2) {
2886 # type doesn't match
2887 return undef;
2888 }
2889 return $3;
2890 }
2891
2892 # get path of entry with given hash at given tree-ish (ref)
2893 # used to get 'from' filename for combined diff (merge commit) for renames
2894 sub git_get_path_by_hash {
2895 my $base = shift || return;
2896 my $hash = shift || return;
2897
2898 local $/ = "\0";
2899
2900 open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2901 or return undef;
2902 while (my $line = <$fd>) {
2903 chomp $line;
2904
2905 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423 gitweb'
2906 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f gitweb/README'
2907 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2908 close $fd;
2909 return $1;
2910 }
2911 }
2912 close $fd;
2913 return undef;
2914 }
2915
2916 ## ......................................................................
2917 ## git utility functions, directly accessing git repository
2918
2919 # get the value of config variable either from file named as the variable
2920 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2921 # configuration variable in the repository config file.
2922 sub git_get_file_or_project_config {
2923 my ($path, $name) = @_;
2924
2925 $git_dir = "$projectroot/$path";
2926 open my $fd, '<', "$git_dir/$name"
2927 or return git_get_project_config($name);
2928 my $conf = <$fd>;
2929 close $fd;
2930 if (defined $conf) {
2931 chomp $conf;
2932 }
2933 return $conf;
2934 }
2935
2936 sub git_get_project_description {
2937 my $path = shift;
2938 return git_get_file_or_project_config($path, 'description');
2939 }
2940
2941 sub git_get_project_category {
2942 my $path = shift;
2943 return git_get_file_or_project_config($path, 'category');
2944 }
2945
2946
2947 # supported formats:
2948 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2949 # - if its contents is a number, use it as tag weight,
2950 # - otherwise add a tag with weight 1
2951 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2952 # the same value multiple times increases tag weight
2953 # * `gitweb.ctag' multi-valued repo config variable
2954 sub git_get_project_ctags {
2955 my $project = shift;
2956 my $ctags = {};
2957
2958 $git_dir = "$projectroot/$project";
2959 if (opendir my $dh, "$git_dir/ctags") {
2960 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2961 foreach my $tagfile (@files) {
2962 open my $ct, '<', $tagfile
2963 or next;
2964 my $val = <$ct>;
2965 chomp $val if $val;
2966 close $ct;
2967
2968 (my $ctag = $tagfile) =~ s#.*/##;
2969 if ($val =~ /^\d+$/) {
2970 $ctags->{$ctag} = $val;
2971 } else {
2972 $ctags->{$ctag} = 1;
2973 }
2974 }
2975 closedir $dh;
2976
2977 } elsif (open my $fh, '<', "$git_dir/ctags") {
2978 while (my $line = <$fh>) {
2979 chomp $line;
2980 $ctags->{$line}++ if $line;
2981 }
2982 close $fh;
2983
2984 } else {
2985 my $taglist = config_to_multi(git_get_project_config('ctag'));
2986 foreach my $tag (@$taglist) {
2987 $ctags->{$tag}++;
2988 }
2989 }
2990
2991 return $ctags;
2992 }
2993
2994 # return hash, where keys are content tags ('ctags'),
2995 # and values are sum of weights of given tag in every project
2996 sub git_gather_all_ctags {
2997 my $projects = shift;
2998 my $ctags = {};
2999
3000 foreach my $p (@$projects) {
3001 foreach my $ct (keys %{$p->{'ctags'}}) {
3002 $ctags->{$ct} += $p->{'ctags'}->{$ct};
3003 }
3004 }
3005
3006 return $ctags;
3007 }
3008
3009 sub git_populate_project_tagcloud {
3010 my $ctags = shift;
3011
3012 # First, merge different-cased tags; tags vote on casing
3013 my %ctags_lc;
3014 foreach (keys %$ctags) {
3015 $ctags_lc{lc $_}->{count} += $ctags->{$_};
3016 if (not $ctags_lc{lc $_}->{topcount}
3017 or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
3018 $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
3019 $ctags_lc{lc $_}->{topname} = $_;
3020 }
3021 }
3022
3023 my $cloud;
3024 my $matched = $input_params{'ctag'};
3025 if (eval { require HTML::TagCloud; 1; }) {
3026 $cloud = HTML::TagCloud->new;
3027 foreach my $ctag (sort keys %ctags_lc) {
3028 # Pad the title with spaces so that the cloud looks
3029 # less crammed.
3030 my $title = esc_html($ctags_lc{$ctag}->{topname});
3031 $title =~ s/ /&nbsp;/g;
3032 $title =~ s/^/&nbsp;/g;
3033 $title =~ s/$/&nbsp;/g;
3034 if (defined $matched && $matched eq $ctag) {
3035 $title = qq(<mark>$title</mark>);
3036 }
3037 $cloud->add($title, href(project=>undef, ctag=>$ctag),
3038 $ctags_lc{$ctag}->{count});
3039 }
3040 } else {
3041 $cloud = {};
3042 foreach my $ctag (keys %ctags_lc) {
3043 my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
3044 if (defined $matched && $matched eq $ctag) {
3045 $title = qq(<mark>$title</mark>);
3046 }
3047 $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
3048 $cloud->{$ctag}{ctag} =
3049 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
3050 }
3051 }
3052 return $cloud;
3053 }
3054
3055 sub git_show_project_tagcloud {
3056 my ($cloud, $count) = @_;
3057 if (ref $cloud eq 'HTML::TagCloud') {
3058 return $cloud->html_and_css($count);
3059 } else {
3060 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
3061 return
3062 '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
3063 join (', ', map {
3064 $cloud->{$_}->{'ctag'}
3065 } splice(@tags, 0, $count)) .
3066 '</div>';
3067 }
3068 }
3069
3070 sub git_get_project_url_list {
3071 my $path = shift;
3072
3073 $git_dir = "$projectroot/$path";
3074 open my $fd, '<', "$git_dir/cloneurl"
3075 or return wantarray ?
3076 @{ config_to_multi(git_get_project_config('url')) } :
3077 config_to_multi(git_get_project_config('url'));
3078 my @git_project_url_list = map { chomp; $_ } <$fd>;
3079 close $fd;
3080
3081 return wantarray ? @git_project_url_list : \@git_project_url_list;
3082 }
3083
3084 sub git_get_projects_list {
3085 my $filter = shift || '';
3086 my $paranoid = shift;
3087 my @list;
3088
3089 if (-d $projects_list) {
3090 # search in directory
3091 my $dir = $projects_list;
3092 # remove the trailing "/"
3093 $dir =~ s!/+$!!;
3094 my $pfxlen = length("$dir");
3095 my $pfxdepth = ($dir =~ tr!/!!);
3096 # when filtering, search only given subdirectory
3097 if ($filter && !$paranoid) {
3098 $dir .= "/$filter";
3099 $dir =~ s!/+$!!;
3100 }
3101
3102 File::Find::find({
3103 follow_fast => 1, # follow symbolic links
3104 follow_skip => 2, # ignore duplicates
3105 dangling_symlinks => 0, # ignore dangling symlinks, silently
3106 wanted => sub {
3107 # global variables
3108 our $project_maxdepth;
3109 our $projectroot;
3110 # skip project-list toplevel, if we get it.
3111 return if (m!^[/.]$!);
3112 # only directories can be git repositories
3113 return unless (-d $_);
3114 # need search permission
3115 return unless (-x $_);
3116 # don't traverse too deep (Find is super slow on os x)
3117 # $project_maxdepth excludes depth of $projectroot
3118 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
3119 $File::Find::prune = 1;
3120 return;
3121 }
3122
3123 my $path = substr($File::Find::name, $pfxlen + 1);
3124 # paranoidly only filter here
3125 if ($paranoid && $filter && $path !~ m!^\Q$filter\E/!) {
3126 next;
3127 }
3128 # we check related file in $projectroot
3129 if (check_export_ok("$projectroot/$path")) {
3130 push @list, { path => $path };
3131 $File::Find::prune = 1;
3132 }
3133 },
3134 }, "$dir");
3135
3136 } elsif (-f $projects_list) {
3137 # read from file(url-encoded):
3138 # 'git%2Fgit.git Linus+Torvalds'
3139 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3140 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3141 open my $fd, '<', $projects_list or return;
3142 PROJECT:
3143 while (my $line = <$fd>) {
3144 chomp $line;
3145 my ($path, $owner) = split ' ', $line;
3146 $path = unescape($path);
3147 $owner = unescape($owner);
3148 if (!defined $path) {
3149 next;
3150 }
3151 # if $filter is rpovided, check if $path begins with $filter
3152 if ($filter && $path !~ m!^\Q$filter\E/!) {
3153 next;
3154 }
3155 if (check_export_ok("$projectroot/$path")) {
3156 my $pr = {
3157 path => $path
3158 };
3159 if ($owner) {
3160 $pr->{'owner'} = to_utf8($owner);
3161 }
3162 push @list, $pr;
3163 }
3164 }
3165 close $fd;
3166 }
3167 return @list;
3168 }
3169
3170 # written with help of Tree::Trie module (Perl Artistic License, GPL compatible)
3171 # as side effects it sets 'forks' field to list of forks for forked projects
3172 sub filter_forks_from_projects_list {
3173 my $projects = shift;
3174
3175 my %trie; # prefix tree of directories (path components)
3176 # generate trie out of those directories that might contain forks
3177 foreach my $pr (@$projects) {
3178 my $path = $pr->{'path'};
3179 $path =~ s/\.git$//; # forks of 'repo.git' are in 'repo/' directory
3180 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
3181 next unless ($path); # skip '.git' repository: tests, git-instaweb
3182 next unless (-d "$projectroot/$path"); # containing directory exists
3183 $pr->{'forks'} = []; # there can be 0 or more forks of project
3184
3185 # add to trie
3186 my @dirs = split('/', $path);
3187 # walk the trie, until either runs out of components or out of trie
3188 my $ref = \%trie;
3189 while (scalar @dirs &&
3190 exists($ref->{$dirs[0]})) {
3191 $ref = $ref->{shift @dirs};
3192 }
3193 # create rest of trie structure from rest of components
3194 foreach my $dir (@dirs) {
3195 $ref = $ref->{$dir} = {};
3196 }
3197 # create end marker, store $pr as a data
3198 $ref->{''} = $pr if (!exists $ref->{''});
3199 }
3200
3201 # filter out forks, by finding shortest prefix match for paths
3202 my @filtered;
3203 PROJECT:
3204 foreach my $pr (@$projects) {
3205 # trie lookup
3206 my $ref = \%trie;
3207 DIR:
3208 foreach my $dir (split('/', $pr->{'path'})) {
3209 if (exists $ref->{''}) {
3210 # found [shortest] prefix, is a fork - skip it
3211 push @{$ref->{''}{'forks'}}, $pr;
3212 next PROJECT;
3213 }
3214 if (!exists $ref->{$dir}) {
3215 # not in trie, cannot have prefix, not a fork
3216 push @filtered, $pr;
3217 next PROJECT;
3218 }
3219 # If the dir is there, we just walk one step down the trie.
3220 $ref = $ref->{$dir};
3221 }
3222 # we ran out of trie
3223 # (shouldn't happen: it's either no match, or end marker)
3224 push @filtered, $pr;
3225 }
3226
3227 return @filtered;
3228 }
3229
3230 # note: fill_project_list_info must be run first,
3231 # for 'descr_long' and 'ctags' to be filled
3232 sub search_projects_list {
3233 my ($projlist, %opts) = @_;
3234 my $tagfilter = $opts{'tagfilter'};
3235 my $search_re = $opts{'search_regexp'};
3236
3237 return @$projlist
3238 unless ($tagfilter || $search_re);
3239
3240 # searching projects require filling to be run before it;
3241 fill_project_list_info($projlist,
3242 $tagfilter ? 'ctags' : (),
3243 $search_re ? ('path', 'descr') : ());
3244 my @projects;
3245 PROJECT:
3246 foreach my $pr (@$projlist) {
3247
3248 if ($tagfilter) {
3249 next unless ref($pr->{'ctags'}) eq 'HASH';
3250 next unless
3251 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
3252 }
3253
3254 if ($search_re) {
3255 next unless
3256 $pr->{'path'} =~ /$search_re/ ||
3257 $pr->{'descr_long'} =~ /$search_re/;
3258 }
3259
3260 push @projects, $pr;
3261 }
3262
3263 return @projects;
3264 }
3265
3266 our $gitweb_project_owner = undef;
3267 sub git_get_project_list_from_file {
3268
3269 return if (defined $gitweb_project_owner);
3270
3271 $gitweb_project_owner = {};
3272 # read from file (url-encoded):
3273 # 'git%2Fgit.git Linus+Torvalds'
3274 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3275 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3276 if (-f $projects_list) {
3277 open(my $fd, '<', $projects_list);
3278 while (my $line = <$fd>) {
3279 chomp $line;
3280 my ($pr, $ow) = split ' ', $line;
3281 $pr = unescape($pr);
3282 $ow = unescape($ow);
3283 $gitweb_project_owner->{$pr} = to_utf8($ow);
3284 }
3285 close $fd;
3286 }
3287 }
3288
3289 sub git_get_project_owner {
3290 my $project = shift;
3291 my $owner;
3292
3293 return undef unless $project;
3294 $git_dir = "$projectroot/$project";
3295
3296 if (!defined $gitweb_project_owner) {
3297 git_get_project_list_from_file();
3298 }
3299
3300 if (exists $gitweb_project_owner->{$project}) {
3301 $owner = $gitweb_project_owner->{$project};
3302 }
3303 if (!defined $owner){
3304 $owner = git_get_project_config('owner');
3305 }
3306 if (!defined $owner) {
3307 $owner = get_file_owner("$git_dir");
3308 }
3309
3310 return $owner;
3311 }
3312
3313 sub git_get_last_activity {
3314 my ($path) = @_;
3315 my $fd;
3316
3317 $git_dir = "$projectroot/$path";
3318 open($fd, "-|", git_cmd(), 'for-each-ref',
3319 '--format=%(committer)',
3320 '--sort=-committerdate',
3321 '--count=1',
3322 map { "refs/$_" } get_branch_refs ()) or return;
3323 my $most_recent = <$fd>;
3324 close $fd or return;
3325 if (defined $most_recent &&
3326 $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
3327 my $timestamp = $1;
3328 my $age = time - $timestamp;
3329 return ($age, age_string($age));
3330 }
3331 return (undef, undef);
3332 }
3333
3334 # Implementation note: when a single remote is wanted, we cannot use 'git
3335 # remote show -n' because that command always work (assuming it's a remote URL
3336 # if it's not defined), and we cannot use 'git remote show' because that would
3337 # try to make a network roundtrip. So the only way to find if that particular
3338 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3339 # and when we find what we want.
3340 sub git_get_remotes_list {
3341 my $wanted = shift;
3342 my %remotes = ();
3343
3344 open my $fd, '-|' , git_cmd(), 'remote', '-v';
3345 return unless $fd;
3346 while (my $remote = <$fd>) {
3347 chomp $remote;
3348 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3349 next if $wanted and not $remote eq $wanted;
3350 my ($url, $key) = ($1, $2);
3351
3352 $remotes{$remote} ||= { 'heads' => () };
3353 $remotes{$remote}{$key} = $url;
3354 }
3355 close $fd or return;
3356 return wantarray ? %remotes : \%remotes;
3357 }
3358
3359 # Takes a hash of remotes as first parameter and fills it by adding the
3360 # available remote heads for each of the indicated remotes.
3361 sub fill_remote_heads {
3362 my $remotes = shift;
3363 my @heads = map { "remotes/$_" } keys %$remotes;
3364 my @remoteheads = git_get_heads_list(undef, @heads);
3365 foreach my $remote (keys %$remotes) {
3366 $remotes->{$remote}{'heads'} = [ grep {
3367 $_->{'name'} =~ s!^$remote/!!
3368 } @remoteheads ];
3369 }
3370 }
3371
3372 sub git_get_references {
3373 my $type = shift || "";
3374 my %refs;
3375 # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3376 # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3377 open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3378 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3379 or return;
3380
3381 while (my $line = <$fd>) {
3382 chomp $line;
3383 if ($line =~ m!^($oid_regex)\srefs/($type.*)$!) {
3384 if (defined $refs{$1}) {
3385 push @{$refs{$1}}, $2;
3386 } else {
3387 $refs{$1} = [ $2 ];
3388 }
3389 }
3390 }
3391 close $fd or return;
3392 return \%refs;
3393 }
3394
3395 sub git_get_rev_name_tags {
3396 my $hash = shift || return undef;
3397
3398 open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3399 or return;
3400 my $name_rev = <$fd>;
3401 close $fd;
3402
3403 if ($name_rev =~ m|^$hash tags/(.*)$|) {
3404 return $1;
3405 } else {
3406 # catches also '$hash undefined' output
3407 return undef;
3408 }
3409 }
3410
3411 ## ----------------------------------------------------------------------
3412 ## parse to hash functions
3413
3414 sub parse_date {
3415 my $epoch = shift;
3416 my $tz = shift || "-0000";
3417
3418 my %date;
3419 my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3420 my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3421 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3422 $date{'hour'} = $hour;
3423 $date{'minute'} = $min;
3424 $date{'mday'} = $mday;
3425 $date{'day'} = $days[$wday];
3426 $date{'month'} = $months[$mon];
3427 $date{'rfc2822'} = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3428 $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3429 $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3430 $mday, $months[$mon], $hour ,$min;
3431 $date{'iso-8601'} = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3432 1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3433
3434 my ($tz_sign, $tz_hour, $tz_min) =
3435 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3436 $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3437 my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3438 ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3439 $date{'hour_local'} = $hour;
3440 $date{'minute_local'} = $min;
3441 $date{'tz_local'} = $tz;
3442 $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3443 1900+$year, $mon+1, $mday,
3444 $hour, $min, $sec, $tz);
3445 return %date;
3446 }
3447
3448 sub hide_mailaddrs_if_private {
3449 my $line = shift;
3450 return $line unless gitweb_check_feature('email-privacy');
3451 $line =~ s/<[^@>]+@[^>]+>/<redacted>/g;
3452 return $line;
3453 }
3454
3455 sub parse_tag {
3456 my $tag_id = shift;
3457 my %tag;
3458 my @comment;
3459
3460 open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3461 $tag{'id'} = $tag_id;
3462 while (my $line = <$fd>) {
3463 chomp $line;
3464 if ($line =~ m/^object ($oid_regex)$/) {
3465 $tag{'object'} = $1;
3466 } elsif ($line =~ m/^type (.+)$/) {
3467 $tag{'type'} = $1;
3468 } elsif ($line =~ m/^tag (.+)$/) {
3469 $tag{'name'} = $1;
3470 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3471 $tag{'author'} = hide_mailaddrs_if_private($1);
3472 $tag{'author_epoch'} = $2;
3473 $tag{'author_tz'} = $3;
3474 if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3475 $tag{'author_name'} = $1;
3476 $tag{'author_email'} = $2;
3477 } else {
3478 $tag{'author_name'} = $tag{'author'};
3479 }
3480 } elsif ($line =~ m/--BEGIN/) {
3481 push @comment, $line;
3482 last;
3483 } elsif ($line eq "") {
3484 last;
3485 }
3486 }
3487 push @comment, <$fd>;
3488 $tag{'comment'} = \@comment;
3489 close $fd or return;
3490 if (!defined $tag{'name'}) {
3491 return
3492 };
3493 return %tag
3494 }
3495
3496 sub parse_commit_text {
3497 my ($commit_text, $withparents) = @_;
3498 my @commit_lines = split '\n', $commit_text;
3499 my %co;
3500
3501 pop @commit_lines; # Remove '\0'
3502
3503 if (! @commit_lines) {
3504 return;
3505 }
3506
3507 my $header = shift @commit_lines;
3508 if ($header !~ m/^$oid_regex/) {
3509 return;
3510 }
3511 ($co{'id'}, my @parents) = split ' ', $header;
3512 while (my $line = shift @commit_lines) {
3513 last if $line eq "\n";
3514 if ($line =~ m/^tree ($oid_regex)$/) {
3515 $co{'tree'} = $1;
3516 } elsif ((!defined $withparents) && ($line =~ m/^parent ($oid_regex)$/)) {
3517 push @parents, $1;
3518 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3519 $co{'author'} = hide_mailaddrs_if_private(to_utf8($1));
3520 $co{'author_epoch'} = $2;
3521 $co{'author_tz'} = $3;
3522 if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3523 $co{'author_name'} = $1;
3524 $co{'author_email'} = $2;
3525 } else {
3526 $co{'author_name'} = $co{'author'};
3527 }
3528 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3529 $co{'committer'} = hide_mailaddrs_if_private(to_utf8($1));
3530 $co{'committer_epoch'} = $2;
3531 $co{'committer_tz'} = $3;
3532 if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3533 $co{'committer_name'} = $1;
3534 $co{'committer_email'} = $2;
3535 } else {
3536 $co{'committer_name'} = $co{'committer'};
3537 }
3538 }
3539 }
3540 if (!defined $co{'tree'}) {
3541 return;
3542 };
3543 $co{'parents'} = \@parents;
3544 $co{'parent'} = $parents[0];
3545
3546 foreach my $title (@commit_lines) {
3547 $title =~ s/^ //;
3548 if ($title ne "") {
3549 $co{'title'} = chop_str($title, 80, 5);
3550 # remove leading stuff of merges to make the interesting part visible
3551 if (length($title) > 50) {
3552 $title =~ s/^Automatic //;
3553 $title =~ s/^merge (of|with) /Merge ... /i;
3554 if (length($title) > 50) {
3555 $title =~ s/(http|rsync):\/\///;
3556 }
3557 if (length($title) > 50) {
3558 $title =~ s/(master|www|rsync)\.//;
3559 }
3560 if (length($title) > 50) {
3561 $title =~ s/kernel.org:?//;
3562 }
3563 if (length($title) > 50) {
3564 $title =~ s/\/pub\/scm//;
3565 }
3566 }
3567 $co{'title_short'} = chop_str($title, 50, 5);
3568 last;
3569 }
3570 }
3571 if (! defined $co{'title'} || $co{'title'} eq "") {
3572 $co{'title'} = $co{'title_short'} = '(no commit message)';
3573 }
3574 # remove added spaces, redact e-mail addresses if applicable.
3575 foreach my $line (@commit_lines) {
3576 $line =~ s/^ //;
3577 $line = hide_mailaddrs_if_private($line);
3578 }
3579 $co{'comment'} = \@commit_lines;
3580
3581 my $age = time - $co{'committer_epoch'};
3582 $co{'age'} = $age;
3583 $co{'age_string'} = age_string($age);
3584 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3585 $co{'age_string_iso8601'} = sprintf "%4i-%02u-%02i %02u:%02u:%02uZ", 1900 + $year, $mon+1, $mday, $hour, $min, $sec;
3586 if ($age > 60*60*24*7*2) {
3587 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3588 $co{'age_string_age'} = "$co{'age_string_iso8601'} ($co{'age_string'})";
3589 } else {
3590 $co{'age_string_date'} = $co{'age_string'};
3591 $co{'age_string_age'} = $co{'age_string_iso8601'};
3592 }
3593 return %co;
3594 }
3595
3596 sub parse_commit {
3597 my ($commit_id) = @_;
3598 my %co;
3599
3600 local $/ = "\0";
3601
3602 open my $fd, "-|", git_cmd(), "rev-list",
3603 "--parents",
3604 "--header",
3605 "--max-count=1",
3606 $commit_id,
3607 "--",
3608 or die_error(500, "Open git-rev-list failed");
3609 %co = parse_commit_text(<$fd>, 1);
3610 close $fd;
3611
3612 return %co;
3613 }
3614
3615 sub parse_commits {
3616 my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3617 my @cos;
3618
3619 $maxcount ||= 1;
3620 $skip ||= 0;
3621
3622 local $/ = "\0";
3623
3624 open my $fd, "-|", git_cmd(), "rev-list",
3625 "--header",
3626 @args,
3627 ("--max-count=" . $maxcount),
3628 ("--skip=" . $skip),
3629 @extra_options,
3630 $commit_id,
3631 "--",
3632 ($filename ? ($filename) : ())
3633 or die_error(500, "Open git-rev-list failed");
3634 while (my $line = <$fd>) {
3635 my %co = parse_commit_text($line);
3636 push @cos, \%co;
3637 }
3638 close $fd;
3639
3640 return wantarray ? @cos : \@cos;
3641 }
3642
3643 # parse line of git-diff-tree "raw" output
3644 sub parse_difftree_raw_line {
3645 my $line = shift;
3646 my %res;
3647
3648 # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M ls-files.c'
3649 # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M rev-tree.c'
3650 if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ($oid_regex) ($oid_regex) (.)([0-9]{0,3})\t(.*)$/) {
3651 $res{'from_mode'} = $1;
3652 $res{'to_mode'} = $2;
3653 $res{'from_id'} = $3;
3654 $res{'to_id'} = $4;
3655 $res{'status'} = $5;
3656 $res{'similarity'} = $6;
3657 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3658 ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3659 } else {
3660 $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3661 }
3662 }
3663 # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3664 # combined diff (for merge commit)
3665 elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:$oid_regex )+)([a-zA-Z]+)\t(.*)$//) {
3666 $res{'nparents'} = length($1);
3667 $res{'from_mode'} = [ split(' ', $2) ];
3668 $res{'to_mode'} = pop @{$res{'from_mode'}};
3669 $res{'from_id'} = [ split(' ', $3) ];
3670 $res{'to_id'} = pop @{$res{'from_id'}};
3671 $res{'status'} = [ split('', $4) ];
3672 $res{'to_file'} = unquote($5);
3673 }
3674 # 'c512b523472485aef4fff9e57b229d9d243c967f'
3675 elsif ($line =~ m/^($oid_regex)$/) {
3676 $res{'commit'} = $1;
3677 }
3678
3679 return wantarray ? %res : \%res;
3680 }
3681
3682 # wrapper: return parsed line of git-diff-tree "raw" output
3683 # (the argument might be raw line, or parsed info)
3684 sub parsed_difftree_line {
3685 my $line_or_ref = shift;
3686
3687 if (ref($line_or_ref) eq "HASH") {
3688 # pre-parsed (or generated by hand)
3689 return $line_or_ref;
3690 } else {
3691 return parse_difftree_raw_line($line_or_ref);
3692 }
3693 }
3694
3695 # parse line of git-ls-tree output
3696 sub parse_ls_tree_line {
3697 my $line = shift;
3698 my %opts = @_;
3699 my %res;
3700
3701 if ($opts{'-l'}) {
3702 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa 16717 panic.c'
3703 $line =~ m/^([0-9]+) (.+) ($oid_regex) +(-|[0-9]+)\t(.+)$/s;
3704
3705 $res{'mode'} = $1;
3706 $res{'type'} = $2;
3707 $res{'hash'} = $3;
3708 $res{'size'} = $4;
3709 if ($opts{'-z'}) {
3710 $res{'name'} = $5;
3711 } else {
3712 $res{'name'} = unquote($5);
3713 }
3714 } else {
3715 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
3716 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t(.+)$/s;
3717
3718 $res{'mode'} = $1;
3719 $res{'type'} = $2;
3720 $res{'hash'} = $3;
3721 if ($opts{'-z'}) {
3722 $res{'name'} = $4;
3723 } else {
3724 $res{'name'} = unquote($4);
3725 }
3726 }
3727
3728 return wantarray ? %res : \%res;
3729 }
3730
3731 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3732 sub parse_from_to_diffinfo {
3733 my ($diffinfo, $from, $to, @parents) = @_;
3734
3735 if ($diffinfo->{'nparents'}) {
3736 # combined diff
3737 $from->{'file'} = [];
3738 $from->{'href'} = [];
3739 fill_from_file_info($diffinfo, @parents)
3740 unless exists $diffinfo->{'from_file'};
3741 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3742 $from->{'file'}[$i] =
3743 defined $diffinfo->{'from_file'}[$i] ?
3744 $diffinfo->{'from_file'}[$i] :
3745 $diffinfo->{'to_file'};
3746 if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3747 $from->{'href'}[$i] = href(action=>"blob",
3748 hash_base=>$parents[$i],
3749 hash=>$diffinfo->{'from_id'}[$i],
3750 file_name=>$from->{'file'}[$i]);