Sync with 1.7.3.4
[git/git.git] / gitweb / gitweb.perl
1 #!/usr/bin/perl
2
3 # gitweb - simple web interface to track changes in git repositories
4 #
5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
6 # (C) 2005, Christian Gierke
7 #
8 # This program is licensed under the GPLv2
9
10 use 5.008;
11 use strict;
12 use warnings;
13 use CGI qw(:standard :escapeHTML -nosticky);
14 use CGI::Util qw(unescape);
15 use CGI::Carp qw(fatalsToBrowser set_message);
16 use Encode;
17 use Fcntl ':mode';
18 use File::Find qw();
19 use File::Basename qw(basename);
20 use Time::HiRes qw(gettimeofday tv_interval);
21 binmode STDOUT, ':utf8';
22
23 our $t0 = [ gettimeofday() ];
24 our $number_of_git_cmds = 0;
25
26 BEGIN {
27 CGI->compile() if $ENV{'MOD_PERL'};
28 }
29
30 our $version = "++GIT_VERSION++";
31
32 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
33 sub evaluate_uri {
34 our $cgi;
35
36 our $my_url = $cgi->url();
37 our $my_uri = $cgi->url(-absolute => 1);
38
39 # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
40 # needed and used only for URLs with nonempty PATH_INFO
41 our $base_url = $my_url;
42
43 # When the script is used as DirectoryIndex, the URL does not contain the name
44 # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
45 # have to do it ourselves. We make $path_info global because it's also used
46 # later on.
47 #
48 # Another issue with the script being the DirectoryIndex is that the resulting
49 # $my_url data is not the full script URL: this is good, because we want
50 # generated links to keep implying the script name if it wasn't explicitly
51 # indicated in the URL we're handling, but it means that $my_url cannot be used
52 # as base URL.
53 # Therefore, if we needed to strip PATH_INFO, then we know that we have
54 # to build the base URL ourselves:
55 our $path_info = $ENV{"PATH_INFO"};
56 if ($path_info) {
57 if ($my_url =~ s,\Q$path_info\E$,, &&
58 $my_uri =~ s,\Q$path_info\E$,, &&
59 defined $ENV{'SCRIPT_NAME'}) {
60 $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
61 }
62 }
63
64 # target of the home link on top of all pages
65 our $home_link = $my_uri || "/";
66 }
67
68 # core git executable to use
69 # this can just be "git" if your webserver has a sensible PATH
70 our $GIT = "++GIT_BINDIR++/git";
71
72 # absolute fs-path which will be prepended to the project path
73 #our $projectroot = "/pub/scm";
74 our $projectroot = "++GITWEB_PROJECTROOT++";
75
76 # fs traversing limit for getting project list
77 # the number is relative to the projectroot
78 our $project_maxdepth = "++GITWEB_PROJECT_MAXDEPTH++";
79
80 # string of the home link on top of all pages
81 our $home_link_str = "++GITWEB_HOME_LINK_STR++";
82
83 # name of your site or organization to appear in page titles
84 # replace this with something more descriptive for clearer bookmarks
85 our $site_name = "++GITWEB_SITENAME++"
86 || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
87
88 # filename of html text to include at top of each page
89 our $site_header = "++GITWEB_SITE_HEADER++";
90 # html text to include at home page
91 our $home_text = "++GITWEB_HOMETEXT++";
92 # filename of html text to include at bottom of each page
93 our $site_footer = "++GITWEB_SITE_FOOTER++";
94
95 # URI of stylesheets
96 our @stylesheets = ("++GITWEB_CSS++");
97 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
98 our $stylesheet = undef;
99 # URI of GIT logo (72x27 size)
100 our $logo = "++GITWEB_LOGO++";
101 # URI of GIT favicon, assumed to be image/png type
102 our $favicon = "++GITWEB_FAVICON++";
103 # URI of gitweb.js (JavaScript code for gitweb)
104 our $javascript = "++GITWEB_JS++";
105
106 # URI and label (title) of GIT logo link
107 #our $logo_url = "http://www.kernel.org/pub/software/scm/git/docs/";
108 #our $logo_label = "git documentation";
109 our $logo_url = "http://git-scm.com/";
110 our $logo_label = "git homepage";
111
112 # source of projects list
113 our $projects_list = "++GITWEB_LIST++";
114
115 # the width (in characters) of the projects list "Description" column
116 our $projects_list_description_width = 25;
117
118 # default order of projects list
119 # valid values are none, project, descr, owner, and age
120 our $default_projects_order = "project";
121
122 # show repository only if this file exists
123 # (only effective if this variable evaluates to true)
124 our $export_ok = "++GITWEB_EXPORT_OK++";
125
126 # show repository only if this subroutine returns true
127 # when given the path to the project, for example:
128 # sub { return -e "$_[0]/git-daemon-export-ok"; }
129 our $export_auth_hook = undef;
130
131 # only allow viewing of repositories also shown on the overview page
132 our $strict_export = "++GITWEB_STRICT_EXPORT++";
133
134 # list of git base URLs used for URL to where fetch project from,
135 # i.e. full URL is "$git_base_url/$project"
136 our @git_base_url_list = grep { $_ ne '' } ("++GITWEB_BASE_URL++");
137
138 # default blob_plain mimetype and default charset for text/plain blob
139 our $default_blob_plain_mimetype = 'text/plain';
140 our $default_text_plain_charset = undef;
141
142 # file to use for guessing MIME types before trying /etc/mime.types
143 # (relative to the current git repository)
144 our $mimetypes_file = undef;
145
146 # assume this charset if line contains non-UTF-8 characters;
147 # it should be valid encoding (see Encoding::Supported(3pm) for list),
148 # for which encoding all byte sequences are valid, for example
149 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
150 # could be even 'utf-8' for the old behavior)
151 our $fallback_encoding = 'latin1';
152
153 # rename detection options for git-diff and git-diff-tree
154 # - default is '-M', with the cost proportional to
155 # (number of removed files) * (number of new files).
156 # - more costly is '-C' (which implies '-M'), with the cost proportional to
157 # (number of changed files + number of removed files) * (number of new files)
158 # - even more costly is '-C', '--find-copies-harder' with cost
159 # (number of files in the original tree) * (number of new files)
160 # - one might want to include '-B' option, e.g. '-B', '-M'
161 our @diff_opts = ('-M'); # taken from git_commit
162
163 # Disables features that would allow repository owners to inject script into
164 # the gitweb domain.
165 our $prevent_xss = 0;
166
167 # Path to the highlight executable to use (must be the one from
168 # http://www.andre-simon.de due to assumptions about parameters and output).
169 # Useful if highlight is not installed on your webserver's PATH.
170 # [Default: highlight]
171 our $highlight_bin = "++HIGHLIGHT_BIN++";
172
173 # information about snapshot formats that gitweb is capable of serving
174 our %known_snapshot_formats = (
175 # name => {
176 # 'display' => display name,
177 # 'type' => mime type,
178 # 'suffix' => filename suffix,
179 # 'format' => --format for git-archive,
180 # 'compressor' => [compressor command and arguments]
181 # (array reference, optional)
182 # 'disabled' => boolean (optional)}
183 #
184 'tgz' => {
185 'display' => 'tar.gz',
186 'type' => 'application/x-gzip',
187 'suffix' => '.tar.gz',
188 'format' => 'tar',
189 'compressor' => ['gzip']},
190
191 'tbz2' => {
192 'display' => 'tar.bz2',
193 'type' => 'application/x-bzip2',
194 'suffix' => '.tar.bz2',
195 'format' => 'tar',
196 'compressor' => ['bzip2']},
197
198 'txz' => {
199 'display' => 'tar.xz',
200 'type' => 'application/x-xz',
201 'suffix' => '.tar.xz',
202 'format' => 'tar',
203 'compressor' => ['xz'],
204 'disabled' => 1},
205
206 'zip' => {
207 'display' => 'zip',
208 'type' => 'application/x-zip',
209 'suffix' => '.zip',
210 'format' => 'zip'},
211 );
212
213 # Aliases so we understand old gitweb.snapshot values in repository
214 # configuration.
215 our %known_snapshot_format_aliases = (
216 'gzip' => 'tgz',
217 'bzip2' => 'tbz2',
218 'xz' => 'txz',
219
220 # backward compatibility: legacy gitweb config support
221 'x-gzip' => undef, 'gz' => undef,
222 'x-bzip2' => undef, 'bz2' => undef,
223 'x-zip' => undef, '' => undef,
224 );
225
226 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
227 # are changed, it may be appropriate to change these values too via
228 # $GITWEB_CONFIG.
229 our %avatar_size = (
230 'default' => 16,
231 'double' => 32
232 );
233
234 # Used to set the maximum load that we will still respond to gitweb queries.
235 # If server load exceed this value then return "503 server busy" error.
236 # If gitweb cannot determined server load, it is taken to be 0.
237 # Leave it undefined (or set to 'undef') to turn off load checking.
238 our $maxload = 300;
239
240 # configuration for 'highlight' (http://www.andre-simon.de/)
241 # match by basename
242 our %highlight_basename = (
243 #'Program' => 'py',
244 #'Library' => 'py',
245 'SConstruct' => 'py', # SCons equivalent of Makefile
246 'Makefile' => 'make',
247 );
248 # match by extension
249 our %highlight_ext = (
250 # main extensions, defining name of syntax;
251 # see files in /usr/share/highlight/langDefs/ directory
252 map { $_ => $_ }
253 qw(py c cpp rb java css php sh pl js tex bib xml awk bat ini spec tcl),
254 # alternate extensions, see /etc/highlight/filetypes.conf
255 'h' => 'c',
256 map { $_ => 'cpp' } qw(cxx c++ cc),
257 map { $_ => 'php' } qw(php3 php4),
258 map { $_ => 'pl' } qw(perl pm), # perhaps also 'cgi'
259 'mak' => 'make',
260 map { $_ => 'xml' } qw(xhtml html htm),
261 );
262
263 # You define site-wide feature defaults here; override them with
264 # $GITWEB_CONFIG as necessary.
265 our %feature = (
266 # feature => {
267 # 'sub' => feature-sub (subroutine),
268 # 'override' => allow-override (boolean),
269 # 'default' => [ default options...] (array reference)}
270 #
271 # if feature is overridable (it means that allow-override has true value),
272 # then feature-sub will be called with default options as parameters;
273 # return value of feature-sub indicates if to enable specified feature
274 #
275 # if there is no 'sub' key (no feature-sub), then feature cannot be
276 # overridden
277 #
278 # use gitweb_get_feature(<feature>) to retrieve the <feature> value
279 # (an array) or gitweb_check_feature(<feature>) to check if <feature>
280 # is enabled
281
282 # Enable the 'blame' blob view, showing the last commit that modified
283 # each line in the file. This can be very CPU-intensive.
284
285 # To enable system wide have in $GITWEB_CONFIG
286 # $feature{'blame'}{'default'} = [1];
287 # To have project specific config enable override in $GITWEB_CONFIG
288 # $feature{'blame'}{'override'} = 1;
289 # and in project config gitweb.blame = 0|1;
290 'blame' => {
291 'sub' => sub { feature_bool('blame', @_) },
292 'override' => 0,
293 'default' => [0]},
294
295 # Enable the 'snapshot' link, providing a compressed archive of any
296 # tree. This can potentially generate high traffic if you have large
297 # project.
298
299 # Value is a list of formats defined in %known_snapshot_formats that
300 # you wish to offer.
301 # To disable system wide have in $GITWEB_CONFIG
302 # $feature{'snapshot'}{'default'} = [];
303 # To have project specific config enable override in $GITWEB_CONFIG
304 # $feature{'snapshot'}{'override'} = 1;
305 # and in project config, a comma-separated list of formats or "none"
306 # to disable. Example: gitweb.snapshot = tbz2,zip;
307 'snapshot' => {
308 'sub' => \&feature_snapshot,
309 'override' => 0,
310 'default' => ['tgz']},
311
312 # Enable text search, which will list the commits which match author,
313 # committer or commit text to a given string. Enabled by default.
314 # Project specific override is not supported.
315 'search' => {
316 'override' => 0,
317 'default' => [1]},
318
319 # Enable grep search, which will list the files in currently selected
320 # tree containing the given string. Enabled by default. This can be
321 # potentially CPU-intensive, of course.
322
323 # To enable system wide have in $GITWEB_CONFIG
324 # $feature{'grep'}{'default'} = [1];
325 # To have project specific config enable override in $GITWEB_CONFIG
326 # $feature{'grep'}{'override'} = 1;
327 # and in project config gitweb.grep = 0|1;
328 'grep' => {
329 'sub' => sub { feature_bool('grep', @_) },
330 'override' => 0,
331 'default' => [1]},
332
333 # Enable the pickaxe search, which will list the commits that modified
334 # a given string in a file. This can be practical and quite faster
335 # alternative to 'blame', but still potentially CPU-intensive.
336
337 # To enable system wide have in $GITWEB_CONFIG
338 # $feature{'pickaxe'}{'default'} = [1];
339 # To have project specific config enable override in $GITWEB_CONFIG
340 # $feature{'pickaxe'}{'override'} = 1;
341 # and in project config gitweb.pickaxe = 0|1;
342 'pickaxe' => {
343 'sub' => sub { feature_bool('pickaxe', @_) },
344 'override' => 0,
345 'default' => [1]},
346
347 # Enable showing size of blobs in a 'tree' view, in a separate
348 # column, similar to what 'ls -l' does. This cost a bit of IO.
349
350 # To disable system wide have in $GITWEB_CONFIG
351 # $feature{'show-sizes'}{'default'} = [0];
352 # To have project specific config enable override in $GITWEB_CONFIG
353 # $feature{'show-sizes'}{'override'} = 1;
354 # and in project config gitweb.showsizes = 0|1;
355 'show-sizes' => {
356 'sub' => sub { feature_bool('showsizes', @_) },
357 'override' => 0,
358 'default' => [1]},
359
360 # Make gitweb use an alternative format of the URLs which can be
361 # more readable and natural-looking: project name is embedded
362 # directly in the path and the query string contains other
363 # auxiliary information. All gitweb installations recognize
364 # URL in either format; this configures in which formats gitweb
365 # generates links.
366
367 # To enable system wide have in $GITWEB_CONFIG
368 # $feature{'pathinfo'}{'default'} = [1];
369 # Project specific override is not supported.
370
371 # Note that you will need to change the default location of CSS,
372 # favicon, logo and possibly other files to an absolute URL. Also,
373 # if gitweb.cgi serves as your indexfile, you will need to force
374 # $my_uri to contain the script name in your $GITWEB_CONFIG.
375 'pathinfo' => {
376 'override' => 0,
377 'default' => [0]},
378
379 # Make gitweb consider projects in project root subdirectories
380 # to be forks of existing projects. Given project $projname.git,
381 # projects matching $projname/*.git will not be shown in the main
382 # projects list, instead a '+' mark will be added to $projname
383 # there and a 'forks' view will be enabled for the project, listing
384 # all the forks. If project list is taken from a file, forks have
385 # to be listed after the main project.
386
387 # To enable system wide have in $GITWEB_CONFIG
388 # $feature{'forks'}{'default'} = [1];
389 # Project specific override is not supported.
390 'forks' => {
391 'override' => 0,
392 'default' => [0]},
393
394 # Insert custom links to the action bar of all project pages.
395 # This enables you mainly to link to third-party scripts integrating
396 # into gitweb; e.g. git-browser for graphical history representation
397 # or custom web-based repository administration interface.
398
399 # The 'default' value consists of a list of triplets in the form
400 # (label, link, position) where position is the label after which
401 # to insert the link and link is a format string where %n expands
402 # to the project name, %f to the project path within the filesystem,
403 # %h to the current hash (h gitweb parameter) and %b to the current
404 # hash base (hb gitweb parameter); %% expands to %.
405
406 # To enable system wide have in $GITWEB_CONFIG e.g.
407 # $feature{'actions'}{'default'} = [('graphiclog',
408 # '/git-browser/by-commit.html?r=%n', 'summary')];
409 # Project specific override is not supported.
410 'actions' => {
411 'override' => 0,
412 'default' => []},
413
414 # Allow gitweb scan project content tags described in ctags/
415 # of project repository, and display the popular Web 2.0-ish
416 # "tag cloud" near the project list. Note that this is something
417 # COMPLETELY different from the normal Git tags.
418
419 # gitweb by itself can show existing tags, but it does not handle
420 # tagging itself; you need an external application for that.
421 # For an example script, check Girocco's cgi/tagproj.cgi.
422 # You may want to install the HTML::TagCloud Perl module to get
423 # a pretty tag cloud instead of just a list of tags.
424
425 # To enable system wide have in $GITWEB_CONFIG
426 # $feature{'ctags'}{'default'} = ['path_to_tag_script'];
427 # Project specific override is not supported.
428 'ctags' => {
429 'override' => 0,
430 'default' => [0]},
431
432 # The maximum number of patches in a patchset generated in patch
433 # view. Set this to 0 or undef to disable patch view, or to a
434 # negative number to remove any limit.
435
436 # To disable system wide have in $GITWEB_CONFIG
437 # $feature{'patches'}{'default'} = [0];
438 # To have project specific config enable override in $GITWEB_CONFIG
439 # $feature{'patches'}{'override'} = 1;
440 # and in project config gitweb.patches = 0|n;
441 # where n is the maximum number of patches allowed in a patchset.
442 'patches' => {
443 'sub' => \&feature_patches,
444 'override' => 0,
445 'default' => [16]},
446
447 # Avatar support. When this feature is enabled, views such as
448 # shortlog or commit will display an avatar associated with
449 # the email of the committer(s) and/or author(s).
450
451 # Currently available providers are gravatar and picon.
452 # If an unknown provider is specified, the feature is disabled.
453
454 # Gravatar depends on Digest::MD5.
455 # Picon currently relies on the indiana.edu database.
456
457 # To enable system wide have in $GITWEB_CONFIG
458 # $feature{'avatar'}{'default'} = ['<provider>'];
459 # where <provider> is either gravatar or picon.
460 # To have project specific config enable override in $GITWEB_CONFIG
461 # $feature{'avatar'}{'override'} = 1;
462 # and in project config gitweb.avatar = <provider>;
463 'avatar' => {
464 'sub' => \&feature_avatar,
465 'override' => 0,
466 'default' => ['']},
467
468 # Enable displaying how much time and how many git commands
469 # it took to generate and display page. Disabled by default.
470 # Project specific override is not supported.
471 'timed' => {
472 'override' => 0,
473 'default' => [0]},
474
475 # Enable turning some links into links to actions which require
476 # JavaScript to run (like 'blame_incremental'). Not enabled by
477 # default. Project specific override is currently not supported.
478 'javascript-actions' => {
479 'override' => 0,
480 'default' => [0]},
481
482 # Syntax highlighting support. This is based on Daniel Svensson's
483 # and Sham Chukoury's work in gitweb-xmms2.git.
484 # It requires the 'highlight' program present in $PATH,
485 # and therefore is disabled by default.
486
487 # To enable system wide have in $GITWEB_CONFIG
488 # $feature{'highlight'}{'default'} = [1];
489
490 'highlight' => {
491 'sub' => sub { feature_bool('highlight', @_) },
492 'override' => 0,
493 'default' => [0]},
494
495 # Enable displaying of remote heads in the heads list
496
497 # To enable system wide have in $GITWEB_CONFIG
498 # $feature{'remote_heads'}{'default'} = [1];
499 # To have project specific config enable override in $GITWEB_CONFIG
500 # $feature{'remote_heads'}{'override'} = 1;
501 # and in project config gitweb.remote_heads = 0|1;
502 'remote_heads' => {
503 'sub' => sub { feature_bool('remote_heads', @_) },
504 'override' => 0,
505 'default' => [0]},
506 );
507
508 sub gitweb_get_feature {
509 my ($name) = @_;
510 return unless exists $feature{$name};
511 my ($sub, $override, @defaults) = (
512 $feature{$name}{'sub'},
513 $feature{$name}{'override'},
514 @{$feature{$name}{'default'}});
515 # project specific override is possible only if we have project
516 our $git_dir; # global variable, declared later
517 if (!$override || !defined $git_dir) {
518 return @defaults;
519 }
520 if (!defined $sub) {
521 warn "feature $name is not overridable";
522 return @defaults;
523 }
524 return $sub->(@defaults);
525 }
526
527 # A wrapper to check if a given feature is enabled.
528 # With this, you can say
529 #
530 # my $bool_feat = gitweb_check_feature('bool_feat');
531 # gitweb_check_feature('bool_feat') or somecode;
532 #
533 # instead of
534 #
535 # my ($bool_feat) = gitweb_get_feature('bool_feat');
536 # (gitweb_get_feature('bool_feat'))[0] or somecode;
537 #
538 sub gitweb_check_feature {
539 return (gitweb_get_feature(@_))[0];
540 }
541
542
543 sub feature_bool {
544 my $key = shift;
545 my ($val) = git_get_project_config($key, '--bool');
546
547 if (!defined $val) {
548 return ($_[0]);
549 } elsif ($val eq 'true') {
550 return (1);
551 } elsif ($val eq 'false') {
552 return (0);
553 }
554 }
555
556 sub feature_snapshot {
557 my (@fmts) = @_;
558
559 my ($val) = git_get_project_config('snapshot');
560
561 if ($val) {
562 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
563 }
564
565 return @fmts;
566 }
567
568 sub feature_patches {
569 my @val = (git_get_project_config('patches', '--int'));
570
571 if (@val) {
572 return @val;
573 }
574
575 return ($_[0]);
576 }
577
578 sub feature_avatar {
579 my @val = (git_get_project_config('avatar'));
580
581 return @val ? @val : @_;
582 }
583
584 # checking HEAD file with -e is fragile if the repository was
585 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
586 # and then pruned.
587 sub check_head_link {
588 my ($dir) = @_;
589 my $headfile = "$dir/HEAD";
590 return ((-e $headfile) ||
591 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
592 }
593
594 sub check_export_ok {
595 my ($dir) = @_;
596 return (check_head_link($dir) &&
597 (!$export_ok || -e "$dir/$export_ok") &&
598 (!$export_auth_hook || $export_auth_hook->($dir)));
599 }
600
601 # process alternate names for backward compatibility
602 # filter out unsupported (unknown) snapshot formats
603 sub filter_snapshot_fmts {
604 my @fmts = @_;
605
606 @fmts = map {
607 exists $known_snapshot_format_aliases{$_} ?
608 $known_snapshot_format_aliases{$_} : $_} @fmts;
609 @fmts = grep {
610 exists $known_snapshot_formats{$_} &&
611 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
612 }
613
614 # If it is set to code reference, it is code that it is to be run once per
615 # request, allowing updating configurations that change with each request,
616 # while running other code in config file only once.
617 #
618 # Otherwise, if it is false then gitweb would process config file only once;
619 # if it is true then gitweb config would be run for each request.
620 our $per_request_config = 1;
621
622 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM);
623 sub evaluate_gitweb_config {
624 our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
625 our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
626 # die if there are errors parsing config file
627 if (-e $GITWEB_CONFIG) {
628 do $GITWEB_CONFIG;
629 die $@ if $@;
630 } elsif (-e $GITWEB_CONFIG_SYSTEM) {
631 do $GITWEB_CONFIG_SYSTEM;
632 die $@ if $@;
633 }
634 }
635
636 # Get loadavg of system, to compare against $maxload.
637 # Currently it requires '/proc/loadavg' present to get loadavg;
638 # if it is not present it returns 0, which means no load checking.
639 sub get_loadavg {
640 if( -e '/proc/loadavg' ){
641 open my $fd, '<', '/proc/loadavg'
642 or return 0;
643 my @load = split(/\s+/, scalar <$fd>);
644 close $fd;
645
646 # The first three columns measure CPU and IO utilization of the last one,
647 # five, and 10 minute periods. The fourth column shows the number of
648 # currently running processes and the total number of processes in the m/n
649 # format. The last column displays the last process ID used.
650 return $load[0] || 0;
651 }
652 # additional checks for load average should go here for things that don't export
653 # /proc/loadavg
654
655 return 0;
656 }
657
658 # version of the core git binary
659 our $git_version;
660 sub evaluate_git_version {
661 our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
662 $number_of_git_cmds++;
663 }
664
665 sub check_loadavg {
666 if (defined $maxload && get_loadavg() > $maxload) {
667 die_error(503, "The load average on the server is too high");
668 }
669 }
670
671 # ======================================================================
672 # input validation and dispatch
673
674 # input parameters can be collected from a variety of sources (presently, CGI
675 # and PATH_INFO), so we define an %input_params hash that collects them all
676 # together during validation: this allows subsequent uses (e.g. href()) to be
677 # agnostic of the parameter origin
678
679 our %input_params = ();
680
681 # input parameters are stored with the long parameter name as key. This will
682 # also be used in the href subroutine to convert parameters to their CGI
683 # equivalent, and since the href() usage is the most frequent one, we store
684 # the name -> CGI key mapping here, instead of the reverse.
685 #
686 # XXX: Warning: If you touch this, check the search form for updating,
687 # too.
688
689 our @cgi_param_mapping = (
690 project => "p",
691 action => "a",
692 file_name => "f",
693 file_parent => "fp",
694 hash => "h",
695 hash_parent => "hp",
696 hash_base => "hb",
697 hash_parent_base => "hpb",
698 page => "pg",
699 order => "o",
700 searchtext => "s",
701 searchtype => "st",
702 snapshot_format => "sf",
703 extra_options => "opt",
704 search_use_regexp => "sr",
705 # this must be last entry (for manipulation from JavaScript)
706 javascript => "js"
707 );
708 our %cgi_param_mapping = @cgi_param_mapping;
709
710 # we will also need to know the possible actions, for validation
711 our %actions = (
712 "blame" => \&git_blame,
713 "blame_incremental" => \&git_blame_incremental,
714 "blame_data" => \&git_blame_data,
715 "blobdiff" => \&git_blobdiff,
716 "blobdiff_plain" => \&git_blobdiff_plain,
717 "blob" => \&git_blob,
718 "blob_plain" => \&git_blob_plain,
719 "commitdiff" => \&git_commitdiff,
720 "commitdiff_plain" => \&git_commitdiff_plain,
721 "commit" => \&git_commit,
722 "forks" => \&git_forks,
723 "heads" => \&git_heads,
724 "history" => \&git_history,
725 "log" => \&git_log,
726 "patch" => \&git_patch,
727 "patches" => \&git_patches,
728 "remotes" => \&git_remotes,
729 "rss" => \&git_rss,
730 "atom" => \&git_atom,
731 "search" => \&git_search,
732 "search_help" => \&git_search_help,
733 "shortlog" => \&git_shortlog,
734 "summary" => \&git_summary,
735 "tag" => \&git_tag,
736 "tags" => \&git_tags,
737 "tree" => \&git_tree,
738 "snapshot" => \&git_snapshot,
739 "object" => \&git_object,
740 # those below don't need $project
741 "opml" => \&git_opml,
742 "project_list" => \&git_project_list,
743 "project_index" => \&git_project_index,
744 );
745
746 # finally, we have the hash of allowed extra_options for the commands that
747 # allow them
748 our %allowed_options = (
749 "--no-merges" => [ qw(rss atom log shortlog history) ],
750 );
751
752 # fill %input_params with the CGI parameters. All values except for 'opt'
753 # should be single values, but opt can be an array. We should probably
754 # build an array of parameters that can be multi-valued, but since for the time
755 # being it's only this one, we just single it out
756 sub evaluate_query_params {
757 our $cgi;
758
759 while (my ($name, $symbol) = each %cgi_param_mapping) {
760 if ($symbol eq 'opt') {
761 $input_params{$name} = [ $cgi->param($symbol) ];
762 } else {
763 $input_params{$name} = $cgi->param($symbol);
764 }
765 }
766 }
767
768 # now read PATH_INFO and update the parameter list for missing parameters
769 sub evaluate_path_info {
770 return if defined $input_params{'project'};
771 return if !$path_info;
772 $path_info =~ s,^/+,,;
773 return if !$path_info;
774
775 # find which part of PATH_INFO is project
776 my $project = $path_info;
777 $project =~ s,/+$,,;
778 while ($project && !check_head_link("$projectroot/$project")) {
779 $project =~ s,/*[^/]*$,,;
780 }
781 return unless $project;
782 $input_params{'project'} = $project;
783
784 # do not change any parameters if an action is given using the query string
785 return if $input_params{'action'};
786 $path_info =~ s,^\Q$project\E/*,,;
787
788 # next, check if we have an action
789 my $action = $path_info;
790 $action =~ s,/.*$,,;
791 if (exists $actions{$action}) {
792 $path_info =~ s,^$action/*,,;
793 $input_params{'action'} = $action;
794 }
795
796 # list of actions that want hash_base instead of hash, but can have no
797 # pathname (f) parameter
798 my @wants_base = (
799 'tree',
800 'history',
801 );
802
803 # we want to catch, among others
804 # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
805 my ($parentrefname, $parentpathname, $refname, $pathname) =
806 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
807
808 # first, analyze the 'current' part
809 if (defined $pathname) {
810 # we got "branch:filename" or "branch:dir/"
811 # we could use git_get_type(branch:pathname), but:
812 # - it needs $git_dir
813 # - it does a git() call
814 # - the convention of terminating directories with a slash
815 # makes it superfluous
816 # - embedding the action in the PATH_INFO would make it even
817 # more superfluous
818 $pathname =~ s,^/+,,;
819 if (!$pathname || substr($pathname, -1) eq "/") {
820 $input_params{'action'} ||= "tree";
821 $pathname =~ s,/$,,;
822 } else {
823 # the default action depends on whether we had parent info
824 # or not
825 if ($parentrefname) {
826 $input_params{'action'} ||= "blobdiff_plain";
827 } else {
828 $input_params{'action'} ||= "blob_plain";
829 }
830 }
831 $input_params{'hash_base'} ||= $refname;
832 $input_params{'file_name'} ||= $pathname;
833 } elsif (defined $refname) {
834 # we got "branch". In this case we have to choose if we have to
835 # set hash or hash_base.
836 #
837 # Most of the actions without a pathname only want hash to be
838 # set, except for the ones specified in @wants_base that want
839 # hash_base instead. It should also be noted that hand-crafted
840 # links having 'history' as an action and no pathname or hash
841 # set will fail, but that happens regardless of PATH_INFO.
842 if (defined $parentrefname) {
843 # if there is parent let the default be 'shortlog' action
844 # (for http://git.example.com/repo.git/A..B links); if there
845 # is no parent, dispatch will detect type of object and set
846 # action appropriately if required (if action is not set)
847 $input_params{'action'} ||= "shortlog";
848 }
849 if ($input_params{'action'} &&
850 grep { $_ eq $input_params{'action'} } @wants_base) {
851 $input_params{'hash_base'} ||= $refname;
852 } else {
853 $input_params{'hash'} ||= $refname;
854 }
855 }
856
857 # next, handle the 'parent' part, if present
858 if (defined $parentrefname) {
859 # a missing pathspec defaults to the 'current' filename, allowing e.g.
860 # someproject/blobdiff/oldrev..newrev:/filename
861 if ($parentpathname) {
862 $parentpathname =~ s,^/+,,;
863 $parentpathname =~ s,/$,,;
864 $input_params{'file_parent'} ||= $parentpathname;
865 } else {
866 $input_params{'file_parent'} ||= $input_params{'file_name'};
867 }
868 # we assume that hash_parent_base is wanted if a path was specified,
869 # or if the action wants hash_base instead of hash
870 if (defined $input_params{'file_parent'} ||
871 grep { $_ eq $input_params{'action'} } @wants_base) {
872 $input_params{'hash_parent_base'} ||= $parentrefname;
873 } else {
874 $input_params{'hash_parent'} ||= $parentrefname;
875 }
876 }
877
878 # for the snapshot action, we allow URLs in the form
879 # $project/snapshot/$hash.ext
880 # where .ext determines the snapshot and gets removed from the
881 # passed $refname to provide the $hash.
882 #
883 # To be able to tell that $refname includes the format extension, we
884 # require the following two conditions to be satisfied:
885 # - the hash input parameter MUST have been set from the $refname part
886 # of the URL (i.e. they must be equal)
887 # - the snapshot format MUST NOT have been defined already (e.g. from
888 # CGI parameter sf)
889 # It's also useless to try any matching unless $refname has a dot,
890 # so we check for that too
891 if (defined $input_params{'action'} &&
892 $input_params{'action'} eq 'snapshot' &&
893 defined $refname && index($refname, '.') != -1 &&
894 $refname eq $input_params{'hash'} &&
895 !defined $input_params{'snapshot_format'}) {
896 # We loop over the known snapshot formats, checking for
897 # extensions. Allowed extensions are both the defined suffix
898 # (which includes the initial dot already) and the snapshot
899 # format key itself, with a prepended dot
900 while (my ($fmt, $opt) = each %known_snapshot_formats) {
901 my $hash = $refname;
902 unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
903 next;
904 }
905 my $sfx = $1;
906 # a valid suffix was found, so set the snapshot format
907 # and reset the hash parameter
908 $input_params{'snapshot_format'} = $fmt;
909 $input_params{'hash'} = $hash;
910 # we also set the format suffix to the one requested
911 # in the URL: this way a request for e.g. .tgz returns
912 # a .tgz instead of a .tar.gz
913 $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
914 last;
915 }
916 }
917 }
918
919 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
920 $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
921 $searchtext, $search_regexp);
922 sub evaluate_and_validate_params {
923 our $action = $input_params{'action'};
924 if (defined $action) {
925 if (!validate_action($action)) {
926 die_error(400, "Invalid action parameter");
927 }
928 }
929
930 # parameters which are pathnames
931 our $project = $input_params{'project'};
932 if (defined $project) {
933 if (!validate_project($project)) {
934 undef $project;
935 die_error(404, "No such project");
936 }
937 }
938
939 our $file_name = $input_params{'file_name'};
940 if (defined $file_name) {
941 if (!validate_pathname($file_name)) {
942 die_error(400, "Invalid file parameter");
943 }
944 }
945
946 our $file_parent = $input_params{'file_parent'};
947 if (defined $file_parent) {
948 if (!validate_pathname($file_parent)) {
949 die_error(400, "Invalid file parent parameter");
950 }
951 }
952
953 # parameters which are refnames
954 our $hash = $input_params{'hash'};
955 if (defined $hash) {
956 if (!validate_refname($hash)) {
957 die_error(400, "Invalid hash parameter");
958 }
959 }
960
961 our $hash_parent = $input_params{'hash_parent'};
962 if (defined $hash_parent) {
963 if (!validate_refname($hash_parent)) {
964 die_error(400, "Invalid hash parent parameter");
965 }
966 }
967
968 our $hash_base = $input_params{'hash_base'};
969 if (defined $hash_base) {
970 if (!validate_refname($hash_base)) {
971 die_error(400, "Invalid hash base parameter");
972 }
973 }
974
975 our @extra_options = @{$input_params{'extra_options'}};
976 # @extra_options is always defined, since it can only be (currently) set from
977 # CGI, and $cgi->param() returns the empty array in array context if the param
978 # is not set
979 foreach my $opt (@extra_options) {
980 if (not exists $allowed_options{$opt}) {
981 die_error(400, "Invalid option parameter");
982 }
983 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
984 die_error(400, "Invalid option parameter for this action");
985 }
986 }
987
988 our $hash_parent_base = $input_params{'hash_parent_base'};
989 if (defined $hash_parent_base) {
990 if (!validate_refname($hash_parent_base)) {
991 die_error(400, "Invalid hash parent base parameter");
992 }
993 }
994
995 # other parameters
996 our $page = $input_params{'page'};
997 if (defined $page) {
998 if ($page =~ m/[^0-9]/) {
999 die_error(400, "Invalid page parameter");
1000 }
1001 }
1002
1003 our $searchtype = $input_params{'searchtype'};
1004 if (defined $searchtype) {
1005 if ($searchtype =~ m/[^a-z]/) {
1006 die_error(400, "Invalid searchtype parameter");
1007 }
1008 }
1009
1010 our $search_use_regexp = $input_params{'search_use_regexp'};
1011
1012 our $searchtext = $input_params{'searchtext'};
1013 our $search_regexp;
1014 if (defined $searchtext) {
1015 if (length($searchtext) < 2) {
1016 die_error(403, "At least two characters are required for search parameter");
1017 }
1018 $search_regexp = $search_use_regexp ? $searchtext : quotemeta $searchtext;
1019 }
1020 }
1021
1022 # path to the current git repository
1023 our $git_dir;
1024 sub evaluate_git_dir {
1025 our $git_dir = "$projectroot/$project" if $project;
1026 }
1027
1028 our (@snapshot_fmts, $git_avatar);
1029 sub configure_gitweb_features {
1030 # list of supported snapshot formats
1031 our @snapshot_fmts = gitweb_get_feature('snapshot');
1032 @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1033
1034 # check that the avatar feature is set to a known provider name,
1035 # and for each provider check if the dependencies are satisfied.
1036 # if the provider name is invalid or the dependencies are not met,
1037 # reset $git_avatar to the empty string.
1038 our ($git_avatar) = gitweb_get_feature('avatar');
1039 if ($git_avatar eq 'gravatar') {
1040 $git_avatar = '' unless (eval { require Digest::MD5; 1; });
1041 } elsif ($git_avatar eq 'picon') {
1042 # no dependencies
1043 } else {
1044 $git_avatar = '';
1045 }
1046 }
1047
1048 # custom error handler: 'die <message>' is Internal Server Error
1049 sub handle_errors_html {
1050 my $msg = shift; # it is already HTML escaped
1051
1052 # to avoid infinite loop where error occurs in die_error,
1053 # change handler to default handler, disabling handle_errors_html
1054 set_message("Error occured when inside die_error:\n$msg");
1055
1056 # you cannot jump out of die_error when called as error handler;
1057 # the subroutine set via CGI::Carp::set_message is called _after_
1058 # HTTP headers are already written, so it cannot write them itself
1059 die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1060 }
1061 set_message(\&handle_errors_html);
1062
1063 # dispatch
1064 sub dispatch {
1065 if (!defined $action) {
1066 if (defined $hash) {
1067 $action = git_get_type($hash);
1068 } elsif (defined $hash_base && defined $file_name) {
1069 $action = git_get_type("$hash_base:$file_name");
1070 } elsif (defined $project) {
1071 $action = 'summary';
1072 } else {
1073 $action = 'project_list';
1074 }
1075 }
1076 if (!defined($actions{$action})) {
1077 die_error(400, "Unknown action");
1078 }
1079 if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1080 !$project) {
1081 die_error(400, "Project needed");
1082 }
1083 $actions{$action}->();
1084 }
1085
1086 sub reset_timer {
1087 our $t0 = [ gettimeofday() ]
1088 if defined $t0;
1089 our $number_of_git_cmds = 0;
1090 }
1091
1092 our $first_request = 1;
1093 sub run_request {
1094 reset_timer();
1095
1096 evaluate_uri();
1097 if ($first_request) {
1098 evaluate_gitweb_config();
1099 evaluate_git_version();
1100 }
1101 if ($per_request_config) {
1102 if (ref($per_request_config) eq 'CODE') {
1103 $per_request_config->();
1104 } elsif (!$first_request) {
1105 evaluate_gitweb_config();
1106 }
1107 }
1108 check_loadavg();
1109
1110 # $projectroot and $projects_list might be set in gitweb config file
1111 $projects_list ||= $projectroot;
1112
1113 evaluate_query_params();
1114 evaluate_path_info();
1115 evaluate_and_validate_params();
1116 evaluate_git_dir();
1117
1118 configure_gitweb_features();
1119
1120 dispatch();
1121 }
1122
1123 our $is_last_request = sub { 1 };
1124 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1125 our $CGI = 'CGI';
1126 our $cgi;
1127 sub configure_as_fcgi {
1128 require CGI::Fast;
1129 our $CGI = 'CGI::Fast';
1130
1131 my $request_number = 0;
1132 # let each child service 100 requests
1133 our $is_last_request = sub { ++$request_number > 100 };
1134 }
1135 sub evaluate_argv {
1136 my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1137 configure_as_fcgi()
1138 if $script_name =~ /\.fcgi$/;
1139
1140 return unless (@ARGV);
1141
1142 require Getopt::Long;
1143 Getopt::Long::GetOptions(
1144 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1145 'nproc|n=i' => sub {
1146 my ($arg, $val) = @_;
1147 return unless eval { require FCGI::ProcManager; 1; };
1148 my $proc_manager = FCGI::ProcManager->new({
1149 n_processes => $val,
1150 });
1151 our $pre_listen_hook = sub { $proc_manager->pm_manage() };
1152 our $pre_dispatch_hook = sub { $proc_manager->pm_pre_dispatch() };
1153 our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1154 },
1155 );
1156 }
1157
1158 sub run {
1159 evaluate_argv();
1160
1161 $first_request = 1;
1162 $pre_listen_hook->()
1163 if $pre_listen_hook;
1164
1165 REQUEST:
1166 while ($cgi = $CGI->new()) {
1167 $pre_dispatch_hook->()
1168 if $pre_dispatch_hook;
1169
1170 run_request();
1171
1172 $post_dispatch_hook->()
1173 if $post_dispatch_hook;
1174 $first_request = 0;
1175
1176 last REQUEST if ($is_last_request->());
1177 }
1178
1179 DONE_GITWEB:
1180 1;
1181 }
1182
1183 run();
1184
1185 if (defined caller) {
1186 # wrapped in a subroutine processing requests,
1187 # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1188 return;
1189 } else {
1190 # pure CGI script, serving single request
1191 exit;
1192 }
1193
1194 ## ======================================================================
1195 ## action links
1196
1197 # possible values of extra options
1198 # -full => 0|1 - use absolute/full URL ($my_uri/$my_url as base)
1199 # -replay => 1 - start from a current view (replay with modifications)
1200 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1201 sub href {
1202 my %params = @_;
1203 # default is to use -absolute url() i.e. $my_uri
1204 my $href = $params{-full} ? $my_url : $my_uri;
1205
1206 $params{'project'} = $project unless exists $params{'project'};
1207
1208 if ($params{-replay}) {
1209 while (my ($name, $symbol) = each %cgi_param_mapping) {
1210 if (!exists $params{$name}) {
1211 $params{$name} = $input_params{$name};
1212 }
1213 }
1214 }
1215
1216 my $use_pathinfo = gitweb_check_feature('pathinfo');
1217 if (defined $params{'project'} &&
1218 (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1219 # try to put as many parameters as possible in PATH_INFO:
1220 # - project name
1221 # - action
1222 # - hash_parent or hash_parent_base:/file_parent
1223 # - hash or hash_base:/filename
1224 # - the snapshot_format as an appropriate suffix
1225
1226 # When the script is the root DirectoryIndex for the domain,
1227 # $href here would be something like http://gitweb.example.com/
1228 # Thus, we strip any trailing / from $href, to spare us double
1229 # slashes in the final URL
1230 $href =~ s,/$,,;
1231
1232 # Then add the project name, if present
1233 $href .= "/".esc_url($params{'project'});
1234 delete $params{'project'};
1235
1236 # since we destructively absorb parameters, we keep this
1237 # boolean that remembers if we're handling a snapshot
1238 my $is_snapshot = $params{'action'} eq 'snapshot';
1239
1240 # Summary just uses the project path URL, any other action is
1241 # added to the URL
1242 if (defined $params{'action'}) {
1243 $href .= "/".esc_url($params{'action'}) unless $params{'action'} eq 'summary';
1244 delete $params{'action'};
1245 }
1246
1247 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1248 # stripping nonexistent or useless pieces
1249 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1250 || $params{'hash_parent'} || $params{'hash'});
1251 if (defined $params{'hash_base'}) {
1252 if (defined $params{'hash_parent_base'}) {
1253 $href .= esc_url($params{'hash_parent_base'});
1254 # skip the file_parent if it's the same as the file_name
1255 if (defined $params{'file_parent'}) {
1256 if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1257 delete $params{'file_parent'};
1258 } elsif ($params{'file_parent'} !~ /\.\./) {
1259 $href .= ":/".esc_url($params{'file_parent'});
1260 delete $params{'file_parent'};
1261 }
1262 }
1263 $href .= "..";
1264 delete $params{'hash_parent'};
1265 delete $params{'hash_parent_base'};
1266 } elsif (defined $params{'hash_parent'}) {
1267 $href .= esc_url($params{'hash_parent'}). "..";
1268 delete $params{'hash_parent'};
1269 }
1270
1271 $href .= esc_url($params{'hash_base'});
1272 if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1273 $href .= ":/".esc_url($params{'file_name'});
1274 delete $params{'file_name'};
1275 }
1276 delete $params{'hash'};
1277 delete $params{'hash_base'};
1278 } elsif (defined $params{'hash'}) {
1279 $href .= esc_url($params{'hash'});
1280 delete $params{'hash'};
1281 }
1282
1283 # If the action was a snapshot, we can absorb the
1284 # snapshot_format parameter too
1285 if ($is_snapshot) {
1286 my $fmt = $params{'snapshot_format'};
1287 # snapshot_format should always be defined when href()
1288 # is called, but just in case some code forgets, we
1289 # fall back to the default
1290 $fmt ||= $snapshot_fmts[0];
1291 $href .= $known_snapshot_formats{$fmt}{'suffix'};
1292 delete $params{'snapshot_format'};
1293 }
1294 }
1295
1296 # now encode the parameters explicitly
1297 my @result = ();
1298 for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1299 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1300 if (defined $params{$name}) {
1301 if (ref($params{$name}) eq "ARRAY") {
1302 foreach my $par (@{$params{$name}}) {
1303 push @result, $symbol . "=" . esc_param($par);
1304 }
1305 } else {
1306 push @result, $symbol . "=" . esc_param($params{$name});
1307 }
1308 }
1309 }
1310 $href .= "?" . join(';', @result) if scalar @result;
1311
1312 return $href;
1313 }
1314
1315
1316 ## ======================================================================
1317 ## validation, quoting/unquoting and escaping
1318
1319 sub validate_action {
1320 my $input = shift || return undef;
1321 return undef unless exists $actions{$input};
1322 return $input;
1323 }
1324
1325 sub validate_project {
1326 my $input = shift || return undef;
1327 if (!validate_pathname($input) ||
1328 !(-d "$projectroot/$input") ||
1329 !check_export_ok("$projectroot/$input") ||
1330 ($strict_export && !project_in_list($input))) {
1331 return undef;
1332 } else {
1333 return $input;
1334 }
1335 }
1336
1337 sub validate_pathname {
1338 my $input = shift || return undef;
1339
1340 # no '.' or '..' as elements of path, i.e. no '.' nor '..'
1341 # at the beginning, at the end, and between slashes.
1342 # also this catches doubled slashes
1343 if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1344 return undef;
1345 }
1346 # no null characters
1347 if ($input =~ m!\0!) {
1348 return undef;
1349 }
1350 return $input;
1351 }
1352
1353 sub validate_refname {
1354 my $input = shift || return undef;
1355
1356 # textual hashes are O.K.
1357 if ($input =~ m/^[0-9a-fA-F]{40}$/) {
1358 return $input;
1359 }
1360 # it must be correct pathname
1361 $input = validate_pathname($input)
1362 or return undef;
1363 # restrictions on ref name according to git-check-ref-format
1364 if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1365 return undef;
1366 }
1367 return $input;
1368 }
1369
1370 # decode sequences of octets in utf8 into Perl's internal form,
1371 # which is utf-8 with utf8 flag set if needed. gitweb writes out
1372 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1373 sub to_utf8 {
1374 my $str = shift;
1375 return undef unless defined $str;
1376 if (utf8::valid($str)) {
1377 utf8::decode($str);
1378 return $str;
1379 } else {
1380 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1381 }
1382 }
1383
1384 # quote unsafe chars, but keep the slash, even when it's not
1385 # correct, but quoted slashes look too horrible in bookmarks
1386 sub esc_param {
1387 my $str = shift;
1388 return undef unless defined $str;
1389 $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1390 $str =~ s/ /\+/g;
1391 return $str;
1392 }
1393
1394 # quote unsafe chars in whole URL, so some characters cannot be quoted
1395 sub esc_url {
1396 my $str = shift;
1397 return undef unless defined $str;
1398 $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1399 $str =~ s/ /\+/g;
1400 return $str;
1401 }
1402
1403 # quote unsafe characters in HTML attributes
1404 sub esc_attr {
1405
1406 # for XHTML conformance escaping '"' to '&quot;' is not enough
1407 return esc_html(@_);
1408 }
1409
1410 # replace invalid utf8 character with SUBSTITUTION sequence
1411 sub esc_html {
1412 my $str = shift;
1413 my %opts = @_;
1414
1415 return undef unless defined $str;
1416
1417 $str = to_utf8($str);
1418 $str = $cgi->escapeHTML($str);
1419 if ($opts{'-nbsp'}) {
1420 $str =~ s/ /&nbsp;/g;
1421 }
1422 $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1423 return $str;
1424 }
1425
1426 # quote control characters and escape filename to HTML
1427 sub esc_path {
1428 my $str = shift;
1429 my %opts = @_;
1430
1431 return undef unless defined $str;
1432
1433 $str = to_utf8($str);
1434 $str = $cgi->escapeHTML($str);
1435 if ($opts{'-nbsp'}) {
1436 $str =~ s/ /&nbsp;/g;
1437 }
1438 $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1439 return $str;
1440 }
1441
1442 # Make control characters "printable", using character escape codes (CEC)
1443 sub quot_cec {
1444 my $cntrl = shift;
1445 my %opts = @_;
1446 my %es = ( # character escape codes, aka escape sequences
1447 "\t" => '\t', # tab (HT)
1448 "\n" => '\n', # line feed (LF)
1449 "\r" => '\r', # carrige return (CR)
1450 "\f" => '\f', # form feed (FF)
1451 "\b" => '\b', # backspace (BS)
1452 "\a" => '\a', # alarm (bell) (BEL)
1453 "\e" => '\e', # escape (ESC)
1454 "\013" => '\v', # vertical tab (VT)
1455 "\000" => '\0', # nul character (NUL)
1456 );
1457 my $chr = ( (exists $es{$cntrl})
1458 ? $es{$cntrl}
1459 : sprintf('\%2x', ord($cntrl)) );
1460 if ($opts{-nohtml}) {
1461 return $chr;
1462 } else {
1463 return "<span class=\"cntrl\">$chr</span>";
1464 }
1465 }
1466
1467 # Alternatively use unicode control pictures codepoints,
1468 # Unicode "printable representation" (PR)
1469 sub quot_upr {
1470 my $cntrl = shift;
1471 my %opts = @_;
1472
1473 my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1474 if ($opts{-nohtml}) {
1475 return $chr;
1476 } else {
1477 return "<span class=\"cntrl\">$chr</span>";
1478 }
1479 }
1480
1481 # git may return quoted and escaped filenames
1482 sub unquote {
1483 my $str = shift;
1484
1485 sub unq {
1486 my $seq = shift;
1487 my %es = ( # character escape codes, aka escape sequences
1488 't' => "\t", # tab (HT, TAB)
1489 'n' => "\n", # newline (NL)
1490 'r' => "\r", # return (CR)
1491 'f' => "\f", # form feed (FF)
1492 'b' => "\b", # backspace (BS)
1493 'a' => "\a", # alarm (bell) (BEL)
1494 'e' => "\e", # escape (ESC)
1495 'v' => "\013", # vertical tab (VT)
1496 );
1497
1498 if ($seq =~ m/^[0-7]{1,3}$/) {
1499 # octal char sequence
1500 return chr(oct($seq));
1501 } elsif (exists $es{$seq}) {
1502 # C escape sequence, aka character escape code
1503 return $es{$seq};
1504 }
1505 # quoted ordinary character
1506 return $seq;
1507 }
1508
1509 if ($str =~ m/^"(.*)"$/) {
1510 # needs unquoting
1511 $str = $1;
1512 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1513 }
1514 return $str;
1515 }
1516
1517 # escape tabs (convert tabs to spaces)
1518 sub untabify {
1519 my $line = shift;
1520
1521 while ((my $pos = index($line, "\t")) != -1) {
1522 if (my $count = (8 - ($pos % 8))) {
1523 my $spaces = ' ' x $count;
1524 $line =~ s/\t/$spaces/;
1525 }
1526 }
1527
1528 return $line;
1529 }
1530
1531 sub project_in_list {
1532 my $project = shift;
1533 my @list = git_get_projects_list();
1534 return @list && scalar(grep { $_->{'path'} eq $project } @list);
1535 }
1536
1537 ## ----------------------------------------------------------------------
1538 ## HTML aware string manipulation
1539
1540 # Try to chop given string on a word boundary between position
1541 # $len and $len+$add_len. If there is no word boundary there,
1542 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1543 # (marking chopped part) would be longer than given string.
1544 sub chop_str {
1545 my $str = shift;
1546 my $len = shift;
1547 my $add_len = shift || 10;
1548 my $where = shift || 'right'; # 'left' | 'center' | 'right'
1549
1550 # Make sure perl knows it is utf8 encoded so we don't
1551 # cut in the middle of a utf8 multibyte char.
1552 $str = to_utf8($str);
1553
1554 # allow only $len chars, but don't cut a word if it would fit in $add_len
1555 # if it doesn't fit, cut it if it's still longer than the dots we would add
1556 # remove chopped character entities entirely
1557
1558 # when chopping in the middle, distribute $len into left and right part
1559 # return early if chopping wouldn't make string shorter
1560 if ($where eq 'center') {
1561 return $str if ($len + 5 >= length($str)); # filler is length 5
1562 $len = int($len/2);
1563 } else {
1564 return $str if ($len + 4 >= length($str)); # filler is length 4
1565 }
1566
1567 # regexps: ending and beginning with word part up to $add_len
1568 my $endre = qr/.{$len}\w{0,$add_len}/;
1569 my $begre = qr/\w{0,$add_len}.{$len}/;
1570
1571 if ($where eq 'left') {
1572 $str =~ m/^(.*?)($begre)$/;
1573 my ($lead, $body) = ($1, $2);
1574 if (length($lead) > 4) {
1575 $lead = " ...";
1576 }
1577 return "$lead$body";
1578
1579 } elsif ($where eq 'center') {
1580 $str =~ m/^($endre)(.*)$/;
1581 my ($left, $str) = ($1, $2);
1582 $str =~ m/^(.*?)($begre)$/;
1583 my ($mid, $right) = ($1, $2);
1584 if (length($mid) > 5) {
1585 $mid = " ... ";
1586 }
1587 return "$left$mid$right";
1588
1589 } else {
1590 $str =~ m/^($endre)(.*)$/;
1591 my $body = $1;
1592 my $tail = $2;
1593 if (length($tail) > 4) {
1594 $tail = "... ";
1595 }
1596 return "$body$tail";
1597 }
1598 }
1599
1600 # takes the same arguments as chop_str, but also wraps a <span> around the
1601 # result with a title attribute if it does get chopped. Additionally, the
1602 # string is HTML-escaped.
1603 sub chop_and_escape_str {
1604 my ($str) = @_;
1605
1606 my $chopped = chop_str(@_);
1607 if ($chopped eq $str) {
1608 return esc_html($chopped);
1609 } else {
1610 $str =~ s/[[:cntrl:]]/?/g;
1611 return $cgi->span({-title=>$str}, esc_html($chopped));
1612 }
1613 }
1614
1615 ## ----------------------------------------------------------------------
1616 ## functions returning short strings
1617
1618 # CSS class for given age value (in seconds)
1619 sub age_class {
1620 my $age = shift;
1621
1622 if (!defined $age) {
1623 return "noage";
1624 } elsif ($age < 60*60*2) {
1625 return "age0";
1626 } elsif ($age < 60*60*24*2) {
1627 return "age1";
1628 } else {
1629 return "age2";
1630 }
1631 }
1632
1633 # convert age in seconds to "nn units ago" string
1634 sub age_string {
1635 my $age = shift;
1636 my $age_str;
1637
1638 if ($age > 60*60*24*365*2) {
1639 $age_str = (int $age/60/60/24/365);
1640 $age_str .= " years ago";
1641 } elsif ($age > 60*60*24*(365/12)*2) {
1642 $age_str = int $age/60/60/24/(365/12);
1643 $age_str .= " months ago";
1644 } elsif ($age > 60*60*24*7*2) {
1645 $age_str = int $age/60/60/24/7;
1646 $age_str .= " weeks ago";
1647 } elsif ($age > 60*60*24*2) {
1648 $age_str = int $age/60/60/24;
1649 $age_str .= " days ago";
1650 } elsif ($age > 60*60*2) {
1651 $age_str = int $age/60/60;
1652 $age_str .= " hours ago";
1653 } elsif ($age > 60*2) {
1654 $age_str = int $age/60;
1655 $age_str .= " min ago";
1656 } elsif ($age > 2) {
1657 $age_str = int $age;
1658 $age_str .= " sec ago";
1659 } else {
1660 $age_str .= " right now";
1661 }
1662 return $age_str;
1663 }
1664
1665 use constant {
1666 S_IFINVALID => 0030000,
1667 S_IFGITLINK => 0160000,
1668 };
1669
1670 # submodule/subproject, a commit object reference
1671 sub S_ISGITLINK {
1672 my $mode = shift;
1673
1674 return (($mode & S_IFMT) == S_IFGITLINK)
1675 }
1676
1677 # convert file mode in octal to symbolic file mode string
1678 sub mode_str {
1679 my $mode = oct shift;
1680
1681 if (S_ISGITLINK($mode)) {
1682 return 'm---------';
1683 } elsif (S_ISDIR($mode & S_IFMT)) {
1684 return 'drwxr-xr-x';
1685 } elsif (S_ISLNK($mode)) {
1686 return 'lrwxrwxrwx';
1687 } elsif (S_ISREG($mode)) {
1688 # git cares only about the executable bit
1689 if ($mode & S_IXUSR) {
1690 return '-rwxr-xr-x';
1691 } else {
1692 return '-rw-r--r--';
1693 };
1694 } else {
1695 return '----------';
1696 }
1697 }
1698
1699 # convert file mode in octal to file type string
1700 sub file_type {
1701 my $mode = shift;
1702
1703 if ($mode !~ m/^[0-7]+$/) {
1704 return $mode;
1705 } else {
1706 $mode = oct $mode;
1707 }
1708
1709 if (S_ISGITLINK($mode)) {
1710 return "submodule";
1711 } elsif (S_ISDIR($mode & S_IFMT)) {
1712 return "directory";
1713 } elsif (S_ISLNK($mode)) {
1714 return "symlink";
1715 } elsif (S_ISREG($mode)) {
1716 return "file";
1717 } else {
1718 return "unknown";
1719 }
1720 }
1721
1722 # convert file mode in octal to file type description string
1723 sub file_type_long {
1724 my $mode = shift;
1725
1726 if ($mode !~ m/^[0-7]+$/) {
1727 return $mode;
1728 } else {
1729 $mode = oct $mode;
1730 }
1731
1732 if (S_ISGITLINK($mode)) {
1733 return "submodule";
1734 } elsif (S_ISDIR($mode & S_IFMT)) {
1735 return "directory";
1736 } elsif (S_ISLNK($mode)) {
1737 return "symlink";
1738 } elsif (S_ISREG($mode)) {
1739 if ($mode & S_IXUSR) {
1740 return "executable";
1741 } else {
1742 return "file";
1743 };
1744 } else {
1745 return "unknown";
1746 }
1747 }
1748
1749
1750 ## ----------------------------------------------------------------------
1751 ## functions returning short HTML fragments, or transforming HTML fragments
1752 ## which don't belong to other sections
1753
1754 # format line of commit message.
1755 sub format_log_line_html {
1756 my $line = shift;
1757
1758 $line = esc_html($line, -nbsp=>1);
1759 $line =~ s{\b([0-9a-fA-F]{8,40})\b}{
1760 $cgi->a({-href => href(action=>"object", hash=>$1),
1761 -class => "text"}, $1);
1762 }eg;
1763
1764 return $line;
1765 }
1766
1767 # format marker of refs pointing to given object
1768
1769 # the destination action is chosen based on object type and current context:
1770 # - for annotated tags, we choose the tag view unless it's the current view
1771 # already, in which case we go to shortlog view
1772 # - for other refs, we keep the current view if we're in history, shortlog or
1773 # log view, and select shortlog otherwise
1774 sub format_ref_marker {
1775 my ($refs, $id) = @_;
1776 my $markers = '';
1777
1778 if (defined $refs->{$id}) {
1779 foreach my $ref (@{$refs->{$id}}) {
1780 # this code exploits the fact that non-lightweight tags are the
1781 # only indirect objects, and that they are the only objects for which
1782 # we want to use tag instead of shortlog as action
1783 my ($type, $name) = qw();
1784 my $indirect = ($ref =~ s/\^\{\}$//);
1785 # e.g. tags/v2.6.11 or heads/next
1786 if ($ref =~ m!^(.*?)s?/(.*)$!) {
1787 $type = $1;
1788 $name = $2;
1789 } else {
1790 $type = "ref";
1791 $name = $ref;
1792 }
1793
1794 my $class = $type;
1795 $class .= " indirect" if $indirect;
1796
1797 my $dest_action = "shortlog";
1798
1799 if ($indirect) {
1800 $dest_action = "tag" unless $action eq "tag";
1801 } elsif ($action =~ /^(history|(short)?log)$/) {
1802 $dest_action = $action;
1803 }
1804
1805 my $dest = "";
1806 $dest .= "refs/" unless $ref =~ m!^refs/!;
1807 $dest .= $ref;
1808
1809 my $link = $cgi->a({
1810 -href => href(
1811 action=>$dest_action,
1812 hash=>$dest
1813 )}, $name);
1814
1815 $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
1816 $link . "</span>";
1817 }
1818 }
1819
1820 if ($markers) {
1821 return ' <span class="refs">'. $markers . '</span>';
1822 } else {
1823 return "";
1824 }
1825 }
1826
1827 # format, perhaps shortened and with markers, title line
1828 sub format_subject_html {
1829 my ($long, $short, $href, $extra) = @_;
1830 $extra = '' unless defined($extra);
1831
1832 if (length($short) < length($long)) {
1833 $long =~ s/[[:cntrl:]]/?/g;
1834 return $cgi->a({-href => $href, -class => "list subject",
1835 -title => to_utf8($long)},
1836 esc_html($short)) . $extra;
1837 } else {
1838 return $cgi->a({-href => $href, -class => "list subject"},
1839 esc_html($long)) . $extra;
1840 }
1841 }
1842
1843 # Rather than recomputing the url for an email multiple times, we cache it
1844 # after the first hit. This gives a visible benefit in views where the avatar
1845 # for the same email is used repeatedly (e.g. shortlog).
1846 # The cache is shared by all avatar engines (currently gravatar only), which
1847 # are free to use it as preferred. Since only one avatar engine is used for any
1848 # given page, there's no risk for cache conflicts.
1849 our %avatar_cache = ();
1850
1851 # Compute the picon url for a given email, by using the picon search service over at
1852 # http://www.cs.indiana.edu/picons/search.html
1853 sub picon_url {
1854 my $email = lc shift;
1855 if (!$avatar_cache{$email}) {
1856 my ($user, $domain) = split('@', $email);
1857 $avatar_cache{$email} =
1858 "http://www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
1859 "$domain/$user/" .
1860 "users+domains+unknown/up/single";
1861 }
1862 return $avatar_cache{$email};
1863 }
1864
1865 # Compute the gravatar url for a given email, if it's not in the cache already.
1866 # Gravatar stores only the part of the URL before the size, since that's the
1867 # one computationally more expensive. This also allows reuse of the cache for
1868 # different sizes (for this particular engine).
1869 sub gravatar_url {
1870 my $email = lc shift;
1871 my $size = shift;
1872 $avatar_cache{$email} ||=
1873 "http://www.gravatar.com/avatar/" .
1874 Digest::MD5::md5_hex($email) . "?s=";
1875 return $avatar_cache{$email} . $size;
1876 }
1877
1878 # Insert an avatar for the given $email at the given $size if the feature
1879 # is enabled.
1880 sub git_get_avatar {
1881 my ($email, %opts) = @_;
1882 my $pre_white = ($opts{-pad_before} ? "&nbsp;" : "");
1883 my $post_white = ($opts{-pad_after} ? "&nbsp;" : "");
1884 $opts{-size} ||= 'default';
1885 my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
1886 my $url = "";
1887 if ($git_avatar eq 'gravatar') {
1888 $url = gravatar_url($email, $size);
1889 } elsif ($git_avatar eq 'picon') {
1890 $url = picon_url($email);
1891 }
1892 # Other providers can be added by extending the if chain, defining $url
1893 # as needed. If no variant puts something in $url, we assume avatars
1894 # are completely disabled/unavailable.
1895 if ($url) {
1896 return $pre_white .
1897 "<img width=\"$size\" " .
1898 "class=\"avatar\" " .
1899 "src=\"".esc_url($url)."\" " .
1900 "alt=\"\" " .
1901 "/>" . $post_white;
1902 } else {
1903 return "";
1904 }
1905 }
1906
1907 sub format_search_author {
1908 my ($author, $searchtype, $displaytext) = @_;
1909 my $have_search = gitweb_check_feature('search');
1910
1911 if ($have_search) {
1912 my $performed = "";
1913 if ($searchtype eq 'author') {
1914 $performed = "authored";
1915 } elsif ($searchtype eq 'committer') {
1916 $performed = "committed";
1917 }
1918
1919 return $cgi->a({-href => href(action=>"search", hash=>$hash,
1920 searchtext=>$author,
1921 searchtype=>$searchtype), class=>"list",
1922 title=>"Search for commits $performed by $author"},
1923 $displaytext);
1924
1925 } else {
1926 return $displaytext;
1927 }
1928 }
1929
1930 # format the author name of the given commit with the given tag
1931 # the author name is chopped and escaped according to the other
1932 # optional parameters (see chop_str).
1933 sub format_author_html {
1934 my $tag = shift;
1935 my $co = shift;
1936 my $author = chop_and_escape_str($co->{'author_name'}, @_);
1937 return "<$tag class=\"author\">" .
1938 format_search_author($co->{'author_name'}, "author",
1939 git_get_avatar($co->{'author_email'}, -pad_after => 1) .
1940 $author) .
1941 "</$tag>";
1942 }
1943
1944 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
1945 sub format_git_diff_header_line {
1946 my $line = shift;
1947 my $diffinfo = shift;
1948 my ($from, $to) = @_;
1949
1950 if ($diffinfo->{'nparents'}) {
1951 # combined diff
1952 $line =~ s!^(diff (.*?) )"?.*$!$1!;
1953 if ($to->{'href'}) {
1954 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
1955 esc_path($to->{'file'}));
1956 } else { # file was deleted (no href)
1957 $line .= esc_path($to->{'file'});
1958 }
1959 } else {
1960 # "ordinary" diff
1961 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
1962 if ($from->{'href'}) {
1963 $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
1964 'a/' . esc_path($from->{'file'}));
1965 } else { # file was added (no href)
1966 $line .= 'a/' . esc_path($from->{'file'});
1967 }
1968 $line .= ' ';
1969 if ($to->{'href'}) {
1970 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
1971 'b/' . esc_path($to->{'file'}));
1972 } else { # file was deleted
1973 $line .= 'b/' . esc_path($to->{'file'});
1974 }
1975 }
1976
1977 return "<div class=\"diff header\">$line</div>\n";
1978 }
1979
1980 # format extended diff header line, before patch itself
1981 sub format_extended_diff_header_line {
1982 my $line = shift;
1983 my $diffinfo = shift;
1984 my ($from, $to) = @_;
1985
1986 # match <path>
1987 if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
1988 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
1989 esc_path($from->{'file'}));
1990 }
1991 if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
1992 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
1993 esc_path($to->{'file'}));
1994 }
1995 # match single <mode>
1996 if ($line =~ m/\s(\d{6})$/) {
1997 $line .= '<span class="info"> (' .
1998 file_type_long($1) .
1999 ')</span>';
2000 }
2001 # match <hash>
2002 if ($line =~ m/^index [0-9a-fA-F]{40},[0-9a-fA-F]{40}/) {
2003 # can match only for combined diff
2004 $line = 'index ';
2005 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2006 if ($from->{'href'}[$i]) {
2007 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2008 -class=>"hash"},
2009 substr($diffinfo->{'from_id'}[$i],0,7));
2010 } else {
2011 $line .= '0' x 7;
2012 }
2013 # separator
2014 $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2015 }
2016 $line .= '..';
2017 if ($to->{'href'}) {
2018 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2019 substr($diffinfo->{'to_id'},0,7));
2020 } else {
2021 $line .= '0' x 7;
2022 }
2023
2024 } elsif ($line =~ m/^index [0-9a-fA-F]{40}..[0-9a-fA-F]{40}/) {
2025 # can match only for ordinary diff
2026 my ($from_link, $to_link);
2027 if ($from->{'href'}) {
2028 $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2029 substr($diffinfo->{'from_id'},0,7));
2030 } else {
2031 $from_link = '0' x 7;
2032 }
2033 if ($to->{'href'}) {
2034 $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2035 substr($diffinfo->{'to_id'},0,7));
2036 } else {
2037 $to_link = '0' x 7;
2038 }
2039 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2040 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2041 }
2042
2043 return $line . "<br/>\n";
2044 }
2045
2046 # format from-file/to-file diff header
2047 sub format_diff_from_to_header {
2048 my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2049 my $line;
2050 my $result = '';
2051
2052 $line = $from_line;
2053 #assert($line =~ m/^---/) if DEBUG;
2054 # no extra formatting for "^--- /dev/null"
2055 if (! $diffinfo->{'nparents'}) {
2056 # ordinary (single parent) diff
2057 if ($line =~ m!^--- "?a/!) {
2058 if ($from->{'href'}) {
2059 $line = '--- a/' .
2060 $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2061 esc_path($from->{'file'}));
2062 } else {
2063 $line = '--- a/' .
2064 esc_path($from->{'file'});
2065 }
2066 }
2067 $result .= qq!<div class="diff from_file">$line</div>\n!;
2068
2069 } else {
2070 # combined diff (merge commit)
2071 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2072 if ($from->{'href'}[$i]) {
2073 $line = '--- ' .
2074 $cgi->a({-href=>href(action=>"blobdiff",
2075 hash_parent=>$diffinfo->{'from_id'}[$i],
2076 hash_parent_base=>$parents[$i],
2077 file_parent=>$from->{'file'}[$i],
2078 hash=>$diffinfo->{'to_id'},
2079 hash_base=>$hash,
2080 file_name=>$to->{'file'}),
2081 -class=>"path",
2082 -title=>"diff" . ($i+1)},
2083 $i+1) .
2084 '/' .
2085 $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2086 esc_path($from->{'file'}[$i]));
2087 } else {
2088 $line = '--- /dev/null';
2089 }
2090 $result .= qq!<div class="diff from_file">$line</div>\n!;
2091 }
2092 }
2093
2094 $line = $to_line;
2095 #assert($line =~ m/^\+\+\+/) if DEBUG;
2096 # no extra formatting for "^+++ /dev/null"
2097 if ($line =~ m!^\+\+\+ "?b/!) {
2098 if ($to->{'href'}) {
2099 $line = '+++ b/' .
2100 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2101 esc_path($to->{'file'}));
2102 } else {
2103 $line = '+++ b/' .
2104 esc_path($to->{'file'});
2105 }
2106 }
2107 $result .= qq!<div class="diff to_file">$line</div>\n!;
2108
2109 return $result;
2110 }
2111
2112 # create note for patch simplified by combined diff
2113 sub format_diff_cc_simplified {
2114 my ($diffinfo, @parents) = @_;
2115 my $result = '';
2116
2117 $result .= "<div class=\"diff header\">" .
2118 "diff --cc ";
2119 if (!is_deleted($diffinfo)) {
2120 $result .= $cgi->a({-href => href(action=>"blob",
2121 hash_base=>$hash,
2122 hash=>$diffinfo->{'to_id'},
2123 file_name=>$diffinfo->{'to_file'}),
2124 -class => "path"},
2125 esc_path($diffinfo->{'to_file'}));
2126 } else {
2127 $result .= esc_path($diffinfo->{'to_file'});
2128 }
2129 $result .= "</div>\n" . # class="diff header"
2130 "<div class=\"diff nodifferences\">" .
2131 "Simple merge" .
2132 "</div>\n"; # class="diff nodifferences"
2133
2134 return $result;
2135 }
2136
2137 # format patch (diff) line (not to be used for diff headers)
2138 sub format_diff_line {
2139 my $line = shift;
2140 my ($from, $to) = @_;
2141 my $diff_class = "";
2142
2143 chomp $line;
2144
2145 if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2146 # combined diff
2147 my $prefix = substr($line, 0, scalar @{$from->{'href'}});
2148 if ($line =~ m/^\@{3}/) {
2149 $diff_class = " chunk_header";
2150 } elsif ($line =~ m/^\\/) {
2151 $diff_class = " incomplete";
2152 } elsif ($prefix =~ tr/+/+/) {
2153 $diff_class = " add";
2154 } elsif ($prefix =~ tr/-/-/) {
2155 $diff_class = " rem";
2156 }
2157 } else {
2158 # assume ordinary diff
2159 my $char = substr($line, 0, 1);
2160 if ($char eq '+') {
2161 $diff_class = " add";
2162 } elsif ($char eq '-') {
2163 $diff_class = " rem";
2164 } elsif ($char eq '@') {
2165 $diff_class = " chunk_header";
2166 } elsif ($char eq "\\") {
2167 $diff_class = " incomplete";
2168 }
2169 }
2170 $line = untabify($line);
2171 if ($from && $to && $line =~ m/^\@{2} /) {
2172 my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2173 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2174
2175 $from_lines = 0 unless defined $from_lines;
2176 $to_lines = 0 unless defined $to_lines;
2177
2178 if ($from->{'href'}) {
2179 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2180 -class=>"list"}, $from_text);
2181 }
2182 if ($to->{'href'}) {
2183 $to_text = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2184 -class=>"list"}, $to_text);
2185 }
2186 $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2187 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2188 return "<div class=\"diff$diff_class\">$line</div>\n";
2189 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2190 my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2191 my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2192
2193 @from_text = split(' ', $ranges);
2194 for (my $i = 0; $i < @from_text; ++$i) {
2195 ($from_start[$i], $from_nlines[$i]) =
2196 (split(',', substr($from_text[$i], 1)), 0);
2197 }
2198
2199 $to_text = pop @from_text;
2200 $to_start = pop @from_start;
2201 $to_nlines = pop @from_nlines;
2202
2203 $line = "<span class=\"chunk_info\">$prefix ";
2204 for (my $i = 0; $i < @from_text; ++$i) {
2205 if ($from->{'href'}[$i]) {
2206 $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2207 -class=>"list"}, $from_text[$i]);
2208 } else {
2209 $line .= $from_text[$i];
2210 }
2211 $line .= " ";
2212 }
2213 if ($to->{'href'}) {
2214 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2215 -class=>"list"}, $to_text);
2216 } else {
2217 $line .= $to_text;
2218 }
2219 $line .= " $prefix</span>" .
2220 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2221 return "<div class=\"diff$diff_class\">$line</div>\n";
2222 }
2223 return "<div class=\"diff$diff_class\">" . esc_html($line, -nbsp=>1) . "</div>\n";
2224 }
2225
2226 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2227 # linked. Pass the hash of the tree/commit to snapshot.
2228 sub format_snapshot_links {
2229 my ($hash) = @_;
2230 my $num_fmts = @snapshot_fmts;
2231 if ($num_fmts > 1) {
2232 # A parenthesized list of links bearing format names.
2233 # e.g. "snapshot (_tar.gz_ _zip_)"
2234 return "snapshot (" . join(' ', map
2235 $cgi->a({
2236 -href => href(
2237 action=>"snapshot",
2238 hash=>$hash,
2239 snapshot_format=>$_
2240 )
2241 }, $known_snapshot_formats{$_}{'display'})
2242 , @snapshot_fmts) . ")";
2243 } elsif ($num_fmts == 1) {
2244 # A single "snapshot" link whose tooltip bears the format name.
2245 # i.e. "_snapshot_"
2246 my ($fmt) = @snapshot_fmts;
2247 return
2248 $cgi->a({
2249 -href => href(
2250 action=>"snapshot",
2251 hash=>$hash,
2252 snapshot_format=>$fmt
2253 ),
2254 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2255 }, "snapshot");
2256 } else { # $num_fmts == 0
2257 return undef;
2258 }
2259 }
2260
2261 ## ......................................................................
2262 ## functions returning values to be passed, perhaps after some
2263 ## transformation, to other functions; e.g. returning arguments to href()
2264
2265 # returns hash to be passed to href to generate gitweb URL
2266 # in -title key it returns description of link
2267 sub get_feed_info {
2268 my $format = shift || 'Atom';
2269 my %res = (action => lc($format));
2270
2271 # feed links are possible only for project views
2272 return unless (defined $project);
2273 # some views should link to OPML, or to generic project feed,
2274 # or don't have specific feed yet (so they should use generic)
2275 return if ($action =~ /^(?:tags|heads|forks|tag|search)$/x);
2276
2277 my $branch;
2278 # branches refs uses 'refs/heads/' prefix (fullname) to differentiate
2279 # from tag links; this also makes possible to detect branch links
2280 if ((defined $hash_base && $hash_base =~ m!^refs/heads/(.*)$!) ||
2281 (defined $hash && $hash =~ m!^refs/heads/(.*)$!)) {
2282 $branch = $1;
2283 }
2284 # find log type for feed description (title)
2285 my $type = 'log';
2286 if (defined $file_name) {
2287 $type = "history of $file_name";
2288 $type .= "/" if ($action eq 'tree');
2289 $type .= " on '$branch'" if (defined $branch);
2290 } else {
2291 $type = "log of $branch" if (defined $branch);
2292 }
2293
2294 $res{-title} = $type;
2295 $res{'hash'} = (defined $branch ? "refs/heads/$branch" : undef);
2296 $res{'file_name'} = $file_name;
2297
2298 return %res;
2299 }
2300
2301 ## ----------------------------------------------------------------------
2302 ## git utility subroutines, invoking git commands
2303
2304 # returns path to the core git executable and the --git-dir parameter as list
2305 sub git_cmd {
2306 $number_of_git_cmds++;
2307 return $GIT, '--git-dir='.$git_dir;
2308 }
2309
2310 # quote the given arguments for passing them to the shell
2311 # quote_command("command", "arg 1", "arg with ' and ! characters")
2312 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2313 # Try to avoid using this function wherever possible.
2314 sub quote_command {
2315 return join(' ',
2316 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2317 }
2318
2319 # get HEAD ref of given project as hash
2320 sub git_get_head_hash {
2321 return git_get_full_hash(shift, 'HEAD');
2322 }
2323
2324 sub git_get_full_hash {
2325 return git_get_hash(@_);
2326 }
2327
2328 sub git_get_short_hash {
2329 return git_get_hash(@_, '--short=7');
2330 }
2331
2332 sub git_get_hash {
2333 my ($project, $hash, @options) = @_;
2334 my $o_git_dir = $git_dir;
2335 my $retval = undef;
2336 $git_dir = "$projectroot/$project";
2337 if (open my $fd, '-|', git_cmd(), 'rev-parse',
2338 '--verify', '-q', @options, $hash) {
2339 $retval = <$fd>;
2340 chomp $retval if defined $retval;
2341 close $fd;
2342 }
2343 if (defined $o_git_dir) {
2344 $git_dir = $o_git_dir;
2345 }
2346 return $retval;
2347 }
2348
2349 # get type of given object
2350 sub git_get_type {
2351 my $hash = shift;
2352
2353 open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2354 my $type = <$fd>;
2355 close $fd or return;
2356 chomp $type;
2357 return $type;
2358 }
2359
2360 # repository configuration
2361 our $config_file = '';
2362 our %config;
2363
2364 # store multiple values for single key as anonymous array reference
2365 # single values stored directly in the hash, not as [ <value> ]
2366 sub hash_set_multi {
2367 my ($hash, $key, $value) = @_;
2368
2369 if (!exists $hash->{$key}) {
2370 $hash->{$key} = $value;
2371 } elsif (!ref $hash->{$key}) {
2372 $hash->{$key} = [ $hash->{$key}, $value ];
2373 } else {
2374 push @{$hash->{$key}}, $value;
2375 }
2376 }
2377
2378 # return hash of git project configuration
2379 # optionally limited to some section, e.g. 'gitweb'
2380 sub git_parse_project_config {
2381 my $section_regexp = shift;
2382 my %config;
2383
2384 local $/ = "\0";
2385
2386 open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2387 or return;
2388
2389 while (my $keyval = <$fh>) {
2390 chomp $keyval;
2391 my ($key, $value) = split(/\n/, $keyval, 2);
2392
2393 hash_set_multi(\%config, $key, $value)
2394 if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2395 }
2396 close $fh;
2397
2398 return %config;
2399 }
2400
2401 # convert config value to boolean: 'true' or 'false'
2402 # no value, number > 0, 'true' and 'yes' values are true
2403 # rest of values are treated as false (never as error)
2404 sub config_to_bool {
2405 my $val = shift;
2406
2407 return 1 if !defined $val; # section.key
2408
2409 # strip leading and trailing whitespace
2410 $val =~ s/^\s+//;
2411 $val =~ s/\s+$//;
2412
2413 return (($val =~ /^\d+$/ && $val) || # section.key = 1
2414 ($val =~ /^(?:true|yes)$/i)); # section.key = true
2415 }
2416
2417 # convert config value to simple decimal number
2418 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2419 # to be multiplied by 1024, 1048576, or 1073741824
2420 sub config_to_int {
2421 my $val = shift;
2422
2423 # strip leading and trailing whitespace
2424 $val =~ s/^\s+//;
2425 $val =~ s/\s+$//;
2426
2427 if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2428 $unit = lc($unit);
2429 # unknown unit is treated as 1
2430 return $num * ($unit eq 'g' ? 1073741824 :
2431 $unit eq 'm' ? 1048576 :
2432 $unit eq 'k' ? 1024 : 1);
2433 }
2434 return $val;
2435 }
2436
2437 # convert config value to array reference, if needed
2438 sub config_to_multi {
2439 my $val = shift;
2440
2441 return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2442 }
2443
2444 sub git_get_project_config {
2445 my ($key, $type) = @_;
2446
2447 return unless defined $git_dir;
2448
2449 # key sanity check
2450 return unless ($key);
2451 $key =~ s/^gitweb\.//;
2452 return if ($key =~ m/\W/);
2453
2454 # type sanity check
2455 if (defined $type) {
2456 $type =~ s/^--//;
2457 $type = undef
2458 unless ($type eq 'bool' || $type eq 'int');
2459 }
2460
2461 # get config
2462 if (!defined $config_file ||
2463 $config_file ne "$git_dir/config") {
2464 %config = git_parse_project_config('gitweb');
2465 $config_file = "$git_dir/config";
2466 }
2467
2468 # check if config variable (key) exists
2469 return unless exists $config{"gitweb.$key"};
2470
2471 # ensure given type
2472 if (!defined $type) {
2473 return $config{"gitweb.$key"};
2474 } elsif ($type eq 'bool') {
2475 # backward compatibility: 'git config --bool' returns true/false
2476 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2477 } elsif ($type eq 'int') {
2478 return config_to_int($config{"gitweb.$key"});
2479 }
2480 return $config{"gitweb.$key"};
2481 }
2482
2483 # get hash of given path at given ref
2484 sub git_get_hash_by_path {
2485 my $base = shift;
2486 my $path = shift || return undef;
2487 my $type = shift;
2488
2489 $path =~ s,/+$,,;
2490
2491 open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2492 or die_error(500, "Open git-ls-tree failed");
2493 my $line = <$fd>;
2494 close $fd or return undef;
2495
2496 if (!defined $line) {
2497 # there is no tree or hash given by $path at $base
2498 return undef;
2499 }
2500
2501 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
2502 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t/;
2503 if (defined $type && $type ne $2) {
2504 # type doesn't match
2505 return undef;
2506 }
2507 return $3;
2508 }
2509
2510 # get path of entry with given hash at given tree-ish (ref)
2511 # used to get 'from' filename for combined diff (merge commit) for renames
2512 sub git_get_path_by_hash {
2513 my $base = shift || return;
2514 my $hash = shift || return;
2515
2516 local $/ = "\0";
2517
2518 open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2519 or return undef;
2520 while (my $line = <$fd>) {
2521 chomp $line;
2522
2523 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423 gitweb'
2524 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f gitweb/README'
2525 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2526 close $fd;
2527 return $1;
2528 }
2529 }
2530 close $fd;
2531 return undef;
2532 }
2533
2534 ## ......................................................................
2535 ## git utility functions, directly accessing git repository
2536
2537 sub git_get_project_description {
2538 my $path = shift;
2539
2540 $git_dir = "$projectroot/$path";
2541 open my $fd, '<', "$git_dir/description"
2542 or return git_get_project_config('description');
2543 my $descr = <$fd>;
2544 close $fd;
2545 if (defined $descr) {
2546 chomp $descr;
2547 }
2548 return $descr;
2549 }
2550
2551 sub git_get_project_ctags {
2552 my $path = shift;
2553 my $ctags = {};
2554
2555 $git_dir = "$projectroot/$path";
2556 opendir my $dh, "$git_dir/ctags"
2557 or return $ctags;
2558 foreach (grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh)) {
2559 open my $ct, '<', $_ or next;
2560 my $val = <$ct>;
2561 chomp $val;
2562 close $ct;
2563 my $ctag = $_; $ctag =~ s#.*/##;
2564 $ctags->{$ctag} = $val;
2565 }
2566 closedir $dh;
2567 $ctags;
2568 }
2569
2570 sub git_populate_project_tagcloud {
2571 my $ctags = shift;
2572
2573 # First, merge different-cased tags; tags vote on casing
2574 my %ctags_lc;
2575 foreach (keys %$ctags) {
2576 $ctags_lc{lc $_}->{count} += $ctags->{$_};
2577 if (not $ctags_lc{lc $_}->{topcount}
2578 or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
2579 $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
2580 $ctags_lc{lc $_}->{topname} = $_;
2581 }
2582 }
2583
2584 my $cloud;
2585 if (eval { require HTML::TagCloud; 1; }) {
2586 $cloud = HTML::TagCloud->new;
2587 foreach (sort keys %ctags_lc) {
2588 # Pad the title with spaces so that the cloud looks
2589 # less crammed.
2590 my $title = $ctags_lc{$_}->{topname};
2591 $title =~ s/ /&nbsp;/g;
2592 $title =~ s/^/&nbsp;/g;
2593 $title =~ s/$/&nbsp;/g;
2594 $cloud->add($title, $home_link."?by_tag=".$_, $ctags_lc{$_}->{count});
2595 }
2596 } else {
2597 $cloud = \%ctags_lc;
2598 }
2599 $cloud;
2600 }
2601
2602 sub git_show_project_tagcloud {
2603 my ($cloud, $count) = @_;
2604 print STDERR ref($cloud)."..\n";
2605 if (ref $cloud eq 'HTML::TagCloud') {
2606 return $cloud->html_and_css($count);
2607 } else {
2608 my @tags = sort { $cloud->{$a}->{count} <=> $cloud->{$b}->{count} } keys %$cloud;
2609 return '<p align="center">' . join (', ', map {
2610 $cgi->a({-href=>"$home_link?by_tag=$_"}, $cloud->{$_}->{topname})
2611 } splice(@tags, 0, $count)) . '</p>';
2612 }
2613 }
2614
2615 sub git_get_project_url_list {
2616 my $path = shift;
2617
2618 $git_dir = "$projectroot/$path";
2619 open my $fd, '<', "$git_dir/cloneurl"
2620 or return wantarray ?
2621 @{ config_to_multi(git_get_project_config('url')) } :
2622 config_to_multi(git_get_project_config('url'));
2623 my @git_project_url_list = map { chomp; $_ } <$fd>;
2624 close $fd;
2625
2626 return wantarray ? @git_project_url_list : \@git_project_url_list;
2627 }
2628
2629 sub git_get_projects_list {
2630 my ($filter) = @_;
2631 my @list;
2632
2633 $filter ||= '';
2634 $filter =~ s/\.git$//;
2635
2636 my $check_forks = gitweb_check_feature('forks');
2637
2638 if (-d $projects_list) {
2639 # search in directory
2640 my $dir = $projects_list . ($filter ? "/$filter" : '');
2641 # remove the trailing "/"
2642 $dir =~ s!/+$!!;
2643 my $pfxlen = length("$dir");
2644 my $pfxdepth = ($dir =~ tr!/!!);
2645
2646 File::Find::find({
2647 follow_fast => 1, # follow symbolic links
2648 follow_skip => 2, # ignore duplicates
2649 dangling_symlinks => 0, # ignore dangling symlinks, silently
2650 wanted => sub {
2651 # global variables
2652 our $project_maxdepth;
2653 our $projectroot;
2654 # skip project-list toplevel, if we get it.
2655 return if (m!^[/.]$!);
2656 # only directories can be git repositories
2657 return unless (-d $_);
2658 # don't traverse too deep (Find is super slow on os x)
2659 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
2660 $File::Find::prune = 1;
2661 return;
2662 }
2663
2664 my $subdir = substr($File::Find::name, $pfxlen + 1);
2665 # we check related file in $projectroot
2666 my $path = ($filter ? "$filter/" : '') . $subdir;
2667 if (check_export_ok("$projectroot/$path")) {
2668 push @list, { path => $path };
2669 $File::Find::prune = 1;
2670 }
2671 },
2672 }, "$dir");
2673
2674 } elsif (-f $projects_list) {
2675 # read from file(url-encoded):
2676 # 'git%2Fgit.git Linus+Torvalds'
2677 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
2678 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
2679 my %paths;
2680 open my $fd, '<', $projects_list or return;
2681 PROJECT:
2682 while (my $line = <$fd>) {
2683 chomp $line;
2684 my ($path, $owner) = split ' ', $line;
2685 $path = unescape($path);
2686 $owner = unescape($owner);
2687 if (!defined $path) {
2688 next;
2689 }
2690 if ($filter ne '') {
2691 # looking for forks;
2692 my $pfx = substr($path, 0, length($filter));
2693 if ($pfx ne $filter) {
2694 next PROJECT;
2695 }
2696 my $sfx = substr($path, length($filter));
2697 if ($sfx !~ /^\/.*\.git$/) {
2698 next PROJECT;
2699 }
2700 } elsif ($check_forks) {
2701 PATH:
2702 foreach my $filter (keys %paths) {
2703 # looking for forks;
2704 my $pfx = substr($path, 0, length($filter));
2705 if ($pfx ne $filter) {
2706 next PATH;
2707 }
2708 my $sfx = substr($path, length($filter));
2709 if ($sfx !~ /^\/.*\.git$/) {
2710 next PATH;
2711 }
2712 # is a fork, don't include it in
2713 # the list
2714 next PROJECT;
2715 }
2716 }
2717 if (check_export_ok("$projectroot/$path")) {
2718 my $pr = {
2719 path => $path,
2720 owner => to_utf8($owner),
2721 };
2722 push @list, $pr;
2723 (my $forks_path = $path) =~ s/\.git$//;
2724 $paths{$forks_path}++;
2725 }
2726 }
2727 close $fd;
2728 }
2729 return @list;
2730 }
2731
2732 our $gitweb_project_owner = undef;
2733 sub git_get_project_list_from_file {
2734
2735 return if (defined $gitweb_project_owner);
2736
2737 $gitweb_project_owner = {};
2738 # read from file (url-encoded):
2739 # 'git%2Fgit.git Linus+Torvalds'
2740 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
2741 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
2742 if (-f $projects_list) {
2743 open(my $fd, '<', $projects_list);
2744 while (my $line = <$fd>) {
2745 chomp $line;
2746 my ($pr, $ow) = split ' ', $line;
2747 $pr = unescape($pr);
2748 $ow = unescape($ow);
2749 $gitweb_project_owner->{$pr} = to_utf8($ow);
2750 }
2751 close $fd;
2752 }
2753 }
2754
2755 sub git_get_project_owner {
2756 my $project = shift;
2757 my $owner;
2758
2759 return undef unless $project;
2760 $git_dir = "$projectroot/$project";
2761
2762 if (!defined $gitweb_project_owner) {
2763 git_get_project_list_from_file();
2764 }
2765
2766 if (exists $gitweb_project_owner->{$project}) {
2767 $owner = $gitweb_project_owner->{$project};
2768 }
2769 if (!defined $owner){
2770 $owner = git_get_project_config('owner');
2771 }
2772 if (!defined $owner) {
2773 $owner = get_file_owner("$git_dir");
2774 }
2775
2776 return $owner;
2777 }
2778
2779 sub git_get_last_activity {
2780 my ($path) = @_;
2781 my $fd;
2782
2783 $git_dir = "$projectroot/$path";
2784 open($fd, "-|", git_cmd(), 'for-each-ref',
2785 '--format=%(committer)',
2786 '--sort=-committerdate',
2787 '--count=1',
2788 'refs/heads') or return;
2789 my $most_recent = <$fd>;
2790 close $fd or return;
2791 if (defined $most_recent &&
2792 $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
2793 my $timestamp = $1;
2794 my $age = time - $timestamp;
2795 return ($age, age_string($age));
2796 }
2797 return (undef, undef);
2798 }
2799
2800 # Implementation note: when a single remote is wanted, we cannot use 'git
2801 # remote show -n' because that command always work (assuming it's a remote URL
2802 # if it's not defined), and we cannot use 'git remote show' because that would
2803 # try to make a network roundtrip. So the only way to find if that particular
2804 # remote is defined is to walk the list provided by 'git remote -v' and stop if
2805 # and when we find what we want.
2806 sub git_get_remotes_list {
2807 my $wanted = shift;
2808 my %remotes = ();
2809
2810 open my $fd, '-|' , git_cmd(), 'remote', '-v';
2811 return unless $fd;
2812 while (my $remote = <$fd>) {
2813 chomp $remote;
2814 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
2815 next if $wanted and not $remote eq $wanted;
2816 my ($url, $key) = ($1, $2);
2817
2818 $remotes{$remote} ||= { 'heads' => () };
2819 $remotes{$remote}{$key} = $url;
2820 }
2821 close $fd or return;
2822 return wantarray ? %remotes : \%remotes;
2823 }
2824
2825 # Takes a hash of remotes as first parameter and fills it by adding the
2826 # available remote heads for each of the indicated remotes.
2827 sub fill_remote_heads {
2828 my $remotes = shift;
2829 my @heads = map { "remotes/$_" } keys %$remotes;
2830 my @remoteheads = git_get_heads_list(undef, @heads);
2831 foreach my $remote (keys %$remotes) {
2832 $remotes->{$remote}{'heads'} = [ grep {
2833 $_->{'name'} =~ s!^$remote/!!
2834 } @remoteheads ];
2835 }
2836 }
2837
2838 sub git_get_references {
2839 my $type = shift || "";
2840 my %refs;
2841 # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
2842 # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
2843 open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
2844 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
2845 or return;
2846
2847 while (my $line = <$fd>) {
2848 chomp $line;
2849 if ($line =~ m!^([0-9a-fA-F]{40})\srefs/($type.*)$!) {
2850 if (defined $refs{$1}) {
2851 push @{$refs{$1}}, $2;
2852 } else {
2853 $refs{$1} = [ $2 ];
2854 }
2855 }
2856 }
2857 close $fd or return;
2858 return \%refs;
2859 }
2860
2861 sub git_get_rev_name_tags {
2862 my $hash = shift || return undef;
2863
2864 open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
2865 or return;
2866 my $name_rev = <$fd>;
2867 close $fd;
2868
2869 if ($name_rev =~ m|^$hash tags/(.*)$|) {
2870 return $1;
2871 } else {
2872 # catches also '$hash undefined' output
2873 return undef;
2874 }
2875 }
2876
2877 ## ----------------------------------------------------------------------
2878 ## parse to hash functions
2879
2880 sub parse_date {
2881 my $epoch = shift;
2882 my $tz = shift || "-0000";
2883
2884 my %date;
2885 my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
2886 my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
2887 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
2888 $date{'hour'} = $hour;
2889 $date{'minute'} = $min;
2890 $date{'mday'} = $mday;
2891 $date{'day'} = $days[$wday];
2892 $date{'month'} = $months[$mon];
2893 $date{'rfc2822'} = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
2894 $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
2895 $date{'mday-time'} = sprintf "%d %s %02d:%02d",
2896 $mday, $months[$mon], $hour ,$min;
2897 $date{'iso-8601'} = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
2898 1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
2899
2900 $tz =~ m/^([+\-][0-9][0-9])([0-9][0-9])$/;
2901 my $local = $epoch + ((int $1 + ($2/60)) * 3600);
2902 ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
2903 $date{'hour_local'} = $hour;
2904 $date{'minute_local'} = $min;
2905 $date{'tz_local'} = $tz;
2906 $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
2907 1900+$year, $mon+1, $mday,
2908 $hour, $min, $sec, $tz);
2909 return %date;
2910 }
2911
2912 sub parse_tag {
2913 my $tag_id = shift;
2914 my %tag;
2915 my @comment;
2916
2917 open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
2918 $tag{'id'} = $tag_id;
2919 while (my $line = <$fd>) {
2920 chomp $line;
2921 if ($line =~ m/^object ([0-9a-fA-F]{40})$/) {
2922 $tag{'object'} = $1;
2923 } elsif ($line =~ m/^type (.+)$/) {
2924 $tag{'type'} = $1;
2925 } elsif ($line =~ m/^tag (.+)$/) {
2926 $tag{'name'} = $1;
2927 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
2928 $tag{'author'} = $1;
2929 $tag{'author_epoch'} = $2;
2930 $tag{'author_tz'} = $3;
2931 if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
2932 $tag{'author_name'} = $1;
2933 $tag{'author_email'} = $2;
2934 } else {
2935 $tag{'author_name'} = $tag{'author'};
2936 }
2937 } elsif ($line =~ m/--BEGIN/) {
2938 push @comment, $line;
2939 last;
2940 } elsif ($line eq "") {
2941 last;
2942 }
2943 }
2944 push @comment, <$fd>;
2945 $tag{'comment'} = \@comment;
2946 close $fd or return;
2947 if (!defined $tag{'name'}) {
2948 return
2949 };
2950 return %tag
2951 }
2952
2953 sub parse_commit_text {
2954 my ($commit_text, $withparents) = @_;
2955 my @commit_lines = split '\n', $commit_text;
2956 my %co;
2957
2958 pop @commit_lines; # Remove '\0'
2959
2960 if (! @commit_lines) {
2961 return;
2962 }
2963
2964 my $header = shift @commit_lines;
2965 if ($header !~ m/^[0-9a-fA-F]{40}/) {
2966 return;
2967 }
2968 ($co{'id'}, my @parents) = split ' ', $header;
2969 while (my $line = shift @commit_lines) {
2970 last if $line eq "\n";
2971 if ($line =~ m/^tree ([0-9a-fA-F]{40})$/) {
2972 $co{'tree'} = $1;
2973 } elsif ((!defined $withparents) && ($line =~ m/^parent ([0-9a-fA-F]{40})$/)) {
2974 push @parents, $1;
2975 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
2976 $co{'author'} = to_utf8($1);
2977 $co{'author_epoch'} = $2;
2978 $co{'author_tz'} = $3;
2979 if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
2980 $co{'author_name'} = $1;
2981 $co{'author_email'} = $2;
2982 } else {
2983 $co{'author_name'} = $co{'author'};
2984 }
2985 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
2986 $co{'committer'} = to_utf8($1);
2987 $co{'committer_epoch'} = $2;
2988 $co{'committer_tz'} = $3;
2989 if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
2990 $co{'committer_name'} = $1;
2991 $co{'committer_email'} = $2;
2992 } else {
2993 $co{'committer_name'} = $co{'committer'};
2994 }
2995 }
2996 }
2997 if (!defined $co{'tree'}) {
2998 return;
2999 };
3000 $co{'parents'} = \@parents;
3001 $co{'parent'} = $parents[0];
3002
3003 foreach my $title (@commit_lines) {
3004 $title =~ s/^ //;
3005 if ($title ne "") {
3006 $co{'title'} = chop_str($title, 80, 5);
3007 # remove leading stuff of merges to make the interesting part visible
3008 if (length($title) > 50) {
3009 $title =~ s/^Automatic //;
3010 $title =~ s/^merge (of|with) /Merge ... /i;
3011 if (length($title) > 50) {
3012 $title =~ s/(http|rsync):\/\///;
3013 }
3014 if (length($title) > 50) {
3015 $title =~ s/(master|www|rsync)\.//;
3016 }
3017 if (length($title) > 50) {
3018 $title =~ s/kernel.org:?//;
3019 }
3020 if (length($title) > 50) {
3021 $title =~ s/\/pub\/scm//;
3022 }
3023 }
3024 $co{'title_short'} = chop_str($title, 50, 5);
3025 last;
3026 }
3027 }
3028 if (! defined $co{'title'} || $co{'title'} eq "") {
3029 $co{'title'} = $co{'title_short'} = '(no commit message)';
3030 }
3031 # remove added spaces
3032 foreach my $line (@commit_lines) {
3033 $line =~ s/^ //;
3034 }
3035 $co{'comment'} = \@commit_lines;
3036
3037 my $age = time - $co{'committer_epoch'};
3038 $co{'age'} = $age;
3039 $co{'age_string'} = age_string($age);
3040 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3041 if ($age > 60*60*24*7*2) {
3042 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3043 $co{'age_string_age'} = $co{'age_string'};
3044 } else {
3045 $co{'age_string_date'} = $co{'age_string'};
3046 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3047 }
3048 return %co;
3049 }
3050
3051 sub parse_commit {
3052 my ($commit_id) = @_;
3053 my %co;
3054
3055 local $/ = "\0";
3056
3057 open my $fd, "-|", git_cmd(), "rev-list",
3058 "--parents",
3059 "--header",
3060 "--max-count=1",
3061 $commit_id,
3062 "--",
3063 or die_error(500, "Open git-rev-list failed");
3064 %co = parse_commit_text(<$fd>, 1);
3065 close $fd;
3066
3067 return %co;
3068 }
3069
3070 sub parse_commits {
3071 my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3072 my @cos;
3073
3074 $maxcount ||= 1;
3075 $skip ||= 0;
3076
3077 local $/ = "\0";
3078
3079 open my $fd, "-|", git_cmd(), "rev-list",
3080 "--header",
3081 @args,
3082 ("--max-count=" . $maxcount),
3083 ("--skip=" . $skip),
3084 @extra_options,
3085 $commit_id,
3086 "--",
3087 ($filename ? ($filename) : ())
3088 or die_error(500, "Open git-rev-list failed");
3089 while (my $line = <$fd>) {
3090 my %co = parse_commit_text($line);
3091 push @cos, \%co;
3092 }
3093 close $fd;
3094
3095 return wantarray ? @cos : \@cos;
3096 }
3097
3098 # parse line of git-diff-tree "raw" output
3099 sub parse_difftree_raw_line {
3100 my $line = shift;
3101 my %res;
3102
3103 # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M ls-files.c'
3104 # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M rev-tree.c'
3105 if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ([0-9a-fA-F]{40}) ([0-9a-fA-F]{40}) (.)([0-9]{0,3})\t(.*)$/) {
3106 $res{'from_mode'} = $1;
3107 $res{'to_mode'} = $2;
3108 $res{'from_id'} = $3;
3109 $res{'to_id'} = $4;
3110 $res{'status'} = $5;
3111 $res{'similarity'} = $6;
3112 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3113 ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3114 } else {
3115 $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3116 }
3117 }
3118 # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3119 # combined diff (for merge commit)
3120 elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:[0-9a-fA-F]{40} )+)([a-zA-Z]+)\t(.*)$//) {
3121 $res{'nparents'} = length($1);
3122 $res{'from_mode'} = [ split(' ', $2) ];
3123 $res{'to_mode'} = pop @{$res{'from_mode'}};
3124 $res{'from_id'} = [ split(' ', $3) ];
3125 $res{'to_id'} = pop @{$res{'from_id'}};
3126 $res{'status'} = [ split('', $4) ];
3127 $res{'to_file'} = unquote($5);
3128 }
3129 # 'c512b523472485aef4fff9e57b229d9d243c967f'
3130 elsif ($line =~ m/^([0-9a-fA-F]{40})$/) {
3131 $res{'commit'} = $1;
3132 }
3133
3134 return wantarray ? %res : \%res;
3135 }
3136
3137 # wrapper: return parsed line of git-diff-tree "raw" output
3138 # (the argument might be raw line, or parsed info)
3139 sub parsed_difftree_line {
3140 my $line_or_ref = shift;
3141
3142 if (ref($line_or_ref) eq "HASH") {
3143 # pre-parsed (or generated by hand)
3144 return $line_or_ref;
3145 } else {
3146 return parse_difftree_raw_line($line_or_ref);
3147 }
3148 }
3149
3150 # parse line of git-ls-tree output
3151 sub parse_ls_tree_line {
3152 my $line = shift;
3153 my %opts = @_;
3154 my %res;
3155
3156 if ($opts{'-l'}) {
3157 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa 16717 panic.c'
3158 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40}) +(-|[0-9]+)\t(.+)$/s;
3159
3160 $res{'mode'} = $1;
3161 $res{'type'} = $2;
3162 $res{'hash'} = $3;
3163 $res{'size'} = $4;
3164 if ($opts{'-z'}) {
3165 $res{'name'} = $5;
3166 } else {
3167 $res{'name'} = unquote($5);
3168 }
3169 } else {
3170 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
3171 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t(.+)$/s;
3172
3173 $res{'mode'} = $1;
3174 $res{'type'} = $2;
3175 $res{'hash'} = $3;
3176 if ($opts{'-z'}) {
3177 $res{'name'} = $4;
3178 } else {
3179 $res{'name'} = unquote($4);
3180 }
3181 }
3182
3183 return wantarray ? %res : \%res;
3184 }
3185
3186 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3187 sub parse_from_to_diffinfo {
3188 my ($diffinfo, $from, $to, @parents) = @_;
3189
3190 if ($diffinfo->{'nparents'}) {
3191 # combined diff
3192 $from->{'file'} = [];
3193 $from->{'href'} = [];
3194 fill_from_file_info($diffinfo, @parents)
3195 unless exists $diffinfo->{'from_file'};
3196 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3197 $from->{'file'}[$i] =
3198 defined $diffinfo->{'from_file'}[$i] ?
3199 $diffinfo->{'from_file'}[$i] :
3200 $diffinfo->{'to_file'};
3201 if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3202 $from->{'href'}[$i] = href(action=>"blob",
3203 hash_base=>$parents[$i],
3204 hash=>$diffinfo->{'from_id'}[$i],
3205 file_name=>$from->{'file'}[$i]);
3206 } else {
3207 $from->{'href'}[$i] = undef;
3208 }
3209 }
3210 } else {
3211 # ordinary (not combined) diff
3212 $from->{'file'} = $diffinfo->{'from_file'};
3213 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3214 $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3215 hash=>$diffinfo->{'from_id'},
3216 file_name=>$from->{'file'});
3217 } else {
3218 delete $from->{'href'};
3219 }
3220 }
3221
3222 $to->{'file'} = $diffinfo->{'to_file'};
3223 if (!is_deleted($diffinfo)) { # file exists in result
3224 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3225 hash=>$diffinfo->{'to_id'},
3226 file_name=>$to->{'file'});
3227 } else {
3228 delete $to->{'href'};
3229 }
3230 }
3231
3232 ## ......................................................................
3233 ## parse to array of hashes functions
3234
3235 sub git_get_heads_list {
3236 my ($limit, @classes) = @_;
3237 @classes = ('heads') unless @classes;
3238 my @patterns = map { "refs/$_" } @classes;
3239 my @headslist;
3240
3241 open my $fd, '-|', git_cmd(), 'for-each-ref',
3242 ($limit ? '--count='.($limit+1) : ()), '--sort=-committerdate',
3243 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3244 @patterns
3245 or return;
3246 while (my $line = <$fd>) {
3247 my %ref_item;
3248
3249 chomp $line;
3250 my ($refinfo, $committerinfo) = split(/\0/, $line);
3251 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3252 my ($committer, $epoch, $tz) =
3253 ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3254 $ref_item{'fullname'} = $name;
3255 $name =~ s!^refs/(?:head|remote)s/!!;
3256
3257 $ref_item{'name'} = $name;
3258 $ref_item{'id'} = $hash;
3259 $ref_item{'title'} = $title || '(no commit message)';
3260 $ref_item{'epoch'} = $epoch;
3261 if ($epoch) {
3262 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3263 } else {
3264 $ref_item{'age'} = "unknown";
3265 }
3266
3267 push @headslist, \%ref_item;
3268 }
3269 close $fd;
3270
3271 return wantarray ? @headslist : \@headslist;
3272 }
3273
3274 sub git_get_tags_list {
3275 my $limit = shift;
3276 my @tagslist;
3277
3278 open my $fd, '-|', git_cmd(), 'for-each-ref',
3279 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3280 '--format=%(objectname) %(objecttype) %(refname) '.
3281 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3282 'refs/tags'
3283 or return;
3284 while (my $line = <$fd>) {
3285 my %ref_item;
3286
3287 chomp $line;
3288 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3289 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3290 my ($creator, $epoch, $tz) =
3291 ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3292 $ref_item{'fullname'} = $name;
3293 $name =~ s!^refs/tags/!!;
3294
3295 $ref_item{'type'} = $type;
3296 $ref_item{'id'} = $id;
3297 $ref_item{'name'} = $name;
3298 if ($type eq "tag") {
3299 $ref_item{'subject'} = $title;
3300 $ref_item{'reftype'} = $reftype;
3301 $ref_item{'refid'} = $refid;
3302 } else {
3303 $ref_item{'reftype'} = $type;
3304 $ref_item{'refid'} = $id;
3305 }
3306
3307 if ($type eq "tag" || $type eq "commit") {
3308 $ref_item{'epoch'} = $epoch;
3309 if ($epoch) {
3310 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3311 } else {
3312 $ref_item{'age'} = "unknown";
3313 }
3314 }
3315
3316 push @tagslist, \%ref_item;
3317 }
3318 close $fd;
3319
3320 return wantarray ? @tagslist : \@tagslist;
3321 }
3322
3323 ## ----------------------------------------------------------------------
3324 ## filesystem-related functions
3325
3326 sub get_file_owner {
3327 my $path = shift;
3328
3329 my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3330 my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3331 if (!defined $gcos) {
3332 return undef;
3333 }
3334 my $owner = $gcos;
3335 $owner =~ s/[,;].*$//;
3336 return to_utf8($owner);
3337 }
3338
3339 # assume that file exists
3340 sub insert_file {
3341 my $filename = shift;
3342
3343 open my $fd, '<', $filename;
3344 print map { to_utf8($_) } <$fd>;
3345 close $fd;
3346 }
3347
3348 ## ......................................................................
3349 ## mimetype related functions
3350
3351 sub mimetype_guess_file {
3352 my $filename = shift;
3353 my $mimemap = shift;
3354 -r $mimemap or return undef;
3355
3356 my %mimemap;
3357 open(my $mh, '<', $mimemap) or return undef;
3358 while (<$mh>) {
3359 next if m/^#/; # skip comments
3360 my ($mimetype, $exts) = split(/\t+/);
3361 if (defined $exts) {
3362 my @exts = split(/\s+/, $exts);
3363 foreach my $ext (@exts) {
3364 $mimemap{$ext} = $mimetype;
3365 }
3366 }
3367 }
3368 close($mh);
3369
3370 $filename =~ /\.([^.]*)$/;
3371 return $mimemap{$1};
3372 }
3373
3374 sub mimetype_guess {
3375 my $filename = shift;
3376 my $mime;
3377 $filename =~ /\./ or return undef;
3378
3379 if ($mimetypes_file) {
3380 my $file = $mimetypes_file;
3381 if ($file !~ m!^/!) { # if it is relative path
3382 # it is relative to project
3383 $file = "$projectroot/$project/$file";
3384 }
3385 $mime = mimetype_guess_file($filename, $file);
3386 }
3387 $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3388 return $mime;
3389 }
3390
3391 sub blob_mimetype {
3392 my $fd = shift;
3393 my $filename = shift;
3394
3395 if ($filename) {
3396 my $mime = mimetype_guess($filename);
3397 $mime and return $mime;
3398 }
3399
3400 # just in case
3401 return $default_blob_plain_mimetype unless $fd;
3402
3403 if (-T $fd) {
3404 return 'text/plain';
3405 } elsif (! $filename) {
3406 return 'application/octet-stream';
3407 } elsif ($filename =~ m/\.png$/i) {
3408 return 'image/png';
3409 } elsif ($filename =~ m/\.gif$/i) {
3410 return 'image/gif';
3411 } elsif ($filename =~ m/\.jpe?g$/i) {
3412 return 'image/jpeg';
3413 } else {
3414 return 'application/octet-stream';
3415 }
3416 }
3417
3418 sub blob_contenttype {
3419 my ($fd, $file_name, $type) = @_;
3420
3421 $type ||= blob_mimetype($fd, $file_name);
3422 if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3423 $type .= "; charset=$default_text_plain_charset";
3424 }
3425
3426 return $type;
3427 }
3428
3429 # guess file syntax for syntax highlighting; return undef if no highlighting
3430 # the name of syntax can (in the future) depend on syntax highlighter used
3431 sub guess_file_syntax {
3432 my ($highlight, $mimetype, $file_name) = @_;
3433 return undef unless ($highlight && defined $file_name);
3434 my $basename = basename($file_name, '.in');
3435 return $highlight_basename{$basename}
3436 if exists $highlight_basename{$basename};
3437
3438 $basename =~ /\.([^.]*)$/;
3439 my $ext = $1 or return undef;
3440 return $highlight_ext{$ext}
3441 if exists $highlight_ext{$ext};
3442
3443 return undef;
3444 }
3445
3446 # run highlighter and return FD of its output,
3447 # or return original FD if no highlighting
3448 sub run_highlighter {
3449 my ($fd, $highlight, $syntax) = @_;
3450 return $fd unless ($highlight && defined $syntax);
3451
3452 close $fd
3453 or die_error(404, "Reading blob failed");
3454 open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
3455 quote_command($highlight_bin).
3456 " --xhtml --fragment --syntax $syntax |"
3457 or die_error(500, "Couldn't open file or run syntax highlighter");
3458 return $fd;
3459 }
3460
3461 ## ======================================================================
3462 ## functions printing HTML: header, footer, error page
3463
3464 sub get_page_title {
3465 my $title = to_utf8($site_name);
3466
3467 return $title unless (defined $project);
3468 $title .= " - " . to_utf8($project);
3469
3470 return $title unless (defined $action);
3471 $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
3472
3473 return $title unless (defined $file_name);
3474 $title .= " - " . esc_path($file_name);
3475 if ($action eq "tree" && $file_name !~ m|/$|) {
3476 $title .= "/";
3477 }
3478
3479 return $title;
3480 }
3481
3482 sub git_header_html {
3483 my $status = shift || "200 OK";
3484 my $expires = shift;
3485 my %opts = @_;
3486
3487 my $title = get_page_title();
3488 my $content_type;
3489 # require explicit support from the UA if we are to send the page as
3490 # 'application/xhtml+xml', otherwise send it as plain old 'text/html'.
3491 # we have to do this because MSIE sometimes globs '*/*', pretending to
3492 # support xhtml+xml but choking when it gets what it asked for.
3493 if (defined $cgi->http('HTTP_ACCEPT') &&
3494 $cgi->http('HTTP_ACCEPT') =~ m/(,|;|\s|^)application\/xhtml\+xml(,|;|\s|$)/ &&
3495 $cgi->Accept('application/xhtml+xml') != 0) {
3496 $content_type = 'application/xhtml+xml';
3497 } else {
3498 $content_type = 'text/html';
3499 }
3500 print $cgi->header(-type=>$content_type, -charset => 'utf-8',
3501 -status=> $status, -expires => $expires)
3502 unless ($opts{'-no_http_header'});
3503 my $mod_perl_version = $ENV{'MOD_PERL'} ? " $ENV{'MOD_PERL'}" : '';
3504 print <<EOF;
3505 <?xml version="1.0" encoding="utf-8"?>
3506 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
3507 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en-US" lang="en-US">
3508 <!-- git web interface version $version, (C) 2005-2006, Kay Sievers <kay.sievers\@vrfy.org>, Christian Gierke -->
3509 <!-- git core binaries version $git_version -->
3510 <head>
3511 <meta http-equiv="content-type" content="$content_type; charset=utf-8"/>
3512 <meta name="generator" content="gitweb/$version git/$git_version$mod_perl_version"/>
3513 <meta name="robots" content="index, nofollow"/>
3514 <title>$title</title>
3515 EOF
3516 # the stylesheet, favicon etc urls won't work correctly with path_info
3517 # unless we set the appropriate base URL
3518 if ($ENV{'PATH_INFO'}) {
3519 print "<base href=\"".esc_url($base_url)."\" />\n";
3520 }
3521 # print out each stylesheet that exist, providing backwards capability
3522 # for those people who defined $stylesheet in a config file
3523 if (defined $stylesheet) {
3524 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
3525 } else {
3526 foreach my $stylesheet (@stylesheets) {
3527 next unless $stylesheet;
3528 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
3529 }
3530 }
3531 if (defined $project) {
3532 my