Sync with 1.7.4.2
[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 sql make),
254 # alternate extensions, see /etc/highlight/filetypes.conf
255 'h' => 'c',
256 map { $_ => 'sh' } qw(bash zsh ksh),
257 map { $_ => 'cpp' } qw(cxx c++ cc),
258 map { $_ => 'php' } qw(php3 php4 php5 phps),
259 map { $_ => 'pl' } qw(perl pm), # perhaps also 'cgi'
260 map { $_ => 'make'} qw(mak mk),
261 map { $_ => 'xml' } qw(xhtml html htm),
262 );
263
264 # You define site-wide feature defaults here; override them with
265 # $GITWEB_CONFIG as necessary.
266 our %feature = (
267 # feature => {
268 # 'sub' => feature-sub (subroutine),
269 # 'override' => allow-override (boolean),
270 # 'default' => [ default options...] (array reference)}
271 #
272 # if feature is overridable (it means that allow-override has true value),
273 # then feature-sub will be called with default options as parameters;
274 # return value of feature-sub indicates if to enable specified feature
275 #
276 # if there is no 'sub' key (no feature-sub), then feature cannot be
277 # overridden
278 #
279 # use gitweb_get_feature(<feature>) to retrieve the <feature> value
280 # (an array) or gitweb_check_feature(<feature>) to check if <feature>
281 # is enabled
282
283 # Enable the 'blame' blob view, showing the last commit that modified
284 # each line in the file. This can be very CPU-intensive.
285
286 # To enable system wide have in $GITWEB_CONFIG
287 # $feature{'blame'}{'default'} = [1];
288 # To have project specific config enable override in $GITWEB_CONFIG
289 # $feature{'blame'}{'override'} = 1;
290 # and in project config gitweb.blame = 0|1;
291 'blame' => {
292 'sub' => sub { feature_bool('blame', @_) },
293 'override' => 0,
294 'default' => [0]},
295
296 # Enable the 'snapshot' link, providing a compressed archive of any
297 # tree. This can potentially generate high traffic if you have large
298 # project.
299
300 # Value is a list of formats defined in %known_snapshot_formats that
301 # you wish to offer.
302 # To disable system wide have in $GITWEB_CONFIG
303 # $feature{'snapshot'}{'default'} = [];
304 # To have project specific config enable override in $GITWEB_CONFIG
305 # $feature{'snapshot'}{'override'} = 1;
306 # and in project config, a comma-separated list of formats or "none"
307 # to disable. Example: gitweb.snapshot = tbz2,zip;
308 'snapshot' => {
309 'sub' => \&feature_snapshot,
310 'override' => 0,
311 'default' => ['tgz']},
312
313 # Enable text search, which will list the commits which match author,
314 # committer or commit text to a given string. Enabled by default.
315 # Project specific override is not supported.
316 'search' => {
317 'override' => 0,
318 'default' => [1]},
319
320 # Enable grep search, which will list the files in currently selected
321 # tree containing the given string. Enabled by default. This can be
322 # potentially CPU-intensive, of course.
323
324 # To enable system wide have in $GITWEB_CONFIG
325 # $feature{'grep'}{'default'} = [1];
326 # To have project specific config enable override in $GITWEB_CONFIG
327 # $feature{'grep'}{'override'} = 1;
328 # and in project config gitweb.grep = 0|1;
329 'grep' => {
330 'sub' => sub { feature_bool('grep', @_) },
331 'override' => 0,
332 'default' => [1]},
333
334 # Enable the pickaxe search, which will list the commits that modified
335 # a given string in a file. This can be practical and quite faster
336 # alternative to 'blame', but still potentially CPU-intensive.
337
338 # To enable system wide have in $GITWEB_CONFIG
339 # $feature{'pickaxe'}{'default'} = [1];
340 # To have project specific config enable override in $GITWEB_CONFIG
341 # $feature{'pickaxe'}{'override'} = 1;
342 # and in project config gitweb.pickaxe = 0|1;
343 'pickaxe' => {
344 'sub' => sub { feature_bool('pickaxe', @_) },
345 'override' => 0,
346 'default' => [1]},
347
348 # Enable showing size of blobs in a 'tree' view, in a separate
349 # column, similar to what 'ls -l' does. This cost a bit of IO.
350
351 # To disable system wide have in $GITWEB_CONFIG
352 # $feature{'show-sizes'}{'default'} = [0];
353 # To have project specific config enable override in $GITWEB_CONFIG
354 # $feature{'show-sizes'}{'override'} = 1;
355 # and in project config gitweb.showsizes = 0|1;
356 'show-sizes' => {
357 'sub' => sub { feature_bool('showsizes', @_) },
358 'override' => 0,
359 'default' => [1]},
360
361 # Make gitweb use an alternative format of the URLs which can be
362 # more readable and natural-looking: project name is embedded
363 # directly in the path and the query string contains other
364 # auxiliary information. All gitweb installations recognize
365 # URL in either format; this configures in which formats gitweb
366 # generates links.
367
368 # To enable system wide have in $GITWEB_CONFIG
369 # $feature{'pathinfo'}{'default'} = [1];
370 # Project specific override is not supported.
371
372 # Note that you will need to change the default location of CSS,
373 # favicon, logo and possibly other files to an absolute URL. Also,
374 # if gitweb.cgi serves as your indexfile, you will need to force
375 # $my_uri to contain the script name in your $GITWEB_CONFIG.
376 'pathinfo' => {
377 'override' => 0,
378 'default' => [0]},
379
380 # Make gitweb consider projects in project root subdirectories
381 # to be forks of existing projects. Given project $projname.git,
382 # projects matching $projname/*.git will not be shown in the main
383 # projects list, instead a '+' mark will be added to $projname
384 # there and a 'forks' view will be enabled for the project, listing
385 # all the forks. If project list is taken from a file, forks have
386 # to be listed after the main project.
387
388 # To enable system wide have in $GITWEB_CONFIG
389 # $feature{'forks'}{'default'} = [1];
390 # Project specific override is not supported.
391 'forks' => {
392 'override' => 0,
393 'default' => [0]},
394
395 # Insert custom links to the action bar of all project pages.
396 # This enables you mainly to link to third-party scripts integrating
397 # into gitweb; e.g. git-browser for graphical history representation
398 # or custom web-based repository administration interface.
399
400 # The 'default' value consists of a list of triplets in the form
401 # (label, link, position) where position is the label after which
402 # to insert the link and link is a format string where %n expands
403 # to the project name, %f to the project path within the filesystem,
404 # %h to the current hash (h gitweb parameter) and %b to the current
405 # hash base (hb gitweb parameter); %% expands to %.
406
407 # To enable system wide have in $GITWEB_CONFIG e.g.
408 # $feature{'actions'}{'default'} = [('graphiclog',
409 # '/git-browser/by-commit.html?r=%n', 'summary')];
410 # Project specific override is not supported.
411 'actions' => {
412 'override' => 0,
413 'default' => []},
414
415 # Allow gitweb scan project content tags described in ctags/
416 # of project repository, and display the popular Web 2.0-ish
417 # "tag cloud" near the project list. Note that this is something
418 # COMPLETELY different from the normal Git tags.
419
420 # gitweb by itself can show existing tags, but it does not handle
421 # tagging itself; you need an external application for that.
422 # For an example script, check Girocco's cgi/tagproj.cgi.
423 # You may want to install the HTML::TagCloud Perl module to get
424 # a pretty tag cloud instead of just a list of tags.
425
426 # To enable system wide have in $GITWEB_CONFIG
427 # $feature{'ctags'}{'default'} = ['path_to_tag_script'];
428 # Project specific override is not supported.
429 'ctags' => {
430 'override' => 0,
431 'default' => [0]},
432
433 # The maximum number of patches in a patchset generated in patch
434 # view. Set this to 0 or undef to disable patch view, or to a
435 # negative number to remove any limit.
436
437 # To disable system wide have in $GITWEB_CONFIG
438 # $feature{'patches'}{'default'} = [0];
439 # To have project specific config enable override in $GITWEB_CONFIG
440 # $feature{'patches'}{'override'} = 1;
441 # and in project config gitweb.patches = 0|n;
442 # where n is the maximum number of patches allowed in a patchset.
443 'patches' => {
444 'sub' => \&feature_patches,
445 'override' => 0,
446 'default' => [16]},
447
448 # Avatar support. When this feature is enabled, views such as
449 # shortlog or commit will display an avatar associated with
450 # the email of the committer(s) and/or author(s).
451
452 # Currently available providers are gravatar and picon.
453 # If an unknown provider is specified, the feature is disabled.
454
455 # Gravatar depends on Digest::MD5.
456 # Picon currently relies on the indiana.edu database.
457
458 # To enable system wide have in $GITWEB_CONFIG
459 # $feature{'avatar'}{'default'} = ['<provider>'];
460 # where <provider> is either gravatar or picon.
461 # To have project specific config enable override in $GITWEB_CONFIG
462 # $feature{'avatar'}{'override'} = 1;
463 # and in project config gitweb.avatar = <provider>;
464 'avatar' => {
465 'sub' => \&feature_avatar,
466 'override' => 0,
467 'default' => ['']},
468
469 # Enable displaying how much time and how many git commands
470 # it took to generate and display page. Disabled by default.
471 # Project specific override is not supported.
472 'timed' => {
473 'override' => 0,
474 'default' => [0]},
475
476 # Enable turning some links into links to actions which require
477 # JavaScript to run (like 'blame_incremental'). Not enabled by
478 # default. Project specific override is currently not supported.
479 'javascript-actions' => {
480 'override' => 0,
481 'default' => [0]},
482
483 # Syntax highlighting support. This is based on Daniel Svensson's
484 # and Sham Chukoury's work in gitweb-xmms2.git.
485 # It requires the 'highlight' program present in $PATH,
486 # and therefore is disabled by default.
487
488 # To enable system wide have in $GITWEB_CONFIG
489 # $feature{'highlight'}{'default'} = [1];
490
491 'highlight' => {
492 'sub' => sub { feature_bool('highlight', @_) },
493 'override' => 0,
494 'default' => [0]},
495
496 # Enable displaying of remote heads in the heads list
497
498 # To enable system wide have in $GITWEB_CONFIG
499 # $feature{'remote_heads'}{'default'} = [1];
500 # To have project specific config enable override in $GITWEB_CONFIG
501 # $feature{'remote_heads'}{'override'} = 1;
502 # and in project config gitweb.remote_heads = 0|1;
503 'remote_heads' => {
504 'sub' => sub { feature_bool('remote_heads', @_) },
505 'override' => 0,
506 'default' => [0]},
507 );
508
509 sub gitweb_get_feature {
510 my ($name) = @_;
511 return unless exists $feature{$name};
512 my ($sub, $override, @defaults) = (
513 $feature{$name}{'sub'},
514 $feature{$name}{'override'},
515 @{$feature{$name}{'default'}});
516 # project specific override is possible only if we have project
517 our $git_dir; # global variable, declared later
518 if (!$override || !defined $git_dir) {
519 return @defaults;
520 }
521 if (!defined $sub) {
522 warn "feature $name is not overridable";
523 return @defaults;
524 }
525 return $sub->(@defaults);
526 }
527
528 # A wrapper to check if a given feature is enabled.
529 # With this, you can say
530 #
531 # my $bool_feat = gitweb_check_feature('bool_feat');
532 # gitweb_check_feature('bool_feat') or somecode;
533 #
534 # instead of
535 #
536 # my ($bool_feat) = gitweb_get_feature('bool_feat');
537 # (gitweb_get_feature('bool_feat'))[0] or somecode;
538 #
539 sub gitweb_check_feature {
540 return (gitweb_get_feature(@_))[0];
541 }
542
543
544 sub feature_bool {
545 my $key = shift;
546 my ($val) = git_get_project_config($key, '--bool');
547
548 if (!defined $val) {
549 return ($_[0]);
550 } elsif ($val eq 'true') {
551 return (1);
552 } elsif ($val eq 'false') {
553 return (0);
554 }
555 }
556
557 sub feature_snapshot {
558 my (@fmts) = @_;
559
560 my ($val) = git_get_project_config('snapshot');
561
562 if ($val) {
563 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
564 }
565
566 return @fmts;
567 }
568
569 sub feature_patches {
570 my @val = (git_get_project_config('patches', '--int'));
571
572 if (@val) {
573 return @val;
574 }
575
576 return ($_[0]);
577 }
578
579 sub feature_avatar {
580 my @val = (git_get_project_config('avatar'));
581
582 return @val ? @val : @_;
583 }
584
585 # checking HEAD file with -e is fragile if the repository was
586 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
587 # and then pruned.
588 sub check_head_link {
589 my ($dir) = @_;
590 my $headfile = "$dir/HEAD";
591 return ((-e $headfile) ||
592 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
593 }
594
595 sub check_export_ok {
596 my ($dir) = @_;
597 return (check_head_link($dir) &&
598 (!$export_ok || -e "$dir/$export_ok") &&
599 (!$export_auth_hook || $export_auth_hook->($dir)));
600 }
601
602 # process alternate names for backward compatibility
603 # filter out unsupported (unknown) snapshot formats
604 sub filter_snapshot_fmts {
605 my @fmts = @_;
606
607 @fmts = map {
608 exists $known_snapshot_format_aliases{$_} ?
609 $known_snapshot_format_aliases{$_} : $_} @fmts;
610 @fmts = grep {
611 exists $known_snapshot_formats{$_} &&
612 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
613 }
614
615 # If it is set to code reference, it is code that it is to be run once per
616 # request, allowing updating configurations that change with each request,
617 # while running other code in config file only once.
618 #
619 # Otherwise, if it is false then gitweb would process config file only once;
620 # if it is true then gitweb config would be run for each request.
621 our $per_request_config = 1;
622
623 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM);
624 sub evaluate_gitweb_config {
625 our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
626 our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
627 # die if there are errors parsing config file
628 if (-e $GITWEB_CONFIG) {
629 do $GITWEB_CONFIG;
630 die $@ if $@;
631 } elsif (-e $GITWEB_CONFIG_SYSTEM) {
632 do $GITWEB_CONFIG_SYSTEM;
633 die $@ if $@;
634 }
635 }
636
637 # Get loadavg of system, to compare against $maxload.
638 # Currently it requires '/proc/loadavg' present to get loadavg;
639 # if it is not present it returns 0, which means no load checking.
640 sub get_loadavg {
641 if( -e '/proc/loadavg' ){
642 open my $fd, '<', '/proc/loadavg'
643 or return 0;
644 my @load = split(/\s+/, scalar <$fd>);
645 close $fd;
646
647 # The first three columns measure CPU and IO utilization of the last one,
648 # five, and 10 minute periods. The fourth column shows the number of
649 # currently running processes and the total number of processes in the m/n
650 # format. The last column displays the last process ID used.
651 return $load[0] || 0;
652 }
653 # additional checks for load average should go here for things that don't export
654 # /proc/loadavg
655
656 return 0;
657 }
658
659 # version of the core git binary
660 our $git_version;
661 sub evaluate_git_version {
662 our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
663 $number_of_git_cmds++;
664 }
665
666 sub check_loadavg {
667 if (defined $maxload && get_loadavg() > $maxload) {
668 die_error(503, "The load average on the server is too high");
669 }
670 }
671
672 # ======================================================================
673 # input validation and dispatch
674
675 # input parameters can be collected from a variety of sources (presently, CGI
676 # and PATH_INFO), so we define an %input_params hash that collects them all
677 # together during validation: this allows subsequent uses (e.g. href()) to be
678 # agnostic of the parameter origin
679
680 our %input_params = ();
681
682 # input parameters are stored with the long parameter name as key. This will
683 # also be used in the href subroutine to convert parameters to their CGI
684 # equivalent, and since the href() usage is the most frequent one, we store
685 # the name -> CGI key mapping here, instead of the reverse.
686 #
687 # XXX: Warning: If you touch this, check the search form for updating,
688 # too.
689
690 our @cgi_param_mapping = (
691 project => "p",
692 action => "a",
693 file_name => "f",
694 file_parent => "fp",
695 hash => "h",
696 hash_parent => "hp",
697 hash_base => "hb",
698 hash_parent_base => "hpb",
699 page => "pg",
700 order => "o",
701 searchtext => "s",
702 searchtype => "st",
703 snapshot_format => "sf",
704 extra_options => "opt",
705 search_use_regexp => "sr",
706 # this must be last entry (for manipulation from JavaScript)
707 javascript => "js"
708 );
709 our %cgi_param_mapping = @cgi_param_mapping;
710
711 # we will also need to know the possible actions, for validation
712 our %actions = (
713 "blame" => \&git_blame,
714 "blame_incremental" => \&git_blame_incremental,
715 "blame_data" => \&git_blame_data,
716 "blobdiff" => \&git_blobdiff,
717 "blobdiff_plain" => \&git_blobdiff_plain,
718 "blob" => \&git_blob,
719 "blob_plain" => \&git_blob_plain,
720 "commitdiff" => \&git_commitdiff,
721 "commitdiff_plain" => \&git_commitdiff_plain,
722 "commit" => \&git_commit,
723 "forks" => \&git_forks,
724 "heads" => \&git_heads,
725 "history" => \&git_history,
726 "log" => \&git_log,
727 "patch" => \&git_patch,
728 "patches" => \&git_patches,
729 "remotes" => \&git_remotes,
730 "rss" => \&git_rss,
731 "atom" => \&git_atom,
732 "search" => \&git_search,
733 "search_help" => \&git_search_help,
734 "shortlog" => \&git_shortlog,
735 "summary" => \&git_summary,
736 "tag" => \&git_tag,
737 "tags" => \&git_tags,
738 "tree" => \&git_tree,
739 "snapshot" => \&git_snapshot,
740 "object" => \&git_object,
741 # those below don't need $project
742 "opml" => \&git_opml,
743 "project_list" => \&git_project_list,
744 "project_index" => \&git_project_index,
745 );
746
747 # finally, we have the hash of allowed extra_options for the commands that
748 # allow them
749 our %allowed_options = (
750 "--no-merges" => [ qw(rss atom log shortlog history) ],
751 );
752
753 # fill %input_params with the CGI parameters. All values except for 'opt'
754 # should be single values, but opt can be an array. We should probably
755 # build an array of parameters that can be multi-valued, but since for the time
756 # being it's only this one, we just single it out
757 sub evaluate_query_params {
758 our $cgi;
759
760 while (my ($name, $symbol) = each %cgi_param_mapping) {
761 if ($symbol eq 'opt') {
762 $input_params{$name} = [ $cgi->param($symbol) ];
763 } else {
764 $input_params{$name} = $cgi->param($symbol);
765 }
766 }
767 }
768
769 # now read PATH_INFO and update the parameter list for missing parameters
770 sub evaluate_path_info {
771 return if defined $input_params{'project'};
772 return if !$path_info;
773 $path_info =~ s,^/+,,;
774 return if !$path_info;
775
776 # find which part of PATH_INFO is project
777 my $project = $path_info;
778 $project =~ s,/+$,,;
779 while ($project && !check_head_link("$projectroot/$project")) {
780 $project =~ s,/*[^/]*$,,;
781 }
782 return unless $project;
783 $input_params{'project'} = $project;
784
785 # do not change any parameters if an action is given using the query string
786 return if $input_params{'action'};
787 $path_info =~ s,^\Q$project\E/*,,;
788
789 # next, check if we have an action
790 my $action = $path_info;
791 $action =~ s,/.*$,,;
792 if (exists $actions{$action}) {
793 $path_info =~ s,^$action/*,,;
794 $input_params{'action'} = $action;
795 }
796
797 # list of actions that want hash_base instead of hash, but can have no
798 # pathname (f) parameter
799 my @wants_base = (
800 'tree',
801 'history',
802 );
803
804 # we want to catch, among others
805 # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
806 my ($parentrefname, $parentpathname, $refname, $pathname) =
807 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
808
809 # first, analyze the 'current' part
810 if (defined $pathname) {
811 # we got "branch:filename" or "branch:dir/"
812 # we could use git_get_type(branch:pathname), but:
813 # - it needs $git_dir
814 # - it does a git() call
815 # - the convention of terminating directories with a slash
816 # makes it superfluous
817 # - embedding the action in the PATH_INFO would make it even
818 # more superfluous
819 $pathname =~ s,^/+,,;
820 if (!$pathname || substr($pathname, -1) eq "/") {
821 $input_params{'action'} ||= "tree";
822 $pathname =~ s,/$,,;
823 } else {
824 # the default action depends on whether we had parent info
825 # or not
826 if ($parentrefname) {
827 $input_params{'action'} ||= "blobdiff_plain";
828 } else {
829 $input_params{'action'} ||= "blob_plain";
830 }
831 }
832 $input_params{'hash_base'} ||= $refname;
833 $input_params{'file_name'} ||= $pathname;
834 } elsif (defined $refname) {
835 # we got "branch". In this case we have to choose if we have to
836 # set hash or hash_base.
837 #
838 # Most of the actions without a pathname only want hash to be
839 # set, except for the ones specified in @wants_base that want
840 # hash_base instead. It should also be noted that hand-crafted
841 # links having 'history' as an action and no pathname or hash
842 # set will fail, but that happens regardless of PATH_INFO.
843 if (defined $parentrefname) {
844 # if there is parent let the default be 'shortlog' action
845 # (for http://git.example.com/repo.git/A..B links); if there
846 # is no parent, dispatch will detect type of object and set
847 # action appropriately if required (if action is not set)
848 $input_params{'action'} ||= "shortlog";
849 }
850 if ($input_params{'action'} &&
851 grep { $_ eq $input_params{'action'} } @wants_base) {
852 $input_params{'hash_base'} ||= $refname;
853 } else {
854 $input_params{'hash'} ||= $refname;
855 }
856 }
857
858 # next, handle the 'parent' part, if present
859 if (defined $parentrefname) {
860 # a missing pathspec defaults to the 'current' filename, allowing e.g.
861 # someproject/blobdiff/oldrev..newrev:/filename
862 if ($parentpathname) {
863 $parentpathname =~ s,^/+,,;
864 $parentpathname =~ s,/$,,;
865 $input_params{'file_parent'} ||= $parentpathname;
866 } else {
867 $input_params{'file_parent'} ||= $input_params{'file_name'};
868 }
869 # we assume that hash_parent_base is wanted if a path was specified,
870 # or if the action wants hash_base instead of hash
871 if (defined $input_params{'file_parent'} ||
872 grep { $_ eq $input_params{'action'} } @wants_base) {
873 $input_params{'hash_parent_base'} ||= $parentrefname;
874 } else {
875 $input_params{'hash_parent'} ||= $parentrefname;
876 }
877 }
878
879 # for the snapshot action, we allow URLs in the form
880 # $project/snapshot/$hash.ext
881 # where .ext determines the snapshot and gets removed from the
882 # passed $refname to provide the $hash.
883 #
884 # To be able to tell that $refname includes the format extension, we
885 # require the following two conditions to be satisfied:
886 # - the hash input parameter MUST have been set from the $refname part
887 # of the URL (i.e. they must be equal)
888 # - the snapshot format MUST NOT have been defined already (e.g. from
889 # CGI parameter sf)
890 # It's also useless to try any matching unless $refname has a dot,
891 # so we check for that too
892 if (defined $input_params{'action'} &&
893 $input_params{'action'} eq 'snapshot' &&
894 defined $refname && index($refname, '.') != -1 &&
895 $refname eq $input_params{'hash'} &&
896 !defined $input_params{'snapshot_format'}) {
897 # We loop over the known snapshot formats, checking for
898 # extensions. Allowed extensions are both the defined suffix
899 # (which includes the initial dot already) and the snapshot
900 # format key itself, with a prepended dot
901 while (my ($fmt, $opt) = each %known_snapshot_formats) {
902 my $hash = $refname;
903 unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
904 next;
905 }
906 my $sfx = $1;
907 # a valid suffix was found, so set the snapshot format
908 # and reset the hash parameter
909 $input_params{'snapshot_format'} = $fmt;
910 $input_params{'hash'} = $hash;
911 # we also set the format suffix to the one requested
912 # in the URL: this way a request for e.g. .tgz returns
913 # a .tgz instead of a .tar.gz
914 $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
915 last;
916 }
917 }
918 }
919
920 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
921 $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
922 $searchtext, $search_regexp);
923 sub evaluate_and_validate_params {
924 our $action = $input_params{'action'};
925 if (defined $action) {
926 if (!validate_action($action)) {
927 die_error(400, "Invalid action parameter");
928 }
929 }
930
931 # parameters which are pathnames
932 our $project = $input_params{'project'};
933 if (defined $project) {
934 if (!validate_project($project)) {
935 undef $project;
936 die_error(404, "No such project");
937 }
938 }
939
940 our $file_name = $input_params{'file_name'};
941 if (defined $file_name) {
942 if (!validate_pathname($file_name)) {
943 die_error(400, "Invalid file parameter");
944 }
945 }
946
947 our $file_parent = $input_params{'file_parent'};
948 if (defined $file_parent) {
949 if (!validate_pathname($file_parent)) {
950 die_error(400, "Invalid file parent parameter");
951 }
952 }
953
954 # parameters which are refnames
955 our $hash = $input_params{'hash'};
956 if (defined $hash) {
957 if (!validate_refname($hash)) {
958 die_error(400, "Invalid hash parameter");
959 }
960 }
961
962 our $hash_parent = $input_params{'hash_parent'};
963 if (defined $hash_parent) {
964 if (!validate_refname($hash_parent)) {
965 die_error(400, "Invalid hash parent parameter");
966 }
967 }
968
969 our $hash_base = $input_params{'hash_base'};
970 if (defined $hash_base) {
971 if (!validate_refname($hash_base)) {
972 die_error(400, "Invalid hash base parameter");
973 }
974 }
975
976 our @extra_options = @{$input_params{'extra_options'}};
977 # @extra_options is always defined, since it can only be (currently) set from
978 # CGI, and $cgi->param() returns the empty array in array context if the param
979 # is not set
980 foreach my $opt (@extra_options) {
981 if (not exists $allowed_options{$opt}) {
982 die_error(400, "Invalid option parameter");
983 }
984 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
985 die_error(400, "Invalid option parameter for this action");
986 }
987 }
988
989 our $hash_parent_base = $input_params{'hash_parent_base'};
990 if (defined $hash_parent_base) {
991 if (!validate_refname($hash_parent_base)) {
992 die_error(400, "Invalid hash parent base parameter");
993 }
994 }
995
996 # other parameters
997 our $page = $input_params{'page'};
998 if (defined $page) {
999 if ($page =~ m/[^0-9]/) {
1000 die_error(400, "Invalid page parameter");
1001 }
1002 }
1003
1004 our $searchtype = $input_params{'searchtype'};
1005 if (defined $searchtype) {
1006 if ($searchtype =~ m/[^a-z]/) {
1007 die_error(400, "Invalid searchtype parameter");
1008 }
1009 }
1010
1011 our $search_use_regexp = $input_params{'search_use_regexp'};
1012
1013 our $searchtext = $input_params{'searchtext'};
1014 our $search_regexp;
1015 if (defined $searchtext) {
1016 if (length($searchtext) < 2) {
1017 die_error(403, "At least two characters are required for search parameter");
1018 }
1019 $search_regexp = $search_use_regexp ? $searchtext : quotemeta $searchtext;
1020 }
1021 }
1022
1023 # path to the current git repository
1024 our $git_dir;
1025 sub evaluate_git_dir {
1026 our $git_dir = "$projectroot/$project" if $project;
1027 }
1028
1029 our (@snapshot_fmts, $git_avatar);
1030 sub configure_gitweb_features {
1031 # list of supported snapshot formats
1032 our @snapshot_fmts = gitweb_get_feature('snapshot');
1033 @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1034
1035 # check that the avatar feature is set to a known provider name,
1036 # and for each provider check if the dependencies are satisfied.
1037 # if the provider name is invalid or the dependencies are not met,
1038 # reset $git_avatar to the empty string.
1039 our ($git_avatar) = gitweb_get_feature('avatar');
1040 if ($git_avatar eq 'gravatar') {
1041 $git_avatar = '' unless (eval { require Digest::MD5; 1; });
1042 } elsif ($git_avatar eq 'picon') {
1043 # no dependencies
1044 } else {
1045 $git_avatar = '';
1046 }
1047 }
1048
1049 # custom error handler: 'die <message>' is Internal Server Error
1050 sub handle_errors_html {
1051 my $msg = shift; # it is already HTML escaped
1052
1053 # to avoid infinite loop where error occurs in die_error,
1054 # change handler to default handler, disabling handle_errors_html
1055 set_message("Error occured when inside die_error:\n$msg");
1056
1057 # you cannot jump out of die_error when called as error handler;
1058 # the subroutine set via CGI::Carp::set_message is called _after_
1059 # HTTP headers are already written, so it cannot write them itself
1060 die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1061 }
1062 set_message(\&handle_errors_html);
1063
1064 # dispatch
1065 sub dispatch {
1066 if (!defined $action) {
1067 if (defined $hash) {
1068 $action = git_get_type($hash);
1069 } elsif (defined $hash_base && defined $file_name) {
1070 $action = git_get_type("$hash_base:$file_name");
1071 } elsif (defined $project) {
1072 $action = 'summary';
1073 } else {
1074 $action = 'project_list';
1075 }
1076 }
1077 if (!defined($actions{$action})) {
1078 die_error(400, "Unknown action");
1079 }
1080 if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1081 !$project) {
1082 die_error(400, "Project needed");
1083 }
1084 $actions{$action}->();
1085 }
1086
1087 sub reset_timer {
1088 our $t0 = [ gettimeofday() ]
1089 if defined $t0;
1090 our $number_of_git_cmds = 0;
1091 }
1092
1093 our $first_request = 1;
1094 sub run_request {
1095 reset_timer();
1096
1097 evaluate_uri();
1098 if ($first_request) {
1099 evaluate_gitweb_config();
1100 evaluate_git_version();
1101 }
1102 if ($per_request_config) {
1103 if (ref($per_request_config) eq 'CODE') {
1104 $per_request_config->();
1105 } elsif (!$first_request) {
1106 evaluate_gitweb_config();
1107 }
1108 }
1109 check_loadavg();
1110
1111 # $projectroot and $projects_list might be set in gitweb config file
1112 $projects_list ||= $projectroot;
1113
1114 evaluate_query_params();
1115 evaluate_path_info();
1116 evaluate_and_validate_params();
1117 evaluate_git_dir();
1118
1119 configure_gitweb_features();
1120
1121 dispatch();
1122 }
1123
1124 our $is_last_request = sub { 1 };
1125 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1126 our $CGI = 'CGI';
1127 our $cgi;
1128 sub configure_as_fcgi {
1129 require CGI::Fast;
1130 our $CGI = 'CGI::Fast';
1131
1132 my $request_number = 0;
1133 # let each child service 100 requests
1134 our $is_last_request = sub { ++$request_number > 100 };
1135 }
1136 sub evaluate_argv {
1137 my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1138 configure_as_fcgi()
1139 if $script_name =~ /\.fcgi$/;
1140
1141 return unless (@ARGV);
1142
1143 require Getopt::Long;
1144 Getopt::Long::GetOptions(
1145 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1146 'nproc|n=i' => sub {
1147 my ($arg, $val) = @_;
1148 return unless eval { require FCGI::ProcManager; 1; };
1149 my $proc_manager = FCGI::ProcManager->new({
1150 n_processes => $val,
1151 });
1152 our $pre_listen_hook = sub { $proc_manager->pm_manage() };
1153 our $pre_dispatch_hook = sub { $proc_manager->pm_pre_dispatch() };
1154 our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1155 },
1156 );
1157 }
1158
1159 sub run {
1160 evaluate_argv();
1161
1162 $first_request = 1;
1163 $pre_listen_hook->()
1164 if $pre_listen_hook;
1165
1166 REQUEST:
1167 while ($cgi = $CGI->new()) {
1168 $pre_dispatch_hook->()
1169 if $pre_dispatch_hook;
1170
1171 run_request();
1172
1173 $post_dispatch_hook->()
1174 if $post_dispatch_hook;
1175 $first_request = 0;
1176
1177 last REQUEST if ($is_last_request->());
1178 }
1179
1180 DONE_GITWEB:
1181 1;
1182 }
1183
1184 run();
1185
1186 if (defined caller) {
1187 # wrapped in a subroutine processing requests,
1188 # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1189 return;
1190 } else {
1191 # pure CGI script, serving single request
1192 exit;
1193 }
1194
1195 ## ======================================================================
1196 ## action links
1197
1198 # possible values of extra options
1199 # -full => 0|1 - use absolute/full URL ($my_uri/$my_url as base)
1200 # -replay => 1 - start from a current view (replay with modifications)
1201 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1202 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1203 sub href {
1204 my %params = @_;
1205 # default is to use -absolute url() i.e. $my_uri
1206 my $href = $params{-full} ? $my_url : $my_uri;
1207
1208 # implicit -replay, must be first of implicit params
1209 $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1210
1211 $params{'project'} = $project unless exists $params{'project'};
1212
1213 if ($params{-replay}) {
1214 while (my ($name, $symbol) = each %cgi_param_mapping) {
1215 if (!exists $params{$name}) {
1216 $params{$name} = $input_params{$name};
1217 }
1218 }
1219 }
1220
1221 my $use_pathinfo = gitweb_check_feature('pathinfo');
1222 if (defined $params{'project'} &&
1223 (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1224 # try to put as many parameters as possible in PATH_INFO:
1225 # - project name
1226 # - action
1227 # - hash_parent or hash_parent_base:/file_parent
1228 # - hash or hash_base:/filename
1229 # - the snapshot_format as an appropriate suffix
1230
1231 # When the script is the root DirectoryIndex for the domain,
1232 # $href here would be something like http://gitweb.example.com/
1233 # Thus, we strip any trailing / from $href, to spare us double
1234 # slashes in the final URL
1235 $href =~ s,/$,,;
1236
1237 # Then add the project name, if present
1238 $href .= "/".esc_path_info($params{'project'});
1239 delete $params{'project'};
1240
1241 # since we destructively absorb parameters, we keep this
1242 # boolean that remembers if we're handling a snapshot
1243 my $is_snapshot = $params{'action'} eq 'snapshot';
1244
1245 # Summary just uses the project path URL, any other action is
1246 # added to the URL
1247 if (defined $params{'action'}) {
1248 $href .= "/".esc_path_info($params{'action'})
1249 unless $params{'action'} eq 'summary';
1250 delete $params{'action'};
1251 }
1252
1253 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1254 # stripping nonexistent or useless pieces
1255 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1256 || $params{'hash_parent'} || $params{'hash'});
1257 if (defined $params{'hash_base'}) {
1258 if (defined $params{'hash_parent_base'}) {
1259 $href .= esc_path_info($params{'hash_parent_base'});
1260 # skip the file_parent if it's the same as the file_name
1261 if (defined $params{'file_parent'}) {
1262 if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1263 delete $params{'file_parent'};
1264 } elsif ($params{'file_parent'} !~ /\.\./) {
1265 $href .= ":/".esc_path_info($params{'file_parent'});
1266 delete $params{'file_parent'};
1267 }
1268 }
1269 $href .= "..";
1270 delete $params{'hash_parent'};
1271 delete $params{'hash_parent_base'};
1272 } elsif (defined $params{'hash_parent'}) {
1273 $href .= esc_path_info($params{'hash_parent'}). "..";
1274 delete $params{'hash_parent'};
1275 }
1276
1277 $href .= esc_path_info($params{'hash_base'});
1278 if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1279 $href .= ":/".esc_path_info($params{'file_name'});
1280 delete $params{'file_name'};
1281 }
1282 delete $params{'hash'};
1283 delete $params{'hash_base'};
1284 } elsif (defined $params{'hash'}) {
1285 $href .= esc_path_info($params{'hash'});
1286 delete $params{'hash'};
1287 }
1288
1289 # If the action was a snapshot, we can absorb the
1290 # snapshot_format parameter too
1291 if ($is_snapshot) {
1292 my $fmt = $params{'snapshot_format'};
1293 # snapshot_format should always be defined when href()
1294 # is called, but just in case some code forgets, we
1295 # fall back to the default
1296 $fmt ||= $snapshot_fmts[0];
1297 $href .= $known_snapshot_formats{$fmt}{'suffix'};
1298 delete $params{'snapshot_format'};
1299 }
1300 }
1301
1302 # now encode the parameters explicitly
1303 my @result = ();
1304 for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1305 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1306 if (defined $params{$name}) {
1307 if (ref($params{$name}) eq "ARRAY") {
1308 foreach my $par (@{$params{$name}}) {
1309 push @result, $symbol . "=" . esc_param($par);
1310 }
1311 } else {
1312 push @result, $symbol . "=" . esc_param($params{$name});
1313 }
1314 }
1315 }
1316 $href .= "?" . join(';', @result) if scalar @result;
1317
1318 # final transformation: trailing spaces must be escaped (URI-encoded)
1319 $href =~ s/(\s+)$/CGI::escape($1)/e;
1320
1321 if ($params{-anchor}) {
1322 $href .= "#".esc_param($params{-anchor});
1323 }
1324
1325 return $href;
1326 }
1327
1328
1329 ## ======================================================================
1330 ## validation, quoting/unquoting and escaping
1331
1332 sub validate_action {
1333 my $input = shift || return undef;
1334 return undef unless exists $actions{$input};
1335 return $input;
1336 }
1337
1338 sub validate_project {
1339 my $input = shift || return undef;
1340 if (!validate_pathname($input) ||
1341 !(-d "$projectroot/$input") ||
1342 !check_export_ok("$projectroot/$input") ||
1343 ($strict_export && !project_in_list($input))) {
1344 return undef;
1345 } else {
1346 return $input;
1347 }
1348 }
1349
1350 sub validate_pathname {
1351 my $input = shift || return undef;
1352
1353 # no '.' or '..' as elements of path, i.e. no '.' nor '..'
1354 # at the beginning, at the end, and between slashes.
1355 # also this catches doubled slashes
1356 if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1357 return undef;
1358 }
1359 # no null characters
1360 if ($input =~ m!\0!) {
1361 return undef;
1362 }
1363 return $input;
1364 }
1365
1366 sub validate_refname {
1367 my $input = shift || return undef;
1368
1369 # textual hashes are O.K.
1370 if ($input =~ m/^[0-9a-fA-F]{40}$/) {
1371 return $input;
1372 }
1373 # it must be correct pathname
1374 $input = validate_pathname($input)
1375 or return undef;
1376 # restrictions on ref name according to git-check-ref-format
1377 if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1378 return undef;
1379 }
1380 return $input;
1381 }
1382
1383 # decode sequences of octets in utf8 into Perl's internal form,
1384 # which is utf-8 with utf8 flag set if needed. gitweb writes out
1385 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1386 sub to_utf8 {
1387 my $str = shift;
1388 return undef unless defined $str;
1389 if (utf8::valid($str)) {
1390 utf8::decode($str);
1391 return $str;
1392 } else {
1393 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1394 }
1395 }
1396
1397 # quote unsafe chars, but keep the slash, even when it's not
1398 # correct, but quoted slashes look too horrible in bookmarks
1399 sub esc_param {
1400 my $str = shift;
1401 return undef unless defined $str;
1402 $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1403 $str =~ s/ /\+/g;
1404 return $str;
1405 }
1406
1407 # the quoting rules for path_info fragment are slightly different
1408 sub esc_path_info {
1409 my $str = shift;
1410 return undef unless defined $str;
1411
1412 # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1413 $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1414
1415 return $str;
1416 }
1417
1418 # quote unsafe chars in whole URL, so some characters cannot be quoted
1419 sub esc_url {
1420 my $str = shift;
1421 return undef unless defined $str;
1422 $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1423 $str =~ s/ /\+/g;
1424 return $str;
1425 }
1426
1427 # quote unsafe characters in HTML attributes
1428 sub esc_attr {
1429
1430 # for XHTML conformance escaping '"' to '&quot;' is not enough
1431 return esc_html(@_);
1432 }
1433
1434 # replace invalid utf8 character with SUBSTITUTION sequence
1435 sub esc_html {
1436 my $str = shift;
1437 my %opts = @_;
1438
1439 return undef unless defined $str;
1440
1441 $str = to_utf8($str);
1442 $str = $cgi->escapeHTML($str);
1443 if ($opts{'-nbsp'}) {
1444 $str =~ s/ /&nbsp;/g;
1445 }
1446 $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1447 return $str;
1448 }
1449
1450 # quote control characters and escape filename to HTML
1451 sub esc_path {
1452 my $str = shift;
1453 my %opts = @_;
1454
1455 return undef unless defined $str;
1456
1457 $str = to_utf8($str);
1458 $str = $cgi->escapeHTML($str);
1459 if ($opts{'-nbsp'}) {
1460 $str =~ s/ /&nbsp;/g;
1461 }
1462 $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1463 return $str;
1464 }
1465
1466 # Make control characters "printable", using character escape codes (CEC)
1467 sub quot_cec {
1468 my $cntrl = shift;
1469 my %opts = @_;
1470 my %es = ( # character escape codes, aka escape sequences
1471 "\t" => '\t', # tab (HT)
1472 "\n" => '\n', # line feed (LF)
1473 "\r" => '\r', # carrige return (CR)
1474 "\f" => '\f', # form feed (FF)
1475 "\b" => '\b', # backspace (BS)
1476 "\a" => '\a', # alarm (bell) (BEL)
1477 "\e" => '\e', # escape (ESC)
1478 "\013" => '\v', # vertical tab (VT)
1479 "\000" => '\0', # nul character (NUL)
1480 );
1481 my $chr = ( (exists $es{$cntrl})
1482 ? $es{$cntrl}
1483 : sprintf('\%2x', ord($cntrl)) );
1484 if ($opts{-nohtml}) {
1485 return $chr;
1486 } else {
1487 return "<span class=\"cntrl\">$chr</span>";
1488 }
1489 }
1490
1491 # Alternatively use unicode control pictures codepoints,
1492 # Unicode "printable representation" (PR)
1493 sub quot_upr {
1494 my $cntrl = shift;
1495 my %opts = @_;
1496
1497 my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1498 if ($opts{-nohtml}) {
1499 return $chr;
1500 } else {
1501 return "<span class=\"cntrl\">$chr</span>";
1502 }
1503 }
1504
1505 # git may return quoted and escaped filenames
1506 sub unquote {
1507 my $str = shift;
1508
1509 sub unq {
1510 my $seq = shift;
1511 my %es = ( # character escape codes, aka escape sequences
1512 't' => "\t", # tab (HT, TAB)
1513 'n' => "\n", # newline (NL)
1514 'r' => "\r", # return (CR)
1515 'f' => "\f", # form feed (FF)
1516 'b' => "\b", # backspace (BS)
1517 'a' => "\a", # alarm (bell) (BEL)
1518 'e' => "\e", # escape (ESC)
1519 'v' => "\013", # vertical tab (VT)
1520 );
1521
1522 if ($seq =~ m/^[0-7]{1,3}$/) {
1523 # octal char sequence
1524 return chr(oct($seq));
1525 } elsif (exists $es{$seq}) {
1526 # C escape sequence, aka character escape code
1527 return $es{$seq};
1528 }
1529 # quoted ordinary character
1530 return $seq;
1531 }
1532
1533 if ($str =~ m/^"(.*)"$/) {
1534 # needs unquoting
1535 $str = $1;
1536 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1537 }
1538 return $str;
1539 }
1540
1541 # escape tabs (convert tabs to spaces)
1542 sub untabify {
1543 my $line = shift;
1544
1545 while ((my $pos = index($line, "\t")) != -1) {
1546 if (my $count = (8 - ($pos % 8))) {
1547 my $spaces = ' ' x $count;
1548 $line =~ s/\t/$spaces/;
1549 }
1550 }
1551
1552 return $line;
1553 }
1554
1555 sub project_in_list {
1556 my $project = shift;
1557 my @list = git_get_projects_list();
1558 return @list && scalar(grep { $_->{'path'} eq $project } @list);
1559 }
1560
1561 ## ----------------------------------------------------------------------
1562 ## HTML aware string manipulation
1563
1564 # Try to chop given string on a word boundary between position
1565 # $len and $len+$add_len. If there is no word boundary there,
1566 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1567 # (marking chopped part) would be longer than given string.
1568 sub chop_str {
1569 my $str = shift;
1570 my $len = shift;
1571 my $add_len = shift || 10;
1572 my $where = shift || 'right'; # 'left' | 'center' | 'right'
1573
1574 # Make sure perl knows it is utf8 encoded so we don't
1575 # cut in the middle of a utf8 multibyte char.
1576 $str = to_utf8($str);
1577
1578 # allow only $len chars, but don't cut a word if it would fit in $add_len
1579 # if it doesn't fit, cut it if it's still longer than the dots we would add
1580 # remove chopped character entities entirely
1581
1582 # when chopping in the middle, distribute $len into left and right part
1583 # return early if chopping wouldn't make string shorter
1584 if ($where eq 'center') {
1585 return $str if ($len + 5 >= length($str)); # filler is length 5
1586 $len = int($len/2);
1587 } else {
1588 return $str if ($len + 4 >= length($str)); # filler is length 4
1589 }
1590
1591 # regexps: ending and beginning with word part up to $add_len
1592 my $endre = qr/.{$len}\w{0,$add_len}/;
1593 my $begre = qr/\w{0,$add_len}.{$len}/;
1594
1595 if ($where eq 'left') {
1596 $str =~ m/^(.*?)($begre)$/;
1597 my ($lead, $body) = ($1, $2);
1598 if (length($lead) > 4) {
1599 $lead = " ...";
1600 }
1601 return "$lead$body";
1602
1603 } elsif ($where eq 'center') {
1604 $str =~ m/^($endre)(.*)$/;
1605 my ($left, $str) = ($1, $2);
1606 $str =~ m/^(.*?)($begre)$/;
1607 my ($mid, $right) = ($1, $2);
1608 if (length($mid) > 5) {
1609 $mid = " ... ";
1610 }
1611 return "$left$mid$right";
1612
1613 } else {
1614 $str =~ m/^($endre)(.*)$/;
1615 my $body = $1;
1616 my $tail = $2;
1617 if (length($tail) > 4) {
1618 $tail = "... ";
1619 }
1620 return "$body$tail";
1621 }
1622 }
1623
1624 # takes the same arguments as chop_str, but also wraps a <span> around the
1625 # result with a title attribute if it does get chopped. Additionally, the
1626 # string is HTML-escaped.
1627 sub chop_and_escape_str {
1628 my ($str) = @_;
1629
1630 my $chopped = chop_str(@_);
1631 if ($chopped eq $str) {
1632 return esc_html($chopped);
1633 } else {
1634 $str =~ s/[[:cntrl:]]/?/g;
1635 return $cgi->span({-title=>$str}, esc_html($chopped));
1636 }
1637 }
1638
1639 ## ----------------------------------------------------------------------
1640 ## functions returning short strings
1641
1642 # CSS class for given age value (in seconds)
1643 sub age_class {
1644 my $age = shift;
1645
1646 if (!defined $age) {
1647 return "noage";
1648 } elsif ($age < 60*60*2) {
1649 return "age0";
1650 } elsif ($age < 60*60*24*2) {
1651 return "age1";
1652 } else {
1653 return "age2";
1654 }
1655 }
1656
1657 # convert age in seconds to "nn units ago" string
1658 sub age_string {
1659 my $age = shift;
1660 my $age_str;
1661
1662 if ($age > 60*60*24*365*2) {
1663 $age_str = (int $age/60/60/24/365);
1664 $age_str .= " years ago";
1665 } elsif ($age > 60*60*24*(365/12)*2) {
1666 $age_str = int $age/60/60/24/(365/12);
1667 $age_str .= " months ago";
1668 } elsif ($age > 60*60*24*7*2) {
1669 $age_str = int $age/60/60/24/7;
1670 $age_str .= " weeks ago";
1671 } elsif ($age > 60*60*24*2) {
1672 $age_str = int $age/60/60/24;
1673 $age_str .= " days ago";
1674 } elsif ($age > 60*60*2) {
1675 $age_str = int $age/60/60;
1676 $age_str .= " hours ago";
1677 } elsif ($age > 60*2) {
1678 $age_str = int $age/60;
1679 $age_str .= " min ago";
1680 } elsif ($age > 2) {
1681 $age_str = int $age;
1682 $age_str .= " sec ago";
1683 } else {
1684 $age_str .= " right now";
1685 }
1686 return $age_str;
1687 }
1688
1689 use constant {
1690 S_IFINVALID => 0030000,
1691 S_IFGITLINK => 0160000,
1692 };
1693
1694 # submodule/subproject, a commit object reference
1695 sub S_ISGITLINK {
1696 my $mode = shift;
1697
1698 return (($mode & S_IFMT) == S_IFGITLINK)
1699 }
1700
1701 # convert file mode in octal to symbolic file mode string
1702 sub mode_str {
1703 my $mode = oct shift;
1704
1705 if (S_ISGITLINK($mode)) {
1706 return 'm---------';
1707 } elsif (S_ISDIR($mode & S_IFMT)) {
1708 return 'drwxr-xr-x';
1709 } elsif (S_ISLNK($mode)) {
1710 return 'lrwxrwxrwx';
1711 } elsif (S_ISREG($mode)) {
1712 # git cares only about the executable bit
1713 if ($mode & S_IXUSR) {
1714 return '-rwxr-xr-x';
1715 } else {
1716 return '-rw-r--r--';
1717 };
1718 } else {
1719 return '----------';
1720 }
1721 }
1722
1723 # convert file mode in octal to file type string
1724 sub file_type {
1725 my $mode = shift;
1726
1727 if ($mode !~ m/^[0-7]+$/) {
1728 return $mode;
1729 } else {
1730 $mode = oct $mode;
1731 }
1732
1733 if (S_ISGITLINK($mode)) {
1734 return "submodule";
1735 } elsif (S_ISDIR($mode & S_IFMT)) {
1736 return "directory";
1737 } elsif (S_ISLNK($mode)) {
1738 return "symlink";
1739 } elsif (S_ISREG($mode)) {
1740 return "file";
1741 } else {
1742 return "unknown";
1743 }
1744 }
1745
1746 # convert file mode in octal to file type description string
1747 sub file_type_long {
1748 my $mode = shift;
1749
1750 if ($mode !~ m/^[0-7]+$/) {
1751 return $mode;
1752 } else {
1753 $mode = oct $mode;
1754 }
1755
1756 if (S_ISGITLINK($mode)) {
1757 return "submodule";
1758 } elsif (S_ISDIR($mode & S_IFMT)) {
1759 return "directory";
1760 } elsif (S_ISLNK($mode)) {
1761 return "symlink";
1762 } elsif (S_ISREG($mode)) {
1763 if ($mode & S_IXUSR) {
1764 return "executable";
1765 } else {
1766 return "file";
1767 };
1768 } else {
1769 return "unknown";
1770 }
1771 }
1772
1773
1774 ## ----------------------------------------------------------------------
1775 ## functions returning short HTML fragments, or transforming HTML fragments
1776 ## which don't belong to other sections
1777
1778 # format line of commit message.
1779 sub format_log_line_html {
1780 my $line = shift;
1781
1782 $line = esc_html($line, -nbsp=>1);
1783 $line =~ s{\b([0-9a-fA-F]{8,40})\b}{
1784 $cgi->a({-href => href(action=>"object", hash=>$1),
1785 -class => "text"}, $1);
1786 }eg;
1787
1788 return $line;
1789 }
1790
1791 # format marker of refs pointing to given object
1792
1793 # the destination action is chosen based on object type and current context:
1794 # - for annotated tags, we choose the tag view unless it's the current view
1795 # already, in which case we go to shortlog view
1796 # - for other refs, we keep the current view if we're in history, shortlog or
1797 # log view, and select shortlog otherwise
1798 sub format_ref_marker {
1799 my ($refs, $id) = @_;
1800 my $markers = '';
1801
1802 if (defined $refs->{$id}) {
1803 foreach my $ref (@{$refs->{$id}}) {
1804 # this code exploits the fact that non-lightweight tags are the
1805 # only indirect objects, and that they are the only objects for which
1806 # we want to use tag instead of shortlog as action
1807 my ($type, $name) = qw();
1808 my $indirect = ($ref =~ s/\^\{\}$//);
1809 # e.g. tags/v2.6.11 or heads/next
1810 if ($ref =~ m!^(.*?)s?/(.*)$!) {
1811 $type = $1;
1812 $name = $2;
1813 } else {
1814 $type = "ref";
1815 $name = $ref;
1816 }
1817
1818 my $class = $type;
1819 $class .= " indirect" if $indirect;
1820
1821 my $dest_action = "shortlog";
1822
1823 if ($indirect) {
1824 $dest_action = "tag" unless $action eq "tag";
1825 } elsif ($action =~ /^(history|(short)?log)$/) {
1826 $dest_action = $action;
1827 }
1828
1829 my $dest = "";
1830 $dest .= "refs/" unless $ref =~ m!^refs/!;
1831 $dest .= $ref;
1832
1833 my $link = $cgi->a({
1834 -href => href(
1835 action=>$dest_action,
1836 hash=>$dest
1837 )}, $name);
1838
1839 $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
1840 $link . "</span>";
1841 }
1842 }
1843
1844 if ($markers) {
1845 return ' <span class="refs">'. $markers . '</span>';
1846 } else {
1847 return "";
1848 }
1849 }
1850
1851 # format, perhaps shortened and with markers, title line
1852 sub format_subject_html {
1853 my ($long, $short, $href, $extra) = @_;
1854 $extra = '' unless defined($extra);
1855
1856 if (length($short) < length($long)) {
1857 $long =~ s/[[:cntrl:]]/?/g;
1858 return $cgi->a({-href => $href, -class => "list subject",
1859 -title => to_utf8($long)},
1860 esc_html($short)) . $extra;
1861 } else {
1862 return $cgi->a({-href => $href, -class => "list subject"},
1863 esc_html($long)) . $extra;
1864 }
1865 }
1866
1867 # Rather than recomputing the url for an email multiple times, we cache it
1868 # after the first hit. This gives a visible benefit in views where the avatar
1869 # for the same email is used repeatedly (e.g. shortlog).
1870 # The cache is shared by all avatar engines (currently gravatar only), which
1871 # are free to use it as preferred. Since only one avatar engine is used for any
1872 # given page, there's no risk for cache conflicts.
1873 our %avatar_cache = ();
1874
1875 # Compute the picon url for a given email, by using the picon search service over at
1876 # http://www.cs.indiana.edu/picons/search.html
1877 sub picon_url {
1878 my $email = lc shift;
1879 if (!$avatar_cache{$email}) {
1880 my ($user, $domain) = split('@', $email);
1881 $avatar_cache{$email} =
1882 "http://www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
1883 "$domain/$user/" .
1884 "users+domains+unknown/up/single";
1885 }
1886 return $avatar_cache{$email};
1887 }
1888
1889 # Compute the gravatar url for a given email, if it's not in the cache already.
1890 # Gravatar stores only the part of the URL before the size, since that's the
1891 # one computationally more expensive. This also allows reuse of the cache for
1892 # different sizes (for this particular engine).
1893 sub gravatar_url {
1894 my $email = lc shift;
1895 my $size = shift;
1896 $avatar_cache{$email} ||=
1897 "http://www.gravatar.com/avatar/" .
1898 Digest::MD5::md5_hex($email) . "?s=";
1899 return $avatar_cache{$email} . $size;
1900 }
1901
1902 # Insert an avatar for the given $email at the given $size if the feature
1903 # is enabled.
1904 sub git_get_avatar {
1905 my ($email, %opts) = @_;
1906 my $pre_white = ($opts{-pad_before} ? "&nbsp;" : "");
1907 my $post_white = ($opts{-pad_after} ? "&nbsp;" : "");
1908 $opts{-size} ||= 'default';
1909 my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
1910 my $url = "";
1911 if ($git_avatar eq 'gravatar') {
1912 $url = gravatar_url($email, $size);
1913 } elsif ($git_avatar eq 'picon') {
1914 $url = picon_url($email);
1915 }
1916 # Other providers can be added by extending the if chain, defining $url
1917 # as needed. If no variant puts something in $url, we assume avatars
1918 # are completely disabled/unavailable.
1919 if ($url) {
1920 return $pre_white .
1921 "<img width=\"$size\" " .
1922 "class=\"avatar\" " .
1923 "src=\"".esc_url($url)."\" " .
1924 "alt=\"\" " .
1925 "/>" . $post_white;
1926 } else {
1927 return "";
1928 }
1929 }
1930
1931 sub format_search_author {
1932 my ($author, $searchtype, $displaytext) = @_;
1933 my $have_search = gitweb_check_feature('search');
1934
1935 if ($have_search) {
1936 my $performed = "";
1937 if ($searchtype eq 'author') {
1938 $performed = "authored";
1939 } elsif ($searchtype eq 'committer') {
1940 $performed = "committed";
1941 }
1942
1943 return $cgi->a({-href => href(action=>"search", hash=>$hash,
1944 searchtext=>$author,
1945 searchtype=>$searchtype), class=>"list",
1946 title=>"Search for commits $performed by $author"},
1947 $displaytext);
1948
1949 } else {
1950 return $displaytext;
1951 }
1952 }
1953
1954 # format the author name of the given commit with the given tag
1955 # the author name is chopped and escaped according to the other
1956 # optional parameters (see chop_str).
1957 sub format_author_html {
1958 my $tag = shift;
1959 my $co = shift;
1960 my $author = chop_and_escape_str($co->{'author_name'}, @_);
1961 return "<$tag class=\"author\">" .
1962 format_search_author($co->{'author_name'}, "author",
1963 git_get_avatar($co->{'author_email'}, -pad_after => 1) .
1964 $author) .
1965 "</$tag>";
1966 }
1967
1968 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
1969 sub format_git_diff_header_line {
1970 my $line = shift;
1971 my $diffinfo = shift;
1972 my ($from, $to) = @_;
1973
1974 if ($diffinfo->{'nparents'}) {
1975 # combined diff
1976 $line =~ s!^(diff (.*?) )"?.*$!$1!;
1977 if ($to->{'href'}) {
1978 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
1979 esc_path($to->{'file'}));
1980 } else { # file was deleted (no href)
1981 $line .= esc_path($to->{'file'});
1982 }
1983 } else {
1984 # "ordinary" diff
1985 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
1986 if ($from->{'href'}) {
1987 $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
1988 'a/' . esc_path($from->{'file'}));
1989 } else { # file was added (no href)
1990 $line .= 'a/' . esc_path($from->{'file'});
1991 }
1992 $line .= ' ';
1993 if ($to->{'href'}) {
1994 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
1995 'b/' . esc_path($to->{'file'}));
1996 } else { # file was deleted
1997 $line .= 'b/' . esc_path($to->{'file'});
1998 }
1999 }
2000
2001 return "<div class=\"diff header\">$line</div>\n";
2002 }
2003
2004 # format extended diff header line, before patch itself
2005 sub format_extended_diff_header_line {
2006 my $line = shift;
2007 my $diffinfo = shift;
2008 my ($from, $to) = @_;
2009
2010 # match <path>
2011 if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2012 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2013 esc_path($from->{'file'}));
2014 }
2015 if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2016 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2017 esc_path($to->{'file'}));
2018 }
2019 # match single <mode>
2020 if ($line =~ m/\s(\d{6})$/) {
2021 $line .= '<span class="info"> (' .
2022 file_type_long($1) .
2023 ')</span>';
2024 }
2025 # match <hash>
2026 if ($line =~ m/^index [0-9a-fA-F]{40},[0-9a-fA-F]{40}/) {
2027 # can match only for combined diff
2028 $line = 'index ';
2029 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2030 if ($from->{'href'}[$i]) {
2031 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2032 -class=>"hash"},
2033 substr($diffinfo->{'from_id'}[$i],0,7));
2034 } else {
2035 $line .= '0' x 7;
2036 }
2037 # separator
2038 $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2039 }
2040 $line .= '..';
2041 if ($to->{'href'}) {
2042 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2043 substr($diffinfo->{'to_id'},0,7));
2044 } else {
2045 $line .= '0' x 7;
2046 }
2047
2048 } elsif ($line =~ m/^index [0-9a-fA-F]{40}..[0-9a-fA-F]{40}/) {
2049 # can match only for ordinary diff
2050 my ($from_link, $to_link);
2051 if ($from->{'href'}) {
2052 $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2053 substr($diffinfo->{'from_id'},0,7));
2054 } else {
2055 $from_link = '0' x 7;
2056 }
2057 if ($to->{'href'}) {
2058 $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2059 substr($diffinfo->{'to_id'},0,7));
2060 } else {
2061 $to_link = '0' x 7;
2062 }
2063 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2064 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2065 }
2066
2067 return $line . "<br/>\n";
2068 }
2069
2070 # format from-file/to-file diff header
2071 sub format_diff_from_to_header {
2072 my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2073 my $line;
2074 my $result = '';
2075
2076 $line = $from_line;
2077 #assert($line =~ m/^---/) if DEBUG;
2078 # no extra formatting for "^--- /dev/null"
2079 if (! $diffinfo->{'nparents'}) {
2080 # ordinary (single parent) diff
2081 if ($line =~ m!^--- "?a/!) {
2082 if ($from->{'href'}) {
2083 $line = '--- a/' .
2084 $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2085 esc_path($from->{'file'}));
2086 } else {
2087 $line = '--- a/' .
2088 esc_path($from->{'file'});
2089 }
2090 }
2091 $result .= qq!<div class="diff from_file">$line</div>\n!;
2092
2093 } else {
2094 # combined diff (merge commit)
2095 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2096 if ($from->{'href'}[$i]) {
2097 $line = '--- ' .
2098 $cgi->a({-href=>href(action=>"blobdiff",
2099 hash_parent=>$diffinfo->{'from_id'}[$i],
2100 hash_parent_base=>$parents[$i],
2101 file_parent=>$from->{'file'}[$i],
2102 hash=>$diffinfo->{'to_id'},
2103 hash_base=>$hash,
2104 file_name=>$to->{'file'}),
2105 -class=>"path",
2106 -title=>"diff" . ($i+1)},
2107 $i+1) .
2108 '/' .
2109 $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2110 esc_path($from->{'file'}[$i]));
2111 } else {
2112 $line = '--- /dev/null';
2113 }
2114 $result .= qq!<div class="diff from_file">$line</div>\n!;
2115 }
2116 }
2117
2118 $line = $to_line;
2119 #assert($line =~ m/^\+\+\+/) if DEBUG;
2120 # no extra formatting for "^+++ /dev/null"
2121 if ($line =~ m!^\+\+\+ "?b/!) {
2122 if ($to->{'href'}) {
2123 $line = '+++ b/' .
2124 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2125 esc_path($to->{'file'}));
2126 } else {
2127 $line = '+++ b/' .
2128 esc_path($to->{'file'});
2129 }
2130 }
2131 $result .= qq!<div class="diff to_file">$line</div>\n!;
2132
2133 return $result;
2134 }
2135
2136 # create note for patch simplified by combined diff
2137 sub format_diff_cc_simplified {
2138 my ($diffinfo, @parents) = @_;
2139 my $result = '';
2140
2141 $result .= "<div class=\"diff header\">" .
2142 "diff --cc ";
2143 if (!is_deleted($diffinfo)) {
2144 $result .= $cgi->a({-href => href(action=>"blob",
2145 hash_base=>$hash,
2146 hash=>$diffinfo->{'to_id'},
2147 file_name=>$diffinfo->{'to_file'}),
2148 -class => "path"},
2149 esc_path($diffinfo->{'to_file'}));
2150 } else {
2151 $result .= esc_path($diffinfo->{'to_file'});
2152 }
2153 $result .= "</div>\n" . # class="diff header"
2154 "<div class=\"diff nodifferences\">" .
2155 "Simple merge" .
2156 "</div>\n"; # class="diff nodifferences"
2157
2158 return $result;
2159 }
2160
2161 # format patch (diff) line (not to be used for diff headers)
2162 sub format_diff_line {
2163 my $line = shift;
2164 my ($from, $to) = @_;
2165 my $diff_class = "";
2166
2167 chomp $line;
2168
2169 if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2170 # combined diff
2171 my $prefix = substr($line, 0, scalar @{$from->{'href'}});
2172 if ($line =~ m/^\@{3}/) {
2173 $diff_class = " chunk_header";
2174 } elsif ($line =~ m/^\\/) {
2175 $diff_class = " incomplete";
2176 } elsif ($prefix =~ tr/+/+/) {
2177 $diff_class = " add";
2178 } elsif ($prefix =~ tr/-/-/) {
2179 $diff_class = " rem";
2180 }
2181 } else {
2182 # assume ordinary diff
2183 my $char = substr($line, 0, 1);
2184 if ($char eq '+') {
2185 $diff_class = " add";
2186 } elsif ($char eq '-') {
2187 $diff_class = " rem";
2188 } elsif ($char eq '@') {
2189 $diff_class = " chunk_header";
2190 } elsif ($char eq "\\") {
2191 $diff_class = " incomplete";
2192 }
2193 }
2194 $line = untabify($line);
2195 if ($from && $to && $line =~ m/^\@{2} /) {
2196 my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2197 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2198
2199 $from_lines = 0 unless defined $from_lines;
2200 $to_lines = 0 unless defined $to_lines;
2201
2202 if ($from->{'href'}) {
2203 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2204 -class=>"list"}, $from_text);
2205 }
2206 if ($to->{'href'}) {
2207 $to_text = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2208 -class=>"list"}, $to_text);
2209 }
2210 $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2211 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2212 return "<div class=\"diff$diff_class\">$line</div>\n";
2213 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2214 my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2215 my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2216
2217 @from_text = split(' ', $ranges);
2218 for (my $i = 0; $i < @from_text; ++$i) {
2219 ($from_start[$i], $from_nlines[$i]) =
2220 (split(',', substr($from_text[$i], 1)), 0);
2221 }
2222
2223 $to_text = pop @from_text;
2224 $to_start = pop @from_start;
2225 $to_nlines = pop @from_nlines;
2226
2227 $line = "<span class=\"chunk_info\">$prefix ";
2228 for (my $i = 0; $i < @from_text; ++$i) {
2229 if ($from->{'href'}[$i]) {
2230 $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2231 -class=>"list"}, $from_text[$i]);
2232 } else {
2233 $line .= $from_text[$i];
2234 }
2235 $line .= " ";
2236 }
2237 if ($to->{'href'}) {
2238 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2239 -class=>"list"}, $to_text);
2240 } else {
2241 $line .= $to_text;
2242 }
2243 $line .= " $prefix</span>" .
2244 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2245 return "<div class=\"diff$diff_class\">$line</div>\n";
2246 }
2247 return "<div class=\"diff$diff_class\">" . esc_html($line, -nbsp=>1) . "</div>\n";
2248 }
2249
2250 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2251 # linked. Pass the hash of the tree/commit to snapshot.
2252 sub format_snapshot_links {
2253 my ($hash) = @_;
2254 my $num_fmts = @snapshot_fmts;
2255 if ($num_fmts > 1) {
2256 # A parenthesized list of links bearing format names.
2257 # e.g. "snapshot (_tar.gz_ _zip_)"
2258 return "snapshot (" . join(' ', map
2259 $cgi->a({
2260 -href => href(
2261 action=>"snapshot",
2262 hash=>$hash,
2263 snapshot_format=>$_
2264 )
2265 }, $known_snapshot_formats{$_}{'display'})
2266 , @snapshot_fmts) . ")";
2267 } elsif ($num_fmts == 1) {
2268 # A single "snapshot" link whose tooltip bears the format name.
2269 # i.e. "_snapshot_"
2270 my ($fmt) = @snapshot_fmts;
2271 return
2272 $cgi->a({
2273 -href => href(
2274 action=>"snapshot",
2275 hash=>$hash,
2276 snapshot_format=>$fmt
2277 ),
2278 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2279 }, "snapshot");
2280 } else { # $num_fmts == 0
2281 return undef;
2282 }
2283 }
2284
2285 ## ......................................................................
2286 ## functions returning values to be passed, perhaps after some
2287 ## transformation, to other functions; e.g. returning arguments to href()
2288
2289 # returns hash to be passed to href to generate gitweb URL
2290 # in -title key it returns description of link
2291 sub get_feed_info {
2292 my $format = shift || 'Atom';
2293 my %res = (action => lc($format));
2294
2295 # feed links are possible only for project views
2296 return unless (defined $project);
2297 # some views should link to OPML, or to generic project feed,
2298 # or don't have specific feed yet (so they should use generic)
2299 return if ($action =~ /^(?:tags|heads|forks|tag|search)$/x);
2300
2301 my $branch;
2302 # branches refs uses 'refs/heads/' prefix (fullname) to differentiate
2303 # from tag links; this also makes possible to detect branch links
2304 if ((defined $hash_base && $hash_base =~ m!^refs/heads/(.*)$!) ||
2305 (defined $hash && $hash =~ m!^refs/heads/(.*)$!)) {
2306 $branch = $1;
2307 }
2308 # find log type for feed description (title)
2309 my $type = 'log';
2310 if (defined $file_name) {
2311 $type = "history of $file_name";
2312 $type .= "/" if ($action eq 'tree');
2313 $type .= " on '$branch'" if (defined $branch);
2314 } else {
2315 $type = "log of $branch" if (defined $branch);
2316 }
2317
2318 $res{-title} = $type;
2319 $res{'hash'} = (defined $branch ? "refs/heads/$branch" : undef);
2320 $res{'file_name'} = $file_name;
2321
2322 return %res;
2323 }
2324
2325 ## ----------------------------------------------------------------------
2326 ## git utility subroutines, invoking git commands
2327
2328 # returns path to the core git executable and the --git-dir parameter as list
2329 sub git_cmd {
2330 $number_of_git_cmds++;
2331 return $GIT, '--git-dir='.$git_dir;
2332 }
2333
2334 # quote the given arguments for passing them to the shell
2335 # quote_command("command", "arg 1", "arg with ' and ! characters")
2336 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2337 # Try to avoid using this function wherever possible.
2338 sub quote_command {
2339 return join(' ',
2340 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2341 }
2342
2343 # get HEAD ref of given project as hash
2344 sub git_get_head_hash {
2345 return git_get_full_hash(shift, 'HEAD');
2346 }
2347
2348 sub git_get_full_hash {
2349 return git_get_hash(@_);
2350 }
2351
2352 sub git_get_short_hash {
2353 return git_get_hash(@_, '--short=7');
2354 }
2355
2356 sub git_get_hash {
2357 my ($project, $hash, @options) = @_;
2358 my $o_git_dir = $git_dir;
2359 my $retval = undef;
2360 $git_dir = "$projectroot/$project";
2361 if (open my $fd, '-|', git_cmd(), 'rev-parse',
2362 '--verify', '-q', @options, $hash) {
2363 $retval = <$fd>;
2364 chomp $retval if defined $retval;
2365 close $fd;
2366 }
2367 if (defined $o_git_dir) {
2368 $git_dir = $o_git_dir;
2369 }
2370 return $retval;
2371 }
2372
2373 # get type of given object
2374 sub git_get_type {
2375 my $hash = shift;
2376
2377 open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2378 my $type = <$fd>;
2379 close $fd or return;
2380 chomp $type;
2381 return $type;
2382 }
2383
2384 # repository configuration
2385 our $config_file = '';
2386 our %config;
2387
2388 # store multiple values for single key as anonymous array reference
2389 # single values stored directly in the hash, not as [ <value> ]
2390 sub hash_set_multi {
2391 my ($hash, $key, $value) = @_;
2392
2393 if (!exists $hash->{$key}) {
2394 $hash->{$key} = $value;
2395 } elsif (!ref $hash->{$key}) {
2396 $hash->{$key} = [ $hash->{$key}, $value ];
2397 } else {
2398 push @{$hash->{$key}}, $value;
2399 }
2400 }
2401
2402 # return hash of git project configuration
2403 # optionally limited to some section, e.g. 'gitweb'
2404 sub git_parse_project_config {
2405 my $section_regexp = shift;
2406 my %config;
2407
2408 local $/ = "\0";
2409
2410 open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2411 or return;
2412
2413 while (my $keyval = <$fh>) {
2414 chomp $keyval;
2415 my ($key, $value) = split(/\n/, $keyval, 2);
2416
2417 hash_set_multi(\%config, $key, $value)
2418 if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2419 }
2420 close $fh;
2421
2422 return %config;
2423 }
2424
2425 # convert config value to boolean: 'true' or 'false'
2426 # no value, number > 0, 'true' and 'yes' values are true
2427 # rest of values are treated as false (never as error)
2428 sub config_to_bool {
2429 my $val = shift;
2430
2431 return 1 if !defined $val; # section.key
2432
2433 # strip leading and trailing whitespace
2434 $val =~ s/^\s+//;
2435 $val =~ s/\s+$//;
2436
2437 return (($val =~ /^\d+$/ && $val) || # section.key = 1
2438 ($val =~ /^(?:true|yes)$/i)); # section.key = true
2439 }
2440
2441 # convert config value to simple decimal number
2442 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2443 # to be multiplied by 1024, 1048576, or 1073741824
2444 sub config_to_int {
2445 my $val = shift;
2446
2447 # strip leading and trailing whitespace
2448 $val =~ s/^\s+//;
2449 $val =~ s/\s+$//;
2450
2451 if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2452 $unit = lc($unit);
2453 # unknown unit is treated as 1
2454 return $num * ($unit eq 'g' ? 1073741824 :
2455 $unit eq 'm' ? 1048576 :
2456 $unit eq 'k' ? 1024 : 1);
2457 }
2458 return $val;
2459 }
2460
2461 # convert config value to array reference, if needed
2462 sub config_to_multi {
2463 my $val = shift;
2464
2465 return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2466 }
2467
2468 sub git_get_project_config {
2469 my ($key, $type) = @_;
2470
2471 return unless defined $git_dir;
2472
2473 # key sanity check
2474 return unless ($key);
2475 $key =~ s/^gitweb\.//;
2476 return if ($key =~ m/\W/);
2477
2478 # type sanity check
2479 if (defined $type) {
2480 $type =~ s/^--//;
2481 $type = undef
2482 unless ($type eq 'bool' || $type eq 'int');
2483 }
2484
2485 # get config
2486 if (!defined $config_file ||
2487 $config_file ne "$git_dir/config") {
2488 %config = git_parse_project_config('gitweb');
2489 $config_file = "$git_dir/config";
2490 }
2491
2492 # check if config variable (key) exists
2493 return unless exists $config{"gitweb.$key"};
2494
2495 # ensure given type
2496 if (!defined $type) {
2497 return $config{"gitweb.$key"};
2498 } elsif ($type eq 'bool') {
2499 # backward compatibility: 'git config --bool' returns true/false
2500 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2501 } elsif ($type eq 'int') {
2502 return config_to_int($config{"gitweb.$key"});
2503 }
2504 return $config{"gitweb.$key"};
2505 }
2506
2507 # get hash of given path at given ref
2508 sub git_get_hash_by_path {
2509 my $base = shift;
2510 my $path = shift || return undef;
2511 my $type = shift;
2512
2513 $path =~ s,/+$,,;
2514
2515 open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2516 or die_error(500, "Open git-ls-tree failed");
2517 my $line = <$fd>;
2518 close $fd or return undef;
2519
2520 if (!defined $line) {
2521 # there is no tree or hash given by $path at $base
2522 return undef;
2523 }
2524
2525 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
2526 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t/;
2527 if (defined $type && $type ne $2) {
2528 # type doesn't match
2529 return undef;
2530 }
2531 return $3;
2532 }
2533
2534 # get path of entry with given hash at given tree-ish (ref)
2535 # used to get 'from' filename for combined diff (merge commit) for renames
2536 sub git_get_path_by_hash {
2537 my $base = shift || return;
2538 my $hash = shift || return;
2539
2540 local $/ = "\0";
2541
2542 open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2543 or return undef;
2544 while (my $line = <$fd>) {
2545 chomp $line;
2546
2547 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423 gitweb'
2548 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f gitweb/README'
2549 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2550 close $fd;
2551 return $1;
2552 }
2553 }
2554 close $fd;
2555 return undef;
2556 }
2557
2558 ## ......................................................................
2559 ## git utility functions, directly accessing git repository
2560
2561 sub git_get_project_description {
2562 my $path = shift;
2563
2564 $git_dir = "$projectroot/$path";
2565 open my $fd, '<', "$git_dir/description"
2566 or return git_get_project_config('description');
2567 my $descr = <$fd>;
2568 close $fd;
2569 if (defined $descr) {
2570 chomp $descr;
2571 }
2572 return $descr;
2573 }
2574
2575 sub git_get_project_ctags {
2576 my $path = shift;
2577 my $ctags = {};
2578
2579 $git_dir = "$projectroot/$path";
2580 opendir my $dh, "$git_dir/ctags"
2581 or return $ctags;
2582 foreach (grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh)) {
2583 open my $ct, '<', $_ or next;
2584 my $val = <$ct>;
2585 chomp $val;
2586 close $ct;
2587 my $ctag = $_; $ctag =~ s#.*/##;
2588 $ctags->{$ctag} = $val;
2589 }
2590 closedir $dh;
2591 $ctags;
2592 }
2593
2594 sub git_populate_project_tagcloud {
2595 my $ctags = shift;
2596
2597 # First, merge different-cased tags; tags vote on casing
2598 my %ctags_lc;
2599 foreach (keys %$ctags) {
2600 $ctags_lc{lc $_}->{count} += $ctags->{$_};
2601 if (not $ctags_lc{lc $_}->{topcount}
2602 or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
2603 $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
2604 $ctags_lc{lc $_}->{topname} = $_;
2605 }
2606 }
2607
2608 my $cloud;
2609 if (eval { require HTML::TagCloud; 1; }) {
2610 $cloud = HTML::TagCloud->new;
2611 foreach (sort keys %ctags_lc) {
2612 # Pad the title with spaces so that the cloud looks
2613 # less crammed.
2614 my $title = $ctags_lc{$_}->{topname};
2615 $title =~ s/ /&nbsp;/g;
2616 $title =~ s/^/&nbsp;/g;
2617 $title =~ s/$/&nbsp;/g;
2618 $cloud->add($title, $home_link."?by_tag=".$_, $ctags_lc{$_}->{count});
2619 }
2620 } else {
2621 $cloud = \%ctags_lc;
2622 }
2623 $cloud;
2624 }
2625
2626 sub git_show_project_tagcloud {
2627 my ($cloud, $count) = @_;
2628 print STDERR ref($cloud)."..\n";
2629 if (ref $cloud eq 'HTML::TagCloud') {
2630 return $cloud->html_and_css($count);
2631 } else {
2632 my @tags = sort { $cloud->{$a}->{count} <=> $cloud->{$b}->{count} } keys %$cloud;
2633 return '<p align="center">' . join (', ', map {
2634 $cgi->a({-href=>"$home_link?by_tag=$_"}, $cloud->{$_}->{topname})
2635 } splice(@tags, 0, $count)) . '</p>';
2636 }
2637 }
2638
2639 sub git_get_project_url_list {
2640 my $path = shift;
2641
2642 $git_dir = "$projectroot/$path";
2643 open my $fd, '<', "$git_dir/cloneurl"
2644 or return wantarray ?
2645 @{ config_to_multi(git_get_project_config('url')) } :
2646 config_to_multi(git_get_project_config('url'));
2647 my @git_project_url_list = map { chomp; $_ } <$fd>;
2648 close $fd;
2649
2650 return wantarray ? @git_project_url_list : \@git_project_url_list;
2651 }
2652
2653 sub git_get_projects_list {
2654 my ($filter) = @_;
2655 my @list;
2656
2657 $filter ||= '';
2658 $filter =~ s/\.git$//;
2659
2660 my $check_forks = gitweb_check_feature('forks');
2661
2662 if (-d $projects_list) {
2663 # search in directory
2664 my $dir = $projects_list . ($filter ? "/$filter" : '');
2665 # remove the trailing "/"
2666 $dir =~ s!/+$!!;
2667 my $pfxlen = length("$dir");
2668 my $pfxdepth = ($dir =~ tr!/!!);
2669
2670 File::Find::find({
2671 follow_fast => 1, # follow symbolic links
2672 follow_skip => 2, # ignore duplicates
2673 dangling_symlinks => 0, # ignore dangling symlinks, silently
2674 wanted => sub {
2675 # global variables
2676 our $project_maxdepth;
2677 our $projectroot;
2678 # skip project-list toplevel, if we get it.
2679 return if (m!^[/.]$!);
2680 # only directories can be git repositories
2681 return unless (-d $_);
2682 # don't traverse too deep (Find is super slow on os x)
2683 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
2684 $File::Find::prune = 1;
2685 return;
2686 }
2687
2688 my $subdir = substr($File::Find::name, $pfxlen + 1);
2689 # we check related file in $projectroot
2690 my $path = ($filter ? "$filter/" : '') . $subdir;
2691 if (check_export_ok("$projectroot/$path")) {
2692 push @list, { path => $path };
2693 $File::Find::prune = 1;
2694 }
2695 },
2696 }, "$dir");
2697
2698 } elsif (-f $projects_list) {
2699 # read from file(url-encoded):
2700 # 'git%2Fgit.git Linus+Torvalds'
2701 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
2702 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
2703 my %paths;
2704 open my $fd, '<', $projects_list or return;
2705 PROJECT:
2706 while (my $line = <$fd>) {
2707 chomp $line;
2708 my ($path, $owner) = split ' ', $line;
2709 $path = unescape($path);
2710 $owner = unescape($owner);
2711 if (!defined $path) {
2712 next;
2713 }
2714 if ($filter ne '') {
2715 # looking for forks;
2716 my $pfx = substr($path, 0, length($filter));
2717 if ($pfx ne $filter) {
2718 next PROJECT;
2719 }
2720 my $sfx = substr($path, length($filter));
2721 if ($sfx !~ /^\/.*\.git$/) {
2722 next PROJECT;
2723 }
2724 } elsif ($check_forks) {
2725 PATH:
2726 foreach my $filter (keys %paths) {
2727 # looking for forks;
2728 my $pfx = substr($path, 0, length($filter));
2729 if ($pfx ne $filter) {
2730 next PATH;
2731 }
2732 my $sfx = substr($path, length($filter));
2733 if ($sfx !~ /^\/.*\.git$/) {
2734 next PATH;
2735 }
2736 # is a fork, don't include it in
2737 # the list
2738 next PROJECT;
2739 }
2740 }
2741 if (check_export_ok("$projectroot/$path")) {
2742 my $pr = {
2743 path => $path,
2744 owner => to_utf8($owner),
2745 };
2746 push @list, $pr;
2747 (my $forks_path = $path) =~ s/\.git$//;
2748 $paths{$forks_path}++;
2749 }
2750 }
2751 close $fd;
2752 }
2753 return @list;
2754 }
2755
2756 our $gitweb_project_owner = undef;
2757 sub git_get_project_list_from_file {
2758
2759 return if (defined $gitweb_project_owner);
2760
2761 $gitweb_project_owner = {};
2762 # read from file (url-encoded):
2763 # 'git%2Fgit.git Linus+Torvalds'
2764 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
2765 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
2766 if (-f $projects_list) {
2767 open(my $fd, '<', $projects_list);
2768 while (my $line = <$fd>) {
2769 chomp $line;
2770 my ($pr, $ow) = split ' ', $line;
2771 $pr = unescape($pr);
2772 $ow = unescape($ow);
2773 $gitweb_project_owner->{$pr} = to_utf8($ow);
2774 }
2775 close $fd;
2776 }
2777 }
2778
2779 sub git_get_project_owner {
2780 my $project = shift;
2781 my $owner;
2782
2783 return undef unless $project;
2784 $git_dir = "$projectroot/$project";
2785
2786 if (!defined $gitweb_project_owner) {
2787 git_get_project_list_from_file();
2788 }
2789
2790 if (exists $gitweb_project_owner->{$project}) {
2791 $owner = $gitweb_project_owner->{$project};
2792 }
2793 if (!defined $owner){
2794 $owner = git_get_project_config('owner');
2795 }
2796 if (!defined $owner) {
2797 $owner = get_file_owner("$git_dir");
2798 }
2799
2800 return $owner;
2801 }
2802
2803 sub git_get_last_activity {
2804 my ($path) = @_;
2805 my $fd;
2806
2807 $git_dir = "$projectroot/$path";
2808 open($fd, "-|", git_cmd(), 'for-each-ref',
2809 '--format=%(committer)',
2810 '--sort=-committerdate',
2811 '--count=1',
2812 'refs/heads') or return;
2813 my $most_recent = <$fd>;
2814 close $fd or return;
2815 if (defined $most_recent &&
2816 $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
2817 my $timestamp = $1;
2818 my $age = time - $timestamp;
2819 return ($age, age_string($age));
2820 }
2821 return (undef, undef);
2822 }
2823
2824 # Implementation note: when a single remote is wanted, we cannot use 'git
2825 # remote show -n' because that command always work (assuming it's a remote URL
2826 # if it's not defined), and we cannot use 'git remote show' because that would
2827 # try to make a network roundtrip. So the only way to find if that particular
2828 # remote is defined is to walk the list provided by 'git remote -v' and stop if
2829 # and when we find what we want.
2830 sub git_get_remotes_list {
2831 my $wanted = shift;
2832 my %remotes = ();
2833
2834 open my $fd, '-|' , git_cmd(), 'remote', '-v';
2835 return unless $fd;
2836 while (my $remote = <$fd>) {
2837 chomp $remote;
2838 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
2839 next if $wanted and not $remote eq $wanted;
2840 my ($url, $key) = ($1, $2);
2841
2842 $remotes{$remote} ||= { 'heads' => () };
2843 $remotes{$remote}{$key} = $url;
2844 }
2845 close $fd or return;
2846 return wantarray ? %remotes : \%remotes;
2847 }
2848
2849 # Takes a hash of remotes as first parameter and fills it by adding the
2850 # available remote heads for each of the indicated remotes.
2851 sub fill_remote_heads {
2852 my $remotes = shift;
2853 my @heads = map { "remotes/$_" } keys %$remotes;
2854 my @remoteheads = git_get_heads_list(undef, @heads);
2855 foreach my $remote (keys %$remotes) {
2856 $remotes->{$remote}{'heads'} = [ grep {
2857 $_->{'name'} =~ s!^$remote/!!
2858 } @remoteheads ];
2859 }
2860 }
2861
2862 sub git_get_references {
2863 my $type = shift || "";
2864 my %refs;
2865 # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
2866 # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
2867 open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
2868 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
2869 or return;
2870
2871 while (my $line = <$fd>) {
2872 chomp $line;
2873 if ($line =~ m!^([0-9a-fA-F]{40})\srefs/($type.*)$!) {
2874 if (defined $refs{$1}) {
2875 push @{$refs{$1}}, $2;
2876 } else {
2877 $refs{$1} = [ $2 ];
2878 }
2879 }
2880 }
2881 close $fd or return;
2882 return \%refs;
2883 }
2884
2885 sub git_get_rev_name_tags {
2886 my $hash = shift || return undef;
2887
2888 open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
2889 or return;
2890 my $name_rev = <$fd>;
2891 close $fd;
2892
2893 if ($name_rev =~ m|^$hash tags/(.*)$|) {
2894 return $1;
2895 } else {
2896 # catches also '$hash undefined' output
2897 return undef;
2898 }
2899 }
2900
2901 ## ----------------------------------------------------------------------
2902 ## parse to hash functions
2903
2904 sub parse_date {
2905 my $epoch = shift;
2906 my $tz = shift || "-0000";
2907
2908 my %date;
2909 my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
2910 my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
2911 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
2912 $date{'hour'} = $hour;
2913 $date{'minute'} = $min;
2914 $date{'mday'} = $mday;
2915 $date{'day'} = $days[$wday];
2916 $date{'month'} = $months[$mon];
2917 $date{'rfc2822'} = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
2918 $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
2919 $date{'mday-time'} = sprintf "%d %s %02d:%02d",
2920 $mday, $months[$mon], $hour ,$min;
2921 $date{'iso-8601'} = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
2922 1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
2923
2924 my ($tz_sign, $tz_hour, $tz_min) =
2925 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
2926 $tz_sign = ($tz_sign eq '-' ? -1 : +1);
2927 my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
2928 ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
2929 $date{'hour_local'} = $hour;
2930 $date{'minute_local'} = $min;
2931 $date{'tz_local'} = $tz;
2932 $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
2933 1900+$year, $mon+1, $mday,
2934 $hour, $min, $sec, $tz);
2935 return %date;
2936 }
2937
2938 sub parse_tag {
2939 my $tag_id = shift;
2940 my %tag;
2941 my @comment;
2942
2943 open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
2944 $tag{'id'} = $tag_id;
2945 while (my $line = <$fd>) {
2946 chomp $line;
2947 if ($line =~ m/^object ([0-9a-fA-F]{40})$/) {
2948 $tag{'object'} = $1;
2949 } elsif ($line =~ m/^type (.+)$/) {
2950 $tag{'type'} = $1;
2951 } elsif ($line =~ m/^tag (.+)$/) {
2952 $tag{'name'} = $1;
2953 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
2954 $tag{'author'} = $1;
2955 $tag{'author_epoch'} = $2;
2956 $tag{'author_tz'} = $3;
2957 if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
2958 $tag{'author_name'} = $1;
2959 $tag{'author_email'} = $2;
2960 } else {
2961 $tag{'author_name'} = $tag{'author'};
2962 }
2963 } elsif ($line =~ m/--BEGIN/) {
2964 push @comment, $line;
2965 last;
2966 } elsif ($line eq "") {
2967 last;
2968 }
2969 }
2970 push @comment, <$fd>;
2971 $tag{'comment'} = \@comment;
2972 close $fd or return;
2973 if (!defined $tag{'name'}) {
2974 return
2975 };
2976 return %tag
2977 }
2978
2979 sub parse_commit_text {
2980 my ($commit_text, $withparents) = @_;
2981 my @commit_lines = split '\n', $commit_text;
2982 my %co;
2983
2984 pop @commit_lines; # Remove '\0'
2985
2986 if (! @commit_lines) {
2987 return;
2988 }
2989
2990 my $header = shift @commit_lines;
2991 if ($header !~ m/^[0-9a-fA-F]{40}/) {
2992 return;
2993 }
2994 ($co{'id'}, my @parents) = split ' ', $header;
2995 while (my $line = shift @commit_lines) {
2996 last if $line eq "\n";
2997 if ($line =~ m/^tree ([0-9a-fA-F]{40})$/) {
2998 $co{'tree'} = $1;
2999 } elsif ((!defined $withparents) && ($line =~ m/^parent ([0-9a-fA-F]{40})$/)) {
3000 push @parents, $1;
3001 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3002 $co{'author'} = to_utf8($1);
3003 $co{'author_epoch'} = $2;
3004 $co{'author_tz'} = $3;
3005 if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3006 $co{'author_name'} = $1;
3007 $co{'author_email'} = $2;
3008 } else {
3009 $co{'author_name'} = $co{'author'};
3010 }
3011 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3012 $co{'committer'} = to_utf8($1);
3013 $co{'committer_epoch'} = $2;
3014 $co{'committer_tz'} = $3;
3015 if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3016 $co{'committer_name'} = $1;
3017 $co{'committer_email'} = $2;
3018 } else {
3019 $co{'committer_name'} = $co{'committer'};
3020 }
3021 }
3022 }
3023 if (!defined $co{'tree'}) {
3024 return;
3025 };
3026 $co{'parents'} = \@parents;
3027 $co{'parent'} = $parents[0];
3028
3029 foreach my $title (@commit_lines) {
3030 $title =~ s/^ //;
3031 if ($title ne "") {
3032 $co{'title'} = chop_str($title, 80, 5);
3033 # remove leading stuff of merges to make the interesting part visible
3034 if (length($title) > 50) {
3035 $title =~ s/^Automatic //;
3036 $title =~ s/^merge (of|with) /Merge ... /i;
3037 if (length($title) > 50) {
3038 $title =~ s/(http|rsync):\/\///;
3039 }
3040 if (length($title) > 50) {
3041 $title =~ s/(master|www|rsync)\.//;
3042 }
3043 if (length($title) > 50) {
3044 $title =~ s/kernel.org:?//;
3045 }
3046 if (length($title) > 50) {
3047 $title =~ s/\/pub\/scm//;
3048 }
3049 }
3050 $co{'title_short'} = chop_str($title, 50, 5);
3051 last;
3052 }
3053 }
3054 if (! defined $co{'title'} || $co{'title'} eq "") {
3055 $co{'title'} = $co{'title_short'} = '(no commit message)';
3056 }
3057 # remove added spaces
3058 foreach my $line (@commit_lines) {
3059 $line =~ s/^ //;
3060 }
3061 $co{'comment'} = \@commit_lines;
3062
3063 my $age = time - $co{'committer_epoch'};
3064 $co{'age'} = $age;
3065 $co{'age_string'} = age_string($age);
3066 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3067 if ($age > 60*60*24*7*2) {
3068 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3069 $co{'age_string_age'} = $co{'age_string'};
3070 } else {
3071 $co{'age_string_date'} = $co{'age_string'};
3072 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3073 }
3074 return %co;
3075 }
3076
3077 sub parse_commit {
3078 my ($commit_id) = @_;
3079 my %co;
3080
3081 local $/ = "\0";
3082
3083 open my $fd, "-|", git_cmd(), "rev-list",
3084 "--parents",
3085 "--header",
3086 "--max-count=1",
3087 $commit_id,
3088 "--",
3089 or die_error(500, "Open git-rev-list failed");
3090 %co = parse_commit_text(<$fd>, 1);
3091 close $fd;
3092
3093 return %co;
3094 }
3095
3096 sub parse_commits {
3097 my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3098 my @cos;
3099
3100 $maxcount ||= 1;
3101 $skip ||= 0;
3102
3103 local $/ = "\0";
3104
3105 open my $fd, "-|", git_cmd(), "rev-list",
3106 "--header",
3107 @args,
3108 ("--max-count=" . $maxcount),
3109 ("--skip=" . $skip),
3110 @extra_options,
3111 $commit_id,
3112 "--",
3113 ($filename ? ($filename) : ())
3114 or die_error(500, "Open git-rev-list failed");
3115 while (my $line = <$fd>) {
3116 my %co = parse_commit_text($line);
3117 push @cos, \%co;
3118 }
3119 close $fd;
3120
3121 return wantarray ? @cos : \@cos;
3122 }
3123
3124 # parse line of git-diff-tree "raw" output
3125 sub parse_difftree_raw_line {
3126 my $line = shift;
3127 my %res;
3128
3129 # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M ls-files.c'
3130 # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M rev-tree.c'
3131 if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ([0-9a-fA-F]{40}) ([0-9a-fA-F]{40}) (.)([0-9]{0,3})\t(.*)$/) {
3132 $res{'from_mode'} = $1;
3133 $res{'to_mode'} = $2;
3134 $res{'from_id'} = $3;
3135 $res{'to_id'} = $4;
3136 $res{'status'} = $5;
3137 $res{'similarity'} = $6;
3138 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3139 ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3140 } else {
3141 $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3142 }
3143 }
3144 # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3145 # combined diff (for merge commit)
3146 elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:[0-9a-fA-F]{40} )+)([a-zA-Z]+)\t(.*)$//) {
3147 $res{'nparents'} = length($1);
3148 $res{'from_mode'} = [ split(' ', $2) ];
3149 $res{'to_mode'} = pop @{$res{'from_mode'}};
3150 $res{'from_id'} = [ split(' ', $3) ];
3151 $res{'to_id'} = pop @{$res{'from_id'}};
3152 $res{'status'} = [ split('', $4) ];
3153 $res{'to_file'} = unquote($5);
3154 }
3155 # 'c512b523472485aef4fff9e57b229d9d243c967f'
3156 elsif ($line =~ m/^([0-9a-fA-F]{40})$/) {
3157 $res{'commit'} = $1;
3158 }
3159
3160 return wantarray ? %res : \%res;
3161 }
3162
3163 # wrapper: return parsed line of git-diff-tree "raw" output
3164 # (the argument might be raw line, or parsed info)
3165 sub parsed_difftree_line {
3166 my $line_or_ref = shift;
3167
3168 if (ref($line_or_ref) eq "HASH") {
3169 # pre-parsed (or generated by hand)
3170 return $line_or_ref;
3171 } else {
3172 return parse_difftree_raw_line($line_or_ref);
3173 }
3174 }
3175
3176 # parse line of git-ls-tree output
3177 sub parse_ls_tree_line {
3178 my $line = shift;
3179 my %opts = @_;
3180 my %res;
3181
3182 if ($opts{'-l'}) {
3183 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa 16717 panic.c'
3184 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40}) +(-|[0-9]+)\t(.+)$/s;
3185
3186 $res{'mode'} = $1;
3187 $res{'type'} = $2;
3188 $res{'hash'} = $3;
3189 $res{'size'} = $4;
3190 if ($opts{'-z'}) {
3191 $res{'name'} = $5;
3192 } else {
3193 $res{'name'} = unquote($5);
3194 }
3195 } else {
3196 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
3197 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t(.+)$/s;
3198
3199 $res{'mode'} = $1;
3200 $res{'type'} = $2;
3201 $res{'hash'} = $3;
3202 if ($opts{'-z'}) {
3203 $res{'name'} = $4;
3204 } else {
3205 $res{'name'} = unquote($4);
3206 }
3207 }
3208
3209 return wantarray ? %res : \%res;
3210 }
3211
3212 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3213 sub parse_from_to_diffinfo {
3214 my ($diffinfo, $from, $to, @parents) = @_;
3215
3216 if ($diffinfo->{'nparents'}) {
3217 # combined diff
3218 $from->{'file'} = [];
3219 $from->{'href'} = [];
3220 fill_from_file_info($diffinfo, @parents)
3221 unless exists $diffinfo->{'from_file'};
3222 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3223 $from->{'file'}[$i] =
3224 defined $diffinfo->{'from_file'}[$i] ?
3225 $diffinfo->{'from_file'}[$i] :
3226 $diffinfo->{'to_file'};
3227 if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3228 $from->{'href'}[$i] = href(action=>"blob",
3229 hash_base=>$parents[$i],
3230 hash=>$diffinfo->{'from_id'}[$i],
3231 file_name=>$from->{'file'}[$i]);
3232 } else {
3233 $from->{'href'}[$i] = undef;
3234 }
3235 }
3236 } else {
3237 # ordinary (not combined) diff
3238 $from->{'file'} = $diffinfo->{'from_file'};
3239 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3240 $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3241 hash=>$diffinfo->{'from_id'},
3242 file_name=>$from->{'file'});
3243 } else {
3244 delete $from->{'href'};
3245 }
3246 }
3247
3248 $to->{'file'} = $diffinfo->{'to_file'};
3249 if (!is_deleted($diffinfo)) { # file exists in result
3250 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3251 hash=>$diffinfo->{'to_id'},
3252 file_name=>$to->{'file'});
3253 } else {
3254 delete $to->{'href'};
3255 }
3256 }
3257
3258 ## ......................................................................
3259 ## parse to array of hashes functions
3260
3261 sub git_get_heads_list {
3262 my ($limit, @classes) = @_;
3263 @classes = ('heads') unless @classes;
3264 my @patterns = map { "refs/$_" } @classes;
3265 my @headslist;
3266
3267 open my $fd, '-|', git_cmd(), 'for-each-ref',
3268 ($limit ? '--count='.($limit+1) : ()), '--sort=-committerdate',
3269 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3270 @patterns
3271 or return;
3272 while (my $line = <$fd>) {
3273 my %ref_item;
3274
3275 chomp $line;
3276 my ($refinfo, $committerinfo) = split(/\0/, $line);
3277 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3278 my ($committer, $epoch, $tz) =
3279 ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3280 $ref_item{'fullname'} = $name;
3281 $name =~ s!^refs/(?:head|remote)s/!!;
3282
3283 $ref_item{'name'} = $name;
3284 $ref_item{'id'} = $hash;
3285 $ref_item{'title'} = $title || '(no commit message)';
3286 $ref_item{'epoch'} = $epoch;
3287 if ($epoch) {
3288 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3289 } else {
3290 $ref_item{'age'} = "unknown";
3291 }
3292
3293 push @headslist, \%ref_item;
3294 }
3295 close $fd;
3296
3297 return wantarray ? @headslist : \@headslist;
3298 }
3299
3300 sub git_get_tags_list {
3301 my $limit = shift;
3302 my @tagslist;
3303
3304 open my $fd, '-|', git_cmd(), 'for-each-ref',
3305 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3306 '--format=%(objectname) %(objecttype) %(refname) '.
3307 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3308 'refs/tags'
3309 or return;
3310 while (my $line = <$fd>) {
3311 my %ref_item;
3312
3313 chomp $line;
3314 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3315 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3316 my ($creator, $epoch, $tz) =
3317 ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3318 $ref_item{'fullname'} = $name;
3319 $name =~ s!^refs/tags/!!;
3320
3321 $ref_item{'type'} = $type;
3322 $ref_item{'id'} = $id;
3323 $ref_item{'name'} = $name;
3324 if ($type eq "tag") {
3325 $ref_item{'subject'} = $title;
3326 $ref_item{'reftype'} = $reftype;
3327 $ref_item{'refid'} = $refid;
3328 } else {
3329 $ref_item{'reftype'} = $type;
3330 $ref_item{'refid'} = $id;
3331 }
3332
3333 if ($type eq "tag" || $type eq "commit") {
3334 $ref_item{'epoch'} = $epoch;
3335 if ($epoch) {
3336 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3337 } else {
3338 $ref_item{'age'} = "unknown";
3339 }
3340 }
3341
3342 push @tagslist, \%ref_item;
3343 }
3344 close $fd;
3345
3346 return wantarray ? @tagslist : \@tagslist;
3347 }
3348
3349 ## ----------------------------------------------------------------------
3350 ## filesystem-related functions
3351
3352 sub get_file_owner {
3353 my $path = shift;
3354
3355 my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3356 my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3357 if (!defined $gcos) {
3358 return undef;
3359 }
3360 my $owner = $gcos;
3361 $owner =~ s/[,;].*$//;
3362 return to_utf8($owner);
3363 }
3364
3365 # assume that file exists
3366 sub insert_file {
3367 my $filename = shift;
3368
3369 open my $fd, '<', $filename;
3370 print map { to_utf8($_) } <$fd>;
3371 close $fd;
3372 }
3373
3374 ## ......................................................................
3375 ## mimetype related functions
3376
3377 sub mimetype_guess_file {
3378 my $filename = shift;
3379 my $mimemap = shift;
3380 -r $mimemap or return undef;
3381
3382 my %mimemap;
3383 open(my $mh, '<', $mimemap) or return undef;
3384 while (<$mh>) {
3385 next if m/^#/; # skip comments
3386 my ($mimetype, $exts) = split(/\t+/);
3387 if (defined $exts) {
3388 my @exts = split(/\s+/, $exts);
3389 foreach my $ext (@exts) {
3390 $mimemap{$ext} = $mimetype;
3391 }
3392 }
3393 }
3394 close($mh);
3395
3396 $filename =~ /\.([^.]*)$/;
3397 return $mimemap{$1};
3398 }
3399
3400 sub mimetype_guess {
3401 my $filename = shift;
3402 my $mime;
3403 $filename =~ /\./ or return undef;
3404
3405 if ($mimetypes_file) {
3406 my $file = $mimetypes_file;
3407 if ($file !~ m!^/!) { # if it is relative path
3408 # it is relative to project
3409 $file = "$projectroot/$project/$file";
3410 }
3411 $mime = mimetype_guess_file($filename, $file);
3412 }
3413 $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3414 return $mime;
3415 }
3416
3417 sub blob_mimetype {
3418 my $fd = shift;
3419 my $filename = shift;
3420
3421 if ($filename) {
3422 my $mime = mimetype_guess($filename);
3423 $mime and return $mime;
3424 }
3425
3426 # just in case
3427 return $default_blob_plain_mimetype unless $fd;
3428
3429 if (-T $fd) {
3430 return 'text/plain';
3431 } elsif (! $filename) {
3432 return 'application/octet-stream';
3433 } elsif ($filename =~ m/\.png$/i) {
3434 return 'image/png';
3435 } elsif ($filename =~ m/\.gif$/i) {
3436 return 'image/gif';
3437 } elsif ($filename =~ m/\.jpe?g$/i) {
3438 return 'image/jpeg';
3439 } else {
3440 return 'application/octet-stream';
3441 }
3442 }
3443
3444 sub blob_contenttype {
3445 my ($fd, $file_name, $type) = @_;
3446
3447 $type ||= blob_mimetype($fd, $file_name);
3448 if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3449 $type .= "; charset=$default_text_plain_charset";
3450 }
3451
3452 return $type;
3453 }
3454
3455 # guess file syntax for syntax highlighting; return undef if no highlighting
3456 # the name of syntax can (in the future) depend on syntax highlighter used
3457 sub guess_file_syntax {
3458 my ($highlight, $mimetype, $file_name) = @_;
3459 return undef unless ($highlight && defined $file_name);
3460 my $basename = basename($file_name, '.in');
3461 return $highlight_basename{$basename}
3462 if exists $highlight_basename{$basename};
3463
3464 $basename =~ /\.([^.]*)$/;
3465 my $ext = $1 or return undef;
3466 return $highlight_ext{$ext}
3467 if exists $highlight_ext{$ext};
3468
3469 return undef;
3470 }
3471
3472 # run highlighter and return FD of its output,
3473 # or return original FD if no highlighting
3474 sub run_highlighter {
3475 my ($fd, $highlight, $syntax) = @_;
3476 return $fd unless ($highlight && defined $syntax);
3477
3478 close $fd;
3479 open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
3480 quote_command($highlight_bin).
3481 " --replace-tabs=8 --fragment --syntax $syntax |"
3482 or die_error(500, "Couldn't open file or run syntax highlighter");
3483 return $fd;
3484 }
3485
3486 ## ======================================================================
3487 ## functions printing HTML: header, footer, error page
3488
3489 sub get_page_title {
3490 my $title = to_utf8($site_name);
3491
3492 return $title unless (defined $project);
3493 $title .= " - " . to_utf8($project);
3494
3495 return $title unless (defined $action);
3496 $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
3497
3498 return $title unless (defined $file_name);
3499 $title .= " - " . esc_path($file_name);
3500 if ($action eq "tree" && $file_name !~ m|/$|) {
3501 $title .= "/";
3502 }
3503
3504 return $title;
3505 }
3506
3507 sub print_feed_meta {
3508 if (defined $project) {
3509 my %href_params = get_feed_info();
3510 if (!exists $href_params{'-title'}) {
3511 $href_params{'-title'} = 'log';
3512 }
3513
3514 foreach my $format (qw(RSS Atom)) {
3515 my $type = lc($format);
3516 my %link_attr = (
3517 '-rel' => 'alternate',
3518 '-title' => esc_attr("$project - $href_params{'-title'} - $format feed"),
3519 '-type' => "application/$type+xml"
3520 );
3521
3522 $href_params{'action'} = $type;
3523 $link_attr{'-href'} = href(%href_params);
3524 print "<link ".
3525 "rel=\"$link_attr{'-rel'}\" ".
3526 "title=\"$link_attr{'-title'}\" ".
3527 "href=\"$link_attr{'-href'}\" ".
3528 "type=\"$link_attr{'-type'}\" ".
3529 "/>\n";
3530
3531 $href_params{'extra_options'} = '--no-merges';
3532 $link_attr{'-href'} = href(%href_params);