adc0f2084a4ea1aeb47f96af6976936f6cde3b0d
[girocco.git] / Girocco / Util.pm
blobadc0f2084a4ea1aeb47f96af6976936f6cde3b0d
1 package Girocco::Util;
3 use strict;
4 use warnings;
6 use Girocco::Config;
7 use Time::Local;
9 BEGIN {
10 use base qw(Exporter);
11 our @EXPORT = qw(scrypt jailed_file sendmail_pipe mailer_pipe
12 lock_file unlock_file valid_tag rand_adjust
13 filedb_atomic_append filedb_atomic_edit filedb_grep
14 filedb_atomic_grep valid_email valid_email_multi
15 valid_repo_url valid_web_url url_base url_path url_server
16 projects_html_list parse_rfc2822_date parse_any_date);
20 sub scrypt {
21 my ($pwd) = @_;
22 crypt($pwd||'', join ('', ('.', '/', 0..9, 'A'..'Z', 'a'..'z')[rand 64, rand 64]));
25 sub jailed_file {
26 my ($filename) = @_;
27 $filename =~ s,^/,,;
28 $Girocco::Config::chroot."/$filename";
31 sub lock_file {
32 my ($path) = @_;
34 $path .= '.lock';
36 use Errno qw(EEXIST);
37 use Fcntl qw(O_WRONLY O_CREAT O_EXCL);
38 use IO::Handle;
39 my $handle = new IO::Handle;
41 unless (sysopen($handle, $path, O_WRONLY|O_CREAT|O_EXCL)) {
42 my $cnt = 0;
43 while (not sysopen($handle, $path, O_WRONLY|O_CREAT|O_EXCL)) {
44 ($! == EEXIST) or die "$path open failed: $!";
45 ($cnt++ < 16) or die "$path open failed: cannot open lockfile";
46 sleep(1);
49 # XXX: filedb-specific
50 chmod 0664, $path or die "$path g+w failed: $!";
52 $handle;
55 sub _is_passwd_file {
56 return defined($_[0]) && $_[0] eq jailed_file('/etc/passwd');
59 sub _run_update_pwd_db {
60 my ($path, $updatearg) = @_;
61 my @cmd = ($Girocco::Config::basedir.'/bin/update-pwd-db', "$path");
62 push(@cmd, $updatearg) if $updatearg;
63 system(@cmd) == 0 or die "update-pwd-db failed: $?";
66 sub unlock_file {
67 my ($path, $noreplace, $updatearg) = @_;
69 if (!$noreplace) {
70 _run_update_pwd_db("$path.lock", $updatearg)
71 if $Girocco::Config::update_pwd_db && _is_passwd_file($path);
72 rename "$path.lock", $path or die "$path unlock failed: $!";
73 } else {
74 unlink "$path.lock" or die "$path unlock failed: $!";
78 sub filedb_atomic_append {
79 my ($file, $line, $updatearg) = @_;
80 my $id = 65536;
82 open my $src, '<', $file or die "$file open for reading failed: $!";
83 my $dst = lock_file($file);
85 while (<$src>) {
86 my $aid = (split /:/)[2];
87 $id = $aid + 1 if ($aid >= $id);
89 print $dst $_ or die "$file(l) write failed: $!";
92 $line =~ s/\\i/$id/g;
93 print $dst "$line\n" or die "$file(l) write failed: $!";
95 close $dst or die "$file(l) close failed: $!";
96 close $src;
98 unlock_file($file, 0, $updatearg);
100 $id;
103 sub filedb_atomic_edit {
104 my ($file, $fn, $updatearg) = @_;
106 open my $src, '<', $file or die "$file open for reading failed: $!";
107 my $dst = lock_file($file);
109 while (<$src>) {
110 print $dst $fn->($_) or die "$file(l) write failed: $!";
113 close $dst or die "$file(l) close failed: $!";
114 close $src;
116 unlock_file($file, 0, $updatearg);
119 sub filedb_atomic_grep {
120 my ($file, $fn) = @_;
121 my @results = ();
123 open my $src, '<', $file or die "$file open for reading failed: $!";
124 my $dst = lock_file($file);
126 while (<$src>) {
127 my $result = $fn->($_);
128 push(@results, $result) if $result;
131 close $dst or die "$file(l) close failed: $!";
132 close $src;
134 unlock_file($file, 1);
135 return @results;
138 sub filedb_grep {
139 my ($file, $fn) = @_;
140 my @results = ();
142 open my $src, '<', $file or die "$file open for reading failed: $!";
144 while (<$src>) {
145 my $result = $fn->($_);
146 push(@results, $result) if $result;
149 close $src;
151 return @results;
154 sub valid_email {
155 local $_ = $_[0];
156 /^[a-zA-Z0-9+._-]+@[a-zA-Z0-9.-]+$/;
158 sub valid_email_multi {
159 local $_ = $_[0];
160 # More relaxed, we just want to avoid too dangerous characters.
161 /^[a-zA-Z0-9+._, @-]+$/;
163 sub valid_web_url {
164 local $_ = $_[0];
165 /^https?:\/\/[a-zA-Z0-9.:-]+(\/[_\%a-zA-Z0-9.\/~:?&=;-]*)?(#[a-zA-Z0-9._-]+)?$/;
167 sub valid_repo_url {
168 my $url = shift || '';
169 # Currently neither username nor password is allowed in the URL and IPv6
170 # literal addresses are not accepted either.
171 $Girocco::Config::mirror_svn &&
172 $url =~ /^svn(\+https?)?:\/\/[a-zA-Z0-9.:-]+(\/[_\%a-zA-Z0-9.\/~-]*)?$/os
173 and return 1;
174 $Girocco::Config::mirror_darcs &&
175 $url =~ /^darcs:\/\/[a-zA-Z0-9.:-]+(\/[_\%a-zA-Z0-9.\/~-]*)?$/os
176 and return 1;
177 $Girocco::Config::mirror_bzr &&
178 $url =~ /^bzr:\/\/[a-zA-Z0-9.:-]+(\/[_\%a-zA-Z0-9.\/~-]*)?$/os
179 and return 1;
180 return $url =~ /^(https?|git):\/\/[a-zA-Z0-9.:-]+(\/[_\%a-zA-Z0-9.\/~-]*)?$/;
182 my %_badtags;
183 BEGIN {
184 %_badtags = (
185 about=>1, after=>1, all=>1, also=>1, an=>1, and=>1, another=>1, any=>1,
186 are=>1, as=>1, at=>1, be=>1, because=>1, been=>1, before=>1, being=>1,
187 between=>1, both=>1, but=>1, by=>1, came=>1, can=>1, come=>1, could=>1,
188 did=>1, do=>1, each=>1, for=>1, from=>1, get=>1, got=>1, had=>1, has=>1,
189 have=>1, he=>1, her=>1, here=>1, him=>1, himself=>1, his=>1, how=>1,
190 if=>1, in=>1, into=>1, is=>1, it=>1, like=>1, make=>1, many=>1, me=>1,
191 might=>1, more=>1, most=>1, much=>1, must=>1, my=>1, never=>1, now=>1,
192 of=>1, on=>1, only=>1, or=>1, other=>1, our=>1, out=>1, over=>1,
193 said=>1, same=>1, see=>1, should=>1, since=>1, some=>1, still=>1,
194 such=>1, take=>1, than=>1, that=>1, the=>1, their=>1, them=>1, then=>1,
195 there=>1, these=>1, they=>1, this=>1, those=>1, through=>1, to=>1,
196 too=>1, under=>1, up=>1, very=>1, was=>1, way=>1, we=>1, well=>1,
197 were=>1, what=>1, where=>1, which=>1, while=>1, who=>1, with=>1,
198 would=>1, you=>1, your=>1
201 # A valid tag must only have [a-zA-Z0-9:.+#_-] characters, must start with a
202 # letter, must not be a noise word and except for 'C' must be more than one
203 # character long and no more than 32 characters long.
204 sub valid_tag {
205 local $_ = $_[0] || '';
206 return 1 if $_ eq 'C'; # Currently only allowed single letter tag
207 return 0 unless /^[a-zA-Z][a-zA-Z0-9:.+#_-]+$/;
208 return 0 if $_badtags{lc($_)};
209 return length($_) <= 32 ? 1 : 0;
212 # If the passed in argument looks like a URL, return only the stuff up through
213 # the host:port part otherwise return the entire argument.
214 sub url_base {
215 my $url = shift || '';
216 # See RFC 3968
217 $url = $1.$2.$3.$4 if $url =~ m,^( [A-Za-z][A-Za-z0-9+.-]*: ) # scheme
218 ( // ) # // separator
219 ((?:[^\@]+\@)?) # optional userinfo
220 ( [^/?#]+ ) # host and port
221 (?:[/?#].*)?$,x; # path and optional query string and/or anchor
222 return $url;
225 # If the passed in argument looks like a URL, return only the stuff following
226 # the host:port part otherwise return the entire argument.
227 sub url_path {
228 my $url = shift || '';
229 my $no_empty = shift || 0;
230 # See RFC 3968
231 $url = $1 if $url =~ m,^(?: [A-Za-z][A-Za-z0-9+.-]*: ) # scheme
232 (?: // ) # // separator
233 (?: [^\@]+\@ )? # optional userinfo
234 (?: [^/?#]+ ) # host and port
235 ((?:[/?#].*)?)$,x; # path and optional query string and/or anchor
236 $url = '/' if $no_empty && $url eq '';
237 return $url;
240 # If both SERVER_NAME and SERVER_PORT are set pass the argument through url_path
241 # and then prefix it with the appropriate scheme (HTTPS=?on), host and port and
242 # return it. If a something that doesn't look like it could be the start of a
243 # URL path comes back from url_path or SERVER_NAME is a link-local IPv6 address
244 # then just return the argument unchanged.
245 sub url_server {
246 my $url = shift || '';
247 my $path = url_path($url);
248 return $url unless $path eq '' || $path =~ m|^[/?#]|;
249 return $url unless $ENV{'SERVER_NAME'} && $ENV{'SERVER_PORT'} &&
250 $ENV{'SERVER_PORT'} =~ /^[1-9][0-9]{0,4}$/;
251 return $url if $ENV{'SERVER_NAME'} =~ /^[[]?fe80:/i;
252 my $server = $ENV{'SERVER_NAME'};
253 # Deal with Apache bug where IPv6 literal server names do not include
254 # the required surrounding '[' and ']' characters
255 $server = '[' . $server . ']' if $server =~ /:/ && $server !~ /^[[]/;
256 my $ishttps = $ENV{'HTTPS'} && $ENV{'HTTPS'} =~ /^on$/i;
257 my $portnum = 0 + $ENV{'SERVER_PORT'};
258 my $port = '';
259 if (($ishttps && $portnum != 443) || (!$ishttps && $portnum != 80)) {
260 $port = ':' . $portnum;
262 return 'http' . ($ishttps ? 's' : '') . '://' . $server . $port . $path;
265 sub _escapeHTML {
266 my $str = shift;
267 $str =~ s/\&/\&amp;/gs;
268 $str =~ s/\</\&lt;/gs;
269 $str =~ s/\>/\&gt;/gs;
270 $str =~ s/\"/\&quot;/gs; #"
271 return $str;
274 # create relative time string from passed in age in seconds
275 sub _rel_age {
276 my $age = shift;
277 my $age_str;
279 if ($age > 60*60*24*365*2) {
280 $age_str = (int $age/60/60/24/365);
281 $age_str .= " years ago";
282 } elsif ($age > 60*60*24*(365/12)*2) {
283 $age_str = int $age/60/60/24/(365/12);
284 $age_str .= " months ago";
285 } elsif ($age > 60*60*24*7*2) {
286 $age_str = int $age/60/60/24/7;
287 $age_str .= " weeks ago";
288 } elsif ($age > 60*60*24*2) {
289 $age_str = int $age/60/60/24;
290 $age_str .= " days ago";
291 } elsif ($age > 60*60*2) {
292 $age_str = int $age/60/60;
293 $age_str .= " hours ago";
294 } elsif ($age > 60*2) {
295 $age_str = int $age/60;
296 $age_str .= " mins ago";
297 } elsif ($age > 2) {
298 $age_str = int $age;
299 $age_str .= " secs ago";
300 } elsif ($age >= 0) {
301 $age_str = "right now";
302 } else {
303 $age_str = "future time";
305 return $age_str;
308 # create relative time string from passed in idle in seconds
309 sub _rel_idle {
310 my $idle_str = _rel_age(shift);
311 $idle_str =~ s/ ago//;
312 $idle_str = "not at all" if $idle_str eq "right now";
313 return $idle_str;
316 sub _strftime {
317 use POSIX qw(strftime);
318 my ($fmt, $secs, $zonesecs) = @_;
319 my ($S,$M,$H,$d,$m,$y) = gmtime($secs + $zonesecs);
320 $zonesecs = int($zonesecs / 60);
321 $fmt =~ s/%z/\$z/g;
322 my $ans = strftime($fmt, $S, $M, $H, $d, $m, $y, -1, -1, -1);
323 my $z;
324 if ($zonesecs < 0) {
325 $z = "-";
326 $zonesecs = -$zonesecs;
327 } else {
328 $z = "+";
330 $z .= sprintf("%02d%02d", int($zonesecs/60), $zonesecs % 60);
331 $ans =~ s/\$z/$z/g;
332 return $ans;
335 # Take a list of project names and produce a nicely formated table that
336 # includes owner links and descriptions. If the list is empty returns ''.
337 # The first argument may be a hash ref that contains options. The following
338 # options are available:
339 # target -- sets the target value of the owner link
340 # emptyok -- if true returns an empty table rather than ''
341 # typecol -- if true include type column with hover info
342 # changed -- if true include a changed and idle column
343 sub projects_html_list {
344 my $options = {};
345 if (defined($_[0]) && ref($_[0]) eq 'HASH') {
346 $options = shift;
348 return '' unless @_ || (defined($options->{emptyok}) && $options->{emptyok});
349 require Girocco::Project;
350 my $count = 0;
351 my $target = '';
352 $target = " target=\""._escapeHTML($options->{target})."\""
353 if defined($options->{target});
354 my $withtype = defined($options->{typecol}) && $options->{typecol};
355 my $withchanged = defined($options->{changed}) && $options->{changed};
356 my $typehead = '';
357 $typehead = '<th>Type</th>' if $withtype;
358 my $chghead = '';
359 $chghead = substr(<<EOT, 0, -1) if $withchanged;
360 <th><span class="hover">Changed<span><span class="head">Changed</span
361 />The last time a ref change was received by this site.</span></span></th
362 ><th><span class="hover">Idle<span><span class="head">Idle</span
363 />The most recent committer time in <i>refs/heads</i>.</span></span></th
366 my $html = <<EOT;
367 <table class='projectlist'><tr><th>Project</th>$typehead$chghead<th class="desc">Description</th></tr>
369 my $trclass = ' class="odd"';
370 foreach (sort({lc($a) cmp lc($b)} @_)) {
371 if (Girocco::Project::does_exist($_, 1)) {
372 my $proj = Girocco::Project->load($_);
373 my $projname = $proj->{name}.".git";
374 my $projdesc = $proj->{desc}||'';
375 utf8::decode($projdesc) if utf8::valid($projdesc);
376 my $typecol = '';
377 if ($withtype) {
378 if ($proj->{mirror}) {
379 $typecol = substr(<<EOT, 0, -1);
380 <td class="type"><span class="hover">mirror<span class="nowrap">@{[_escapeHTML($proj->{url})]}</span></span></td>
382 } else {
383 my $users = @{$proj->{users}};
384 $users .= ' user';
385 $users .= 's' unless @{$proj->{users}} == 1;
386 my $userlist = join(', ', sort({lc($a) cmp lc($b)} @{$proj->{users}}));
387 my $spncls = length($userlist) > 25 ? '' : ' class="nowrap"';
388 $typecol = $userlist ? substr(<<EOT, 0, -1) : substr(<<EOT, 0, -1);
389 <td class="type"><span class="hover">$users<span$spncls>$userlist</span></span></td>
391 <td class="type">$users</td>
395 my $changecol = '';
396 if ($withchanged) {
397 my $rel = '';
398 my $changetime = $proj->{lastchange};
399 if ($changetime) {
400 $rel = "<span class=\"hover\">" .
401 _rel_age(time - parse_rfc2822_date($changetime)) .
402 "<span class=\"nowrap\">$changetime</span></span>";
403 } else {
404 $rel = "no commits";
406 $changecol = substr(<<EOT, 0, -1);
407 <td class="change">$rel</td>
409 my $idletime = $proj->{lastactivity};
410 my ($idlesecs, $tz);
411 $idlesecs = parse_any_date($idletime, \$tz) if $idletime;
412 if ($idlesecs) {
413 my $idle2822 = _strftime("%a, %d %b %Y %T %z", $idlesecs, $tz);
414 $rel = "<span class=\"hover\">" .
415 _rel_idle(time - $idlesecs) .
416 "<span class=\"nowrap\">$idle2822</span></span>";
417 } else {
418 $rel = "no commits";
420 $changecol .= substr(<<EOT, 0, -1);
421 <td class="idle">$rel</td>
424 $html .= <<EOT;
425 <tr$trclass><td><a href="@{[url_path($Girocco::Config::gitweburl)]}/$projname"$target
426 >@{[_escapeHTML($projname)]}</td>$typecol$changecol<td>@{[_escapeHTML($projdesc)]}</td></tr>
428 $trclass = $trclass ? '' : ' class="odd"';
429 ++$count;
432 $html .= <<EOT;
433 </table>
435 return ($count || (defined($options->{emptyok}) && $options->{emptyok})) ? $html : '';
438 my %_month_names;
439 BEGIN {
440 %_month_names = (
441 jan => 0, feb => 1, mar => 2, apr => 3, may => 4, jun => 5,
442 jul => 6, aug => 7, sep => 8, oct => 9, nov => 10, dec => 11
446 # Should be in "date '+%a, %d %b %Y %T %z'" format as saved to lastgc, lastrefresh and lastchange
447 # The leading "%a, " is optional, returns undef if unrecognized date. This is also known as
448 # RFC 2822 date format and git's '%cD', '%aD' and --date=rfc2822 format.
449 # If the second argument is a SCALAR ref, its value will be set to the TZ offset in seconds
450 sub parse_rfc2822_date {
451 my $dstr = shift || '';
452 my $tzoff = shift || '';
453 $dstr = $1 if $dstr =~/^[^\s]+,\s*(.*)$/;
454 return undef unless $dstr =~
455 /^\s*(\d{1,2})\s+([A-Za-z]{3})\s+(\d{4})\s+(\d{1,2}):(\d{2}):(\d{2})\s+([+-]\d{4})\s*$/;
456 my ($d,$b,$Y,$H,$M,$S,$z) = ($1,$2,$3,$4,$5,$6,$7);
457 my $m = $_month_names{lc($b)};
458 return undef unless defined($m);
459 my $seconds = timegm(0+$S, 0+$M, 0+$H, 0+$d, 0+$m, $Y-1900);
460 my $offset = 60 * (60 * (0+substr($z,1,2)) + (0+substr($z,3,2)));
461 $offset = -$offset if substr($z,0,1) eq '-';
462 $$tzoff = $offset if ref($tzoff) eq 'SCALAR';
463 return $seconds - $offset;
466 # Will parse any supported date format. Actually there are three formats
467 # currently supported:
468 # 1. RFC 2822 (uses parse_rfc2822_date)
469 # 2. RFC 3339 / ISO 8601 (T may be ' ', 'Z' is optional, ':' optional in TZ)
470 # 3. unix seconds since epoch with optional +/- trailing TZ (may not have a ':')
471 # Returns undef if unsupported date.
472 # If the second argument is a SCALAR ref, its value will be set to the TZ offset in seconds
473 sub parse_any_date {
474 my $dstr = shift || '';
475 my $tzoff = shift || '';
476 if ($dstr =~ /^\s*([-+]?\d+)(?:\s+([-+]\d{4}))?\s*$/) {
477 # Unix timestamp
478 my $ts = 0 + $1;
479 my $off = 0;
480 if ($2) {
481 my $z = $2;
482 $off = 60 * (60 * (0+substr($z,1,2)) + (0+substr($z,3,2)));
483 $off = -$off if substr($z,0,1) eq '-';
485 $$tzoff = $off if ref($tzoff) eq 'SCALAR';
486 return $ts;
488 if ($dstr =~ /^\s*(\d{4})-(\d{2})-(\d{2})[Tt ](\d{2}):(\d{2}):(\d{2})(?:[ ]([Zz]|(?:[-+]\d{2}:?\d{2})))?\s*$/) {
489 my ($Y,$m,$d,$H,$M,$S,$z) = ($1,$2,$3,$4,$5,$6,$7||'');
490 my $seconds = timegm(0+$S, 0+$M, 0+$H, 0+$d, $m-1, $Y-1900);
491 $z =~ s/://;
492 my $off = 0;
493 if (uc($z) ne 'Z') {
494 $off = 60 * (60 * (0+substr($z,1,2)) + (0+substr($z,3,2)));
495 $off = -$off if substr($z,0,1) eq '-';
497 $$tzoff = $off if ref($tzoff) eq 'SCALAR';
498 return $seconds - $off;
500 return parse_rfc2822_date($dstr, $tzoff);
503 # Input is a number such as a minute interval
504 # Return value is a random number between the input and 1.25*input
505 # This can be used to randomize the update and gc operations a bit to avoid
506 # having them all end up all clustered together
507 sub rand_adjust {
508 my $input = shift || 0;
509 return $input unless $input;
510 return $input + int(rand(0.25 * $input));
513 # Open a pipe to a new sendmail process. The '-i' option is always passed to
514 # the new process followed by any addtional arguments passed in. Note that
515 # the sendmail process is only expected to understand the '-i', '-t' and '-f'
516 # options. Using any other options via this function is not guaranteed to work.
517 # A list of recipients may follow the options. Combining a list of recipients
518 # with the '-t' option is not recommended.
519 sub sendmail_pipe {
520 return undef unless @_;
521 die "\$Girocco::Config::sendmail_bin is unset or not executable!\n"
522 unless $Girocco::Config::sendmail_bin && -x $Girocco::Config::sendmail_bin;
523 my $result = open(my $pipe, '|-', $Girocco::Config::sendmail_bin, '-i', @_);
524 return $result ? $pipe : undef;
527 # Open a pipe that works similarly to a mailer such as /usr/bin/mail in that
528 # if the first argument is '-s', a subject line will be automatically added
529 # (using the second argument as the subject). Any remaining arguments are
530 # expected to be recipient addresses that will be added to an explicit To:
531 # line as well as passed on to sendmail_pipe. In addition an
532 # "Auto-Submitted: auto-generated" header is always added as well as a suitable
533 # "From:" header.
534 sub mailer_pipe {
535 my $subject = undef;
536 if (@_ >= 2 && $_[0] eq '-s') {
537 shift;
538 $subject = shift;
540 my $tolist = join(", ", @_);
541 unshift(@_, '-f', $Girocco::Config::sender) if $Girocco::Config::sender;
542 my $pipe = sendmail_pipe(@_);
543 if ($pipe) {
544 print $pipe "From: \"$Girocco::Config::name\" ",
545 "($Girocco::Config::title) ",
546 "<$Girocco::Config::admin>\n";
547 print $pipe "To: $tolist\n";
548 print $pipe "Subject: $subject\n" if defined($subject);
549 print $pipe "Auto-Submitted: auto-generated\n";
550 print $pipe "\n";
552 return $pipe;