8fcd4985e8933172737c726cf6386888479601bd
[girocco.git] / Girocco / Util.pm
blob8fcd4985e8933172737c726cf6386888479601bd
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 local $_ = $_[0];
169 /^(https?|git|svn(\+http)?|svn(\+https)?|darcs|bzr):\/\/[a-zA-Z0-9.:-]+(\/[_\%a-zA-Z0-9.\/~-]*)?$/;
171 my %_badtags;
172 BEGIN {
173 %_badtags = (
174 about=>1, after=>1, all=>1, also=>1, an=>1, and=>1, another=>1, any=>1,
175 are=>1, as=>1, at=>1, be=>1, because=>1, been=>1, before=>1, being=>1,
176 between=>1, both=>1, but=>1, by=>1, came=>1, can=>1, come=>1, could=>1,
177 did=>1, do=>1, each=>1, for=>1, from=>1, get=>1, got=>1, had=>1, has=>1,
178 have=>1, he=>1, her=>1, here=>1, him=>1, himself=>1, his=>1, how=>1,
179 if=>1, in=>1, into=>1, is=>1, it=>1, like=>1, make=>1, many=>1, me=>1,
180 might=>1, more=>1, most=>1, much=>1, must=>1, my=>1, never=>1, now=>1,
181 of=>1, on=>1, only=>1, or=>1, other=>1, our=>1, out=>1, over=>1,
182 said=>1, same=>1, see=>1, should=>1, since=>1, some=>1, still=>1,
183 such=>1, take=>1, than=>1, that=>1, the=>1, their=>1, them=>1, then=>1,
184 there=>1, these=>1, they=>1, this=>1, those=>1, through=>1, to=>1,
185 too=>1, under=>1, up=>1, very=>1, was=>1, way=>1, we=>1, well=>1,
186 were=>1, what=>1, where=>1, which=>1, while=>1, who=>1, with=>1,
187 would=>1, you=>1, your=>1
190 # A valid tag must only have [a-zA-Z0-9:.+#_-] characters, must start with a
191 # letter, must not be a noise word and except for 'C' must be more than one
192 # character long and no more than 32 characters long.
193 sub valid_tag {
194 local $_ = $_[0] || '';
195 return 1 if $_ eq 'C'; # Currently only allowed single letter tag
196 return 0 unless /^[a-zA-Z][a-zA-Z0-9:.+#_-]+$/;
197 return 0 if $_badtags{lc($_)};
198 return length($_) <= 32 ? 1 : 0;
201 # If the passed in argument looks like a URL, return only the stuff up through
202 # the host:port part otherwise return the entire argument.
203 sub url_base {
204 my $url = shift || '';
205 # See RFC 3968
206 $url = $1.$2.$3.$4 if $url =~ m,^( [A-Za-z][A-Za-z0-9+.-]*: ) # scheme
207 ( // ) # // separator
208 ((?:[^\@]+\@)?) # optional userinfo
209 ( [^/?#]+ ) # host and port
210 (?:[/?#].*)?$,x; # path and optional query string and/or anchor
211 return $url;
214 # If the passed in argument looks like a URL, return only the stuff following
215 # the host:port part otherwise return the entire argument.
216 sub url_path {
217 my $url = shift || '';
218 my $no_empty = shift || 0;
219 # See RFC 3968
220 $url = $1 if $url =~ m,^(?: [A-Za-z][A-Za-z0-9+.-]*: ) # scheme
221 (?: // ) # // separator
222 (?: [^\@]+\@ )? # optional userinfo
223 (?: [^/?#]+ ) # host and port
224 ((?:[/?#].*)?)$,x; # path and optional query string and/or anchor
225 $url = '/' if $no_empty && $url eq '';
226 return $url;
229 # If both SERVER_NAME and SERVER_PORT are set pass the argument through url_path
230 # and then prefix it with the appropriate scheme (HTTPS=?on), host and port and
231 # return it. If a something that doesn't look like it could be the start of a
232 # URL path comes back from url_path or SERVER_NAME is a link-local IPv6 address
233 # then just return the argument unchanged.
234 sub url_server {
235 my $url = shift || '';
236 my $path = url_path($url);
237 return $url unless $path eq '' || $path =~ m|^[/?#]|;
238 return $url unless $ENV{'SERVER_NAME'} && $ENV{'SERVER_PORT'} &&
239 $ENV{'SERVER_PORT'} =~ /^[1-9][0-9]{0,4}$/;
240 return $url if $ENV{'SERVER_NAME'} =~ /^[[]?fe80:/i;
241 my $server = $ENV{'SERVER_NAME'};
242 # Deal with Apache bug where IPv6 literal server names do not include
243 # the required surrounding '[' and ']' characters
244 $server = '[' . $server . ']' if $server =~ /:/ && $server !~ /^[[]/;
245 my $ishttps = $ENV{'HTTPS'} && $ENV{'HTTPS'} =~ /^on$/i;
246 my $portnum = 0 + $ENV{'SERVER_PORT'};
247 my $port = '';
248 if (($ishttps && $portnum != 443) || (!$ishttps && $portnum != 80)) {
249 $port = ':' . $portnum;
251 return 'http' . ($ishttps ? 's' : '') . '://' . $server . $port . $path;
254 sub _escapeHTML {
255 my $str = shift;
256 $str =~ s/\&/\&amp;/gs;
257 $str =~ s/\</\&lt;/gs;
258 $str =~ s/\>/\&gt;/gs;
259 $str =~ s/\"/\&quot;/gs; #"
260 return $str;
263 # create relative time string from passed in age in seconds
264 sub _rel_age {
265 my $age = shift;
266 my $age_str;
268 if ($age > 60*60*24*365*2) {
269 $age_str = (int $age/60/60/24/365);
270 $age_str .= " years ago";
271 } elsif ($age > 60*60*24*(365/12)*2) {
272 $age_str = int $age/60/60/24/(365/12);
273 $age_str .= " months ago";
274 } elsif ($age > 60*60*24*7*2) {
275 $age_str = int $age/60/60/24/7;
276 $age_str .= " weeks ago";
277 } elsif ($age > 60*60*24*2) {
278 $age_str = int $age/60/60/24;
279 $age_str .= " days ago";
280 } elsif ($age > 60*60*2) {
281 $age_str = int $age/60/60;
282 $age_str .= " hours ago";
283 } elsif ($age > 60*2) {
284 $age_str = int $age/60;
285 $age_str .= " mins ago";
286 } elsif ($age > 2) {
287 $age_str = int $age;
288 $age_str .= " secs ago";
289 } elsif ($age >= 0) {
290 $age_str = "right now";
291 } else {
292 $age_str = "future time";
294 return $age_str;
297 # create relative time string from passed in idle in seconds
298 sub _rel_idle {
299 my $idle_str = _rel_age(shift);
300 $idle_str =~ s/ ago//;
301 $idle_str = "not at all" if $idle_str eq "right now";
302 return $idle_str;
305 sub _strftime {
306 use POSIX qw(strftime);
307 my ($fmt, $secs, $zonesecs) = @_;
308 my ($S,$M,$H,$d,$m,$y) = gmtime($secs + $zonesecs);
309 $zonesecs = int($zonesecs / 60);
310 $fmt =~ s/%z/\$z/g;
311 my $ans = strftime($fmt, $S, $M, $H, $d, $m, $y, -1, -1, -1);
312 my $z;
313 if ($zonesecs < 0) {
314 $z = "-";
315 $zonesecs = -$zonesecs;
316 } else {
317 $z = "+";
319 $z .= sprintf("%02d%02d", int($zonesecs/60), $zonesecs % 60);
320 $ans =~ s/\$z/$z/g;
321 return $ans;
324 # Take a list of project names and produce a nicely formated table that
325 # includes owner links and descriptions. If the list is empty returns ''.
326 # The first argument may be a hash ref that contains options. The following
327 # options are available:
328 # target -- sets the target value of the owner link
329 # emptyok -- if true returns an empty table rather than ''
330 # typecol -- if true include type column with hover info
331 # changed -- if true include a changed and idle column
332 sub projects_html_list {
333 my $options = {};
334 if (defined($_[0]) && ref($_[0]) eq 'HASH') {
335 $options = shift;
337 return '' unless @_ || (defined($options->{emptyok}) && $options->{emptyok});
338 require Girocco::Project;
339 my $count = 0;
340 my $target = '';
341 $target = " target=\""._escapeHTML($options->{target})."\""
342 if defined($options->{target});
343 my $withtype = defined($options->{typecol}) && $options->{typecol};
344 my $withchanged = defined($options->{changed}) && $options->{changed};
345 my $typehead = '';
346 $typehead = '<th>Type</th>' if $withtype;
347 my $chghead = '';
348 $chghead = substr(<<EOT, 0, -1) if $withchanged;
349 <th><span class="hover">Changed<span><span class="head">Changed</span
350 />The last time a ref change was received by this site.</span></span></th
351 ><th><span class="hover">Idle<span><span class="head">Idle</span
352 />The most recent committer time in <i>refs/heads</i>.</span></span></th
355 my $html = <<EOT;
356 <table class='projectlist'><tr><th>Project</th>$typehead$chghead<th class="desc">Description</th></tr>
358 my $trclass = ' class="odd"';
359 foreach (sort({lc($a) cmp lc($b)} @_)) {
360 if (Girocco::Project::does_exist($_, 1)) {
361 my $proj = Girocco::Project->load($_);
362 my $projname = $proj->{name}.".git";
363 my $projdesc = $proj->{desc}||'';
364 utf8::decode($projdesc) if utf8::valid($projdesc);
365 my $typecol = '';
366 if ($withtype) {
367 if ($proj->{mirror}) {
368 $typecol = substr(<<EOT, 0, -1);
369 <td class="type"><span class="hover">mirror<span class="nowrap">@{[_escapeHTML($proj->{url})]}</span></span></td>
371 } else {
372 my $users = @{$proj->{users}};
373 $users .= ' user';
374 $users .= 's' unless @{$proj->{users}} == 1;
375 my $userlist = join(', ', sort({lc($a) cmp lc($b)} @{$proj->{users}}));
376 my $spncls = length($userlist) > 25 ? '' : ' class="nowrap"';
377 $typecol = $userlist ? substr(<<EOT, 0, -1) : substr(<<EOT, 0, -1);
378 <td class="type"><span class="hover">$users<span$spncls>$userlist</span></span></td>
380 <td class="type">$users</td>
384 my $changecol = '';
385 if ($withchanged) {
386 my $rel = '';
387 my $changetime = $proj->{lastchange};
388 if ($changetime) {
389 $rel = "<span class=\"hover\">" .
390 _rel_age(time - parse_rfc2822_date($changetime)) .
391 "<span class=\"nowrap\">$changetime</span></span>";
392 } else {
393 $rel = "no commits";
395 $changecol = substr(<<EOT, 0, -1);
396 <td class="change">$rel</td>
398 my $idletime = $proj->{lastactivity};
399 my ($idlesecs, $tz);
400 $idlesecs = parse_any_date($idletime, \$tz) if $idletime;
401 if ($idlesecs) {
402 my $idle2822 = _strftime("%a, %d %b %Y %T %z", $idlesecs, $tz);
403 $rel = "<span class=\"hover\">" .
404 _rel_idle(time - $idlesecs) .
405 "<span class=\"nowrap\">$idle2822</span></span>";
406 } else {
407 $rel = "no commits";
409 $changecol .= substr(<<EOT, 0, -1);
410 <td class="idle">$rel</td>
413 $html .= <<EOT;
414 <tr$trclass><td><a href="@{[url_path($Girocco::Config::gitweburl)]}/$projname"$target
415 >@{[_escapeHTML($projname)]}</td>$typecol$changecol<td>@{[_escapeHTML($projdesc)]}</td></tr>
417 $trclass = $trclass ? '' : ' class="odd"';
418 ++$count;
421 $html .= <<EOT;
422 </table>
424 return ($count || (defined($options->{emptyok}) && $options->{emptyok})) ? $html : '';
427 my %_month_names;
428 BEGIN {
429 %_month_names = (
430 jan => 0, feb => 1, mar => 2, apr => 3, may => 4, jun => 5,
431 jul => 6, aug => 7, sep => 8, oct => 9, nov => 10, dec => 11
435 # Should be in "date '+%a, %d %b %Y %T %z'" format as saved to lastgc, lastrefresh and lastchange
436 # The leading "%a, " is optional, returns undef if unrecognized date. This is also known as
437 # RFC 2822 date format and git's '%cD', '%aD' and --date=rfc2822 format.
438 # If the second argument is a SCALAR ref, its value will be set to the TZ offset in seconds
439 sub parse_rfc2822_date {
440 my $dstr = shift || '';
441 my $tzoff = shift || '';
442 $dstr = $1 if $dstr =~/^[^\s]+,\s*(.*)$/;
443 return undef unless $dstr =~
444 /^\s*(\d{1,2})\s+([A-Za-z]{3})\s+(\d{4})\s+(\d{1,2}):(\d{2}):(\d{2})\s+([+-]\d{4})\s*$/;
445 my ($d,$b,$Y,$H,$M,$S,$z) = ($1,$2,$3,$4,$5,$6,$7);
446 my $m = $_month_names{lc($b)};
447 return undef unless defined($m);
448 my $seconds = timegm(0+$S, 0+$M, 0+$H, 0+$d, 0+$m, $Y-1900);
449 my $offset = 60 * (60 * (0+substr($z,1,2)) + (0+substr($z,3,2)));
450 $offset = -$offset if substr($z,0,1) eq '-';
451 $$tzoff = $offset if ref($tzoff) eq 'SCALAR';
452 return $seconds - $offset;
455 # Will parse any supported date format. Actually there are three formats
456 # currently supported:
457 # 1. RFC 2822 (uses parse_rfc2822_date)
458 # 2. RFC 3339 / ISO 8601 (T may be ' ', 'Z' is optional, ':' optional in TZ)
459 # 3. unix seconds since epoch with optional +/- trailing TZ (may not have a ':')
460 # Returns undef if unsupported date.
461 # If the second argument is a SCALAR ref, its value will be set to the TZ offset in seconds
462 sub parse_any_date {
463 my $dstr = shift || '';
464 my $tzoff = shift || '';
465 if ($dstr =~ /^\s*([-+]?\d+)(?:\s+([-+]\d{4}))?\s*$/) {
466 # Unix timestamp
467 my $ts = 0 + $1;
468 my $off = 0;
469 if ($2) {
470 my $z = $2;
471 $off = 60 * (60 * (0+substr($z,1,2)) + (0+substr($z,3,2)));
472 $off = -$off if substr($z,0,1) eq '-';
474 $$tzoff = $off if ref($tzoff) eq 'SCALAR';
475 return $ts;
477 if ($dstr =~ /^\s*(\d{4})-(\d{2})-(\d{2})[Tt ](\d{2}):(\d{2}):(\d{2})(?:[ ]([Zz]|(?:[-+]\d{2}:?\d{2})))?\s*$/) {
478 my ($Y,$m,$d,$H,$M,$S,$z) = ($1,$2,$3,$4,$5,$6,$7||'');
479 my $seconds = timegm(0+$S, 0+$M, 0+$H, 0+$d, $m-1, $Y-1900);
480 $z =~ s/://;
481 my $off = 0;
482 if (uc($z) ne 'Z') {
483 $off = 60 * (60 * (0+substr($z,1,2)) + (0+substr($z,3,2)));
484 $off = -$off if substr($z,0,1) eq '-';
486 $$tzoff = $off if ref($tzoff) eq 'SCALAR';
487 return $seconds - $off;
489 return parse_rfc2822_date($dstr, $tzoff);
492 # Input is a number such as a minute interval
493 # Return value is a random number between the input and 1.25*input
494 # This can be used to randomize the update and gc operations a bit to avoid
495 # having them all end up all clustered together
496 sub rand_adjust {
497 my $input = shift || 0;
498 return $input unless $input;
499 return $input + int(rand(0.25 * $input));
502 # Open a pipe to a new sendmail process. The '-i' option is always passed to
503 # the new process followed by any addtional arguments passed in. Note that
504 # the sendmail process is only expected to understand the '-i', '-t' and '-f'
505 # options. Using any other options via this function is not guaranteed to work.
506 # A list of recipients may follow the options. Combining a list of recipients
507 # with the '-t' option is not recommended.
508 sub sendmail_pipe {
509 return undef unless @_;
510 die "\$Girocco::Config::sendmail_bin is unset or not executable!\n"
511 unless $Girocco::Config::sendmail_bin && -x $Girocco::Config::sendmail_bin;
512 my $result = open(my $pipe, '|-', $Girocco::Config::sendmail_bin, '-i', @_);
513 return $result ? $pipe : undef;
516 # Open a pipe that works similarly to a mailer such as /usr/bin/mail in that
517 # if the first argument is '-s', a subject line will be automatically added
518 # (using the second argument as the subject). Any remaining arguments are
519 # expected to be recipient addresses that will be added to an explicit To:
520 # line as well as passed on to sendmail_pipe. In addition an
521 # "Auto-Submitted: auto-generated" header is always added as well as a suitable
522 # "From:" header.
523 sub mailer_pipe {
524 my $subject = undef;
525 if (@_ >= 2 && $_[0] eq '-s') {
526 shift;
527 $subject = shift;
529 my $tolist = join(", ", @_);
530 unshift(@_, '-f', $Girocco::Config::sender) if $Girocco::Config::sender;
531 my $pipe = sendmail_pipe(@_);
532 if ($pipe) {
533 print $pipe "From: \"$Girocco::Config::name\" ",
534 "($Girocco::Config::title) ",
535 "<$Girocco::Config::admin>\n";
536 print $pipe "To: $tolist\n";
537 print $pipe "Subject: $subject\n" if defined($subject);
538 print $pipe "Auto-Submitted: auto-generated\n";
539 print $pipe "\n";
541 return $pipe;