]>
Commit | Line | Data |
---|---|---|
428c995c JN |
1 | #! /usr/bin/perl |
2 | ||
3 | # Copyright (C) 2011 | |
4 | # Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr> | |
5 | # Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr> | |
6 | # Claire Fousse <claire.fousse@ensimag.imag.fr> | |
7 | # David Amouyal <david.amouyal@ensimag.imag.fr> | |
8 | # Matthieu Moy <matthieu.moy@grenoble-inp.fr> | |
9 | # License: GPL v2 or later | |
10 | ||
11 | # Gateway between Git and MediaWiki. | |
9a7365cf | 12 | # Documentation & bugtracker: https://github.com/moy/Git-Mediawiki/ |
428c995c JN |
13 | |
14 | use strict; | |
15 | use MediaWiki::API; | |
da608b12 | 16 | use Git; |
192f7a08 BP |
17 | use Git::Mediawiki qw(clean_filename smudge_filename connect_maybe |
18 | EMPTY HTTP_CODE_OK); | |
428c995c | 19 | use DateTime::Format::ISO8601; |
668eec6f | 20 | use warnings; |
428c995c | 21 | |
721a533f | 22 | # By default, use UTF-8 to communicate with Git and the user |
86e95ef2 CM |
23 | binmode STDERR, ':encoding(UTF-8)'; |
24 | binmode STDOUT, ':encoding(UTF-8)'; | |
428c995c JN |
25 | |
26 | use URI::Escape; | |
428c995c | 27 | |
428c995c | 28 | # It's not always possible to delete pages (may require some |
2582ab18 | 29 | # privileges). Deleted pages are replaced with this content. |
428c995c JN |
30 | use constant DELETED_CONTENT => "[[Category:Deleted]]\n"; |
31 | ||
32 | # It's not possible to create empty pages. New empty files in Git are | |
33 | # sent with this content instead. | |
34 | use constant EMPTY_CONTENT => "<!-- empty page -->\n"; | |
35 | ||
36 | # used to reflect file creation or deletion in diff. | |
86e95ef2 | 37 | use constant NULL_SHA1 => '0000000000000000000000000000000000000000'; |
428c995c | 38 | |
dcb1ea62 MM |
39 | # Used on Git's side to reflect empty edit messages on the wiki |
40 | use constant EMPTY_MESSAGE => '*Empty MediaWiki Message*'; | |
41 | ||
fed56c06 CM |
42 | # Number of pages taken into account at once in submodule get_mw_page_list |
43 | use constant SLICE_SIZE => 50; | |
44 | ||
45 | # Number of linked mediafile to get at once in get_linked_mediafiles | |
46 | # The query is split in small batches because of the MW API limit of | |
47 | # the number of links to be returned (500 links max). | |
48 | use constant BATCH_SIZE => 10; | |
49 | ||
5ada8687 CM |
50 | if (@ARGV != 2) { |
51 | exit_error_usage(); | |
52 | } | |
53 | ||
428c995c JN |
54 | my $remotename = $ARGV[0]; |
55 | my $url = $ARGV[1]; | |
56 | ||
57 | # Accept both space-separated and multiple keys in config file. | |
58 | # Spaces should be written as _ anyway because we'll use chomp. | |
86e95ef2 | 59 | my @tracked_pages = split(/[ \n]/, run_git("config --get-all remote.${remotename}.pages")); |
428c995c JN |
60 | chomp(@tracked_pages); |
61 | ||
62 | # Just like @tracked_pages, but for MediaWiki categories. | |
86e95ef2 | 63 | my @tracked_categories = split(/[ \n]/, run_git("config --get-all remote.${remotename}.categories")); |
428c995c JN |
64 | chomp(@tracked_categories); |
65 | ||
5d9798ae K |
66 | # Just like @tracked_categories, but for MediaWiki namespaces. |
67 | my @tracked_namespaces = split(/[ \n]/, run_git("config --get-all remote.${remotename}.namespaces")); | |
cc923380 | 68 | for (@tracked_namespaces) { s/_/ /g; } |
5d9798ae K |
69 | chomp(@tracked_namespaces); |
70 | ||
2045e293 | 71 | # Import media files on pull |
86e95ef2 | 72 | my $import_media = run_git("config --get --bool remote.${remotename}.mediaimport"); |
6a9e55b0 | 73 | chomp($import_media); |
86e95ef2 | 74 | $import_media = ($import_media eq 'true'); |
6a9e55b0 | 75 | |
2045e293 | 76 | # Export media files on push |
86e95ef2 | 77 | my $export_media = run_git("config --get --bool remote.${remotename}.mediaexport"); |
2045e293 | 78 | chomp($export_media); |
86e95ef2 | 79 | $export_media = !($export_media eq 'false'); |
2045e293 | 80 | |
86e95ef2 | 81 | my $wiki_login = run_git("config --get remote.${remotename}.mwLogin"); |
03670c8b | 82 | # Note: mwPassword is discouraged. Use the credential system instead. |
86e95ef2 CM |
83 | my $wiki_passwd = run_git("config --get remote.${remotename}.mwPassword"); |
84 | my $wiki_domain = run_git("config --get remote.${remotename}.mwDomain"); | |
428c995c JN |
85 | chomp($wiki_login); |
86 | chomp($wiki_passwd); | |
1d6abac9 | 87 | chomp($wiki_domain); |
428c995c JN |
88 | |
89 | # Import only last revisions (both for clone and fetch) | |
86e95ef2 | 90 | my $shallow_import = run_git("config --get --bool remote.${remotename}.shallow"); |
428c995c | 91 | chomp($shallow_import); |
86e95ef2 | 92 | $shallow_import = ($shallow_import eq 'true'); |
428c995c | 93 | |
5a29217d MM |
94 | # Fetch (clone and pull) by revisions instead of by pages. This behavior |
95 | # is more efficient when we have a wiki with lots of pages and we fetch | |
96 | # the revisions quite often so that they concern only few pages. | |
97 | # Possible values: | |
98 | # - by_rev: perform one query per new revision on the remote wiki | |
99 | # - by_page: query each tracked page for new revision | |
86e95ef2 | 100 | my $fetch_strategy = run_git("config --get remote.${remotename}.fetchStrategy"); |
b8b4e1b3 | 101 | if (!$fetch_strategy) { |
86e95ef2 | 102 | $fetch_strategy = run_git('config --get mediawiki.fetchStrategy'); |
5a29217d MM |
103 | } |
104 | chomp($fetch_strategy); | |
b8b4e1b3 | 105 | if (!$fetch_strategy) { |
86e95ef2 | 106 | $fetch_strategy = 'by_page'; |
5a29217d MM |
107 | } |
108 | ||
0afd29e2 CM |
109 | # Remember the timestamp corresponding to a revision id. |
110 | my %basetimestamps; | |
111 | ||
93f0d338 MM |
112 | # Dumb push: don't update notes and mediawiki ref to reflect the last push. |
113 | # | |
114 | # Configurable with mediawiki.dumbPush, or per-remote with | |
115 | # remote.<remotename>.dumbPush. | |
116 | # | |
117 | # This means the user will have to re-import the just-pushed | |
118 | # revisions. On the other hand, this means that the Git revisions | |
119 | # corresponding to MediaWiki revisions are all imported from the wiki, | |
120 | # regardless of whether they were initially created in Git or from the | |
121 | # web interface, hence all users will get the same history (i.e. if | |
122 | # the push from Git to MediaWiki loses some information, everybody | |
123 | # will get the history with information lost). If the import is | |
124 | # deterministic, this means everybody gets the same sha1 for each | |
125 | # MediaWiki revision. | |
86e95ef2 | 126 | my $dumb_push = run_git("config --get --bool remote.${remotename}.dumbPush"); |
b8b4e1b3 | 127 | if (!$dumb_push) { |
86e95ef2 | 128 | $dumb_push = run_git('config --get --bool mediawiki.dumbPush'); |
93f0d338 MM |
129 | } |
130 | chomp($dumb_push); | |
86e95ef2 | 131 | $dumb_push = ($dumb_push eq 'true'); |
93f0d338 | 132 | |
428c995c | 133 | my $wiki_name = $url; |
857f21a3 | 134 | $wiki_name =~ s{[^/]*://}{}; |
9fb79503 MM |
135 | # If URL is like http://user:password@example.com/, we clearly don't |
136 | # want the password in $wiki_name. While we're there, also remove user | |
137 | # and '@' sign, to avoid author like MWUser@HTTPUser@host.com | |
138 | $wiki_name =~ s/^.*@//; | |
428c995c JN |
139 | |
140 | # Commands parser | |
428c995c JN |
141 | while (<STDIN>) { |
142 | chomp; | |
6a316bee CM |
143 | |
144 | if (!parse_command($_)) { | |
428c995c JN |
145 | last; |
146 | } | |
147 | ||
148 | BEGIN { $| = 1 } # flush STDOUT, to make sure the previous | |
149 | # command is fully processed. | |
150 | } | |
151 | ||
152 | ########################## Functions ############################## | |
153 | ||
5ada8687 CM |
154 | ## error handling |
155 | sub exit_error_usage { | |
156 | die "ERROR: git-remote-mediawiki module was not called with a correct number of\n" . | |
157 | "parameters\n" . | |
158 | "You may obtain this error because you attempted to run the git-remote-mediawiki\n" . | |
159 | "module directly.\n" . | |
160 | "This module can be used the following way:\n" . | |
161 | "\tgit clone mediawiki://<address of a mediawiki>\n" . | |
162 | "Then, use git commit, push and pull as with every normal git repository.\n"; | |
163 | } | |
164 | ||
6a316bee CM |
165 | sub parse_command { |
166 | my ($line) = @_; | |
167 | my @cmd = split(/ /, $line); | |
168 | if (!defined $cmd[0]) { | |
169 | return 0; | |
170 | } | |
86e95ef2 | 171 | if ($cmd[0] eq 'capabilities') { |
6a316bee CM |
172 | die("Too many arguments for capabilities\n") |
173 | if (defined($cmd[1])); | |
174 | mw_capabilities(); | |
86e95ef2 | 175 | } elsif ($cmd[0] eq 'list') { |
6a316bee CM |
176 | die("Too many arguments for list\n") if (defined($cmd[2])); |
177 | mw_list($cmd[1]); | |
86e95ef2 | 178 | } elsif ($cmd[0] eq 'import') { |
d8e7c67e CM |
179 | die("Invalid argument for import\n") |
180 | if ($cmd[1] eq EMPTY); | |
181 | die("Too many arguments for import\n") | |
182 | if (defined($cmd[2])); | |
6a316bee | 183 | mw_import($cmd[1]); |
86e95ef2 | 184 | } elsif ($cmd[0] eq 'option') { |
d8e7c67e CM |
185 | die("Invalid arguments for option\n") |
186 | if ($cmd[1] eq EMPTY || $cmd[2] eq EMPTY); | |
6a316bee | 187 | die("Too many arguments for option\n") |
d8e7c67e | 188 | if (defined($cmd[3])); |
6a316bee | 189 | mw_option($cmd[1],$cmd[2]); |
86e95ef2 | 190 | } elsif ($cmd[0] eq 'push') { |
6a316bee CM |
191 | mw_push($cmd[1]); |
192 | } else { | |
e83d36b6 | 193 | print {*STDERR} "Unknown command. Aborting...\n"; |
6a316bee CM |
194 | return 0; |
195 | } | |
196 | return 1; | |
197 | } | |
198 | ||
428c995c JN |
199 | # MediaWiki API instance, created lazily. |
200 | my $mediawiki; | |
201 | ||
0aa03212 MM |
202 | sub fatal_mw_error { |
203 | my $action = shift; | |
204 | print STDERR "fatal: could not $action.\n"; | |
205 | print STDERR "fatal: '$url' does not appear to be a mediawiki\n"; | |
206 | if ($url =~ /^https/) { | |
207 | print STDERR "fatal: make sure '$url/api.php' is a valid page\n"; | |
208 | print STDERR "fatal: and the SSL certificate is correct.\n"; | |
209 | } else { | |
210 | print STDERR "fatal: make sure '$url/api.php' is a valid page.\n"; | |
211 | } | |
212 | print STDERR "fatal: (error " . | |
213 | $mediawiki->{error}->{code} . ': ' . | |
214 | $mediawiki->{error}->{details} . ")\n"; | |
215 | exit 1; | |
216 | } | |
217 | ||
9cb74f36 MM |
218 | ## Functions for listing pages on the remote wiki |
219 | sub get_mw_tracked_pages { | |
220 | my $pages = shift; | |
6df7e0df | 221 | get_mw_page_list(\@tracked_pages, $pages); |
1aff8c62 | 222 | return; |
6df7e0df MM |
223 | } |
224 | ||
225 | sub get_mw_page_list { | |
226 | my $page_list = shift; | |
227 | my $pages = shift; | |
d49a0384 | 228 | my @some_pages = @{$page_list}; |
9cb74f36 | 229 | while (@some_pages) { |
fed56c06 | 230 | my $last_page = SLICE_SIZE; |
b835baf6 CM |
231 | if ($#some_pages < $last_page) { |
232 | $last_page = $#some_pages; | |
9cb74f36 | 233 | } |
b835baf6 | 234 | my @slice = @some_pages[0..$last_page]; |
9cb74f36 | 235 | get_mw_first_pages(\@slice, $pages); |
fed56c06 | 236 | @some_pages = @some_pages[(SLICE_SIZE + 1)..$#some_pages]; |
9cb74f36 | 237 | } |
1aff8c62 | 238 | return; |
9cb74f36 MM |
239 | } |
240 | ||
241 | sub get_mw_tracked_categories { | |
242 | my $pages = shift; | |
243 | foreach my $category (@tracked_categories) { | |
244 | if (index($category, ':') < 0) { | |
245 | # Mediawiki requires the Category | |
246 | # prefix, but let's not force the user | |
247 | # to specify it. | |
86e95ef2 | 248 | $category = "Category:${category}"; |
9cb74f36 MM |
249 | } |
250 | my $mw_pages = $mediawiki->list( { | |
251 | action => 'query', | |
252 | list => 'categorymembers', | |
253 | cmtitle => $category, | |
254 | cmlimit => 'max' } ) | |
255 | || die $mediawiki->{error}->{code} . ': ' | |
8a43b36a | 256 | . $mediawiki->{error}->{details} . "\n"; |
9cb74f36 MM |
257 | foreach my $page (@{$mw_pages}) { |
258 | $pages->{$page->{title}} = $page; | |
259 | } | |
260 | } | |
1aff8c62 | 261 | return; |
9cb74f36 MM |
262 | } |
263 | ||
5d9798ae K |
264 | sub get_mw_tracked_namespaces { |
265 | my $pages = shift; | |
55fefa9e | 266 | foreach my $local_namespace (sort @tracked_namespaces) { |
da2a1809 AB |
267 | my $namespace_id; |
268 | if ($local_namespace eq "(Main)") { | |
269 | $namespace_id = 0; | |
270 | } else { | |
271 | $namespace_id = get_mw_namespace_id($local_namespace); | |
272 | } | |
db336435 AB |
273 | # virtual namespaces don't support allpages |
274 | next if !defined($namespace_id) || $namespace_id < 0; | |
5d9798ae K |
275 | my $mw_pages = $mediawiki->list( { |
276 | action => 'query', | |
277 | list => 'allpages', | |
db336435 | 278 | apnamespace => $namespace_id, |
5d9798ae K |
279 | aplimit => 'max' } ) |
280 | || die $mediawiki->{error}->{code} . ': ' | |
281 | . $mediawiki->{error}->{details} . "\n"; | |
94c9acbf | 282 | print {*STDERR} "$#{$mw_pages} found in namespace $local_namespace ($namespace_id)\n"; |
5d9798ae K |
283 | foreach my $page (@{$mw_pages}) { |
284 | $pages->{$page->{title}} = $page; | |
285 | } | |
286 | } | |
287 | return; | |
288 | } | |
289 | ||
9cb74f36 MM |
290 | sub get_mw_all_pages { |
291 | my $pages = shift; | |
292 | # No user-provided list, get the list of pages from the API. | |
293 | my $mw_pages = $mediawiki->list({ | |
294 | action => 'query', | |
295 | list => 'allpages', | |
296 | aplimit => 'max' | |
297 | }); | |
298 | if (!defined($mw_pages)) { | |
0aa03212 | 299 | fatal_mw_error("get the list of wiki pages"); |
9cb74f36 MM |
300 | } |
301 | foreach my $page (@{$mw_pages}) { | |
302 | $pages->{$page->{title}} = $page; | |
303 | } | |
1aff8c62 | 304 | return; |
9cb74f36 MM |
305 | } |
306 | ||
307 | # queries the wiki for a set of pages. Meant to be used within a loop | |
308 | # querying the wiki for slices of page list. | |
428c995c JN |
309 | sub get_mw_first_pages { |
310 | my $some_pages = shift; | |
311 | my @some_pages = @{$some_pages}; | |
312 | ||
313 | my $pages = shift; | |
314 | ||
315 | # pattern 'page1|page2|...' required by the API | |
316 | my $titles = join('|', @some_pages); | |
317 | ||
318 | my $mw_pages = $mediawiki->api({ | |
319 | action => 'query', | |
320 | titles => $titles, | |
321 | }); | |
322 | if (!defined($mw_pages)) { | |
0aa03212 | 323 | fatal_mw_error("query the list of wiki pages"); |
428c995c JN |
324 | } |
325 | while (my ($id, $page) = each(%{$mw_pages->{query}->{pages}})) { | |
326 | if ($id < 0) { | |
e83d36b6 | 327 | print {*STDERR} "Warning: page $page->{title} not found on wiki\n"; |
428c995c JN |
328 | } else { |
329 | $pages->{$page->{title}} = $page; | |
330 | } | |
331 | } | |
1aff8c62 | 332 | return; |
428c995c JN |
333 | } |
334 | ||
9cb74f36 | 335 | # Get the list of pages to be fetched according to configuration. |
428c995c | 336 | sub get_mw_pages { |
192f7a08 | 337 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
428c995c | 338 | |
e83d36b6 | 339 | print {*STDERR} "Listing pages on remote wiki...\n"; |
f690ddfd | 340 | |
428c995c JN |
341 | my %pages; # hash on page titles to avoid duplicates |
342 | my $user_defined; | |
343 | if (@tracked_pages) { | |
344 | $user_defined = 1; | |
345 | # The user provided a list of pages titles, but we | |
346 | # still need to query the API to get the page IDs. | |
9cb74f36 | 347 | get_mw_tracked_pages(\%pages); |
428c995c JN |
348 | } |
349 | if (@tracked_categories) { | |
350 | $user_defined = 1; | |
9cb74f36 | 351 | get_mw_tracked_categories(\%pages); |
428c995c | 352 | } |
5d9798ae K |
353 | if (@tracked_namespaces) { |
354 | $user_defined = 1; | |
355 | get_mw_tracked_namespaces(\%pages); | |
356 | } | |
428c995c | 357 | if (!$user_defined) { |
9cb74f36 | 358 | get_mw_all_pages(\%pages); |
428c995c | 359 | } |
6a9e55b0 | 360 | if ($import_media) { |
e83d36b6 | 361 | print {*STDERR} "Getting media files for selected pages...\n"; |
6a9e55b0 PV |
362 | if ($user_defined) { |
363 | get_linked_mediafiles(\%pages); | |
364 | } else { | |
365 | get_all_mediafiles(\%pages); | |
428c995c JN |
366 | } |
367 | } | |
e83d36b6 | 368 | print {*STDERR} (scalar keys %pages) . " pages found.\n"; |
b1ede9a9 | 369 | return %pages; |
428c995c JN |
370 | } |
371 | ||
b3d98595 NKT |
372 | # usage: $out = run_git("command args"); |
373 | # $out = run_git("command args", "raw"); # don't interpret output as UTF-8. | |
428c995c | 374 | sub run_git { |
b3d98595 | 375 | my $args = shift; |
86e95ef2 CM |
376 | my $encoding = (shift || 'encoding(UTF-8)'); |
377 | open(my $git, "-|:${encoding}", "git ${args}") | |
378 | or die "Unable to fork: $!\n"; | |
4f1b7883 CM |
379 | my $res = do { |
380 | local $/ = undef; | |
381 | <$git> | |
382 | }; | |
428c995c JN |
383 | close($git); |
384 | ||
385 | return $res; | |
386 | } | |
387 | ||
388 | ||
6a9e55b0 PV |
389 | sub get_all_mediafiles { |
390 | my $pages = shift; | |
391 | # Attach list of all pages for media files from the API, | |
392 | # they are in a different namespace, only one namespace | |
393 | # can be queried at the same moment | |
394 | my $mw_pages = $mediawiki->list({ | |
395 | action => 'query', | |
396 | list => 'allpages', | |
86e95ef2 | 397 | apnamespace => get_mw_namespace_id('File'), |
6a9e55b0 PV |
398 | aplimit => 'max' |
399 | }); | |
400 | if (!defined($mw_pages)) { | |
e83d36b6 CM |
401 | print {*STDERR} "fatal: could not get the list of pages for media files.\n"; |
402 | print {*STDERR} "fatal: '$url' does not appear to be a mediawiki\n"; | |
403 | print {*STDERR} "fatal: make sure '$url/api.php' is a valid page.\n"; | |
6a9e55b0 PV |
404 | exit 1; |
405 | } | |
406 | foreach my $page (@{$mw_pages}) { | |
407 | $pages->{$page->{title}} = $page; | |
408 | } | |
1aff8c62 | 409 | return; |
6a9e55b0 PV |
410 | } |
411 | ||
412 | sub get_linked_mediafiles { | |
413 | my $pages = shift; | |
81f6a7a4 | 414 | my @titles = map { $_->{title} } values(%{$pages}); |
6a9e55b0 | 415 | |
fed56c06 | 416 | my $batch = BATCH_SIZE; |
6a9e55b0 PV |
417 | while (@titles) { |
418 | if ($#titles < $batch) { | |
419 | $batch = $#titles; | |
420 | } | |
421 | my @slice = @titles[0..$batch]; | |
422 | ||
423 | # pattern 'page1|page2|...' required by the API | |
424 | my $mw_titles = join('|', @slice); | |
425 | ||
426 | # Media files could be included or linked from | |
427 | # a page, get all related | |
428 | my $query = { | |
429 | action => 'query', | |
430 | prop => 'links|images', | |
431 | titles => $mw_titles, | |
86e95ef2 | 432 | plnamespace => get_mw_namespace_id('File'), |
6a9e55b0 PV |
433 | pllimit => 'max' |
434 | }; | |
435 | my $result = $mediawiki->api($query); | |
436 | ||
437 | while (my ($id, $page) = each(%{$result->{query}->{pages}})) { | |
6df7e0df | 438 | my @media_titles; |
6a9e55b0 | 439 | if (defined($page->{links})) { |
81f6a7a4 CM |
440 | my @link_titles |
441 | = map { $_->{title} } @{$page->{links}}; | |
6df7e0df | 442 | push(@media_titles, @link_titles); |
6a9e55b0 PV |
443 | } |
444 | if (defined($page->{images})) { | |
81f6a7a4 CM |
445 | my @image_titles |
446 | = map { $_->{title} } @{$page->{images}}; | |
6df7e0df | 447 | push(@media_titles, @image_titles); |
6a9e55b0 | 448 | } |
6df7e0df MM |
449 | if (@media_titles) { |
450 | get_mw_page_list(\@media_titles, $pages); | |
6a9e55b0 PV |
451 | } |
452 | } | |
453 | ||
454 | @titles = @titles[($batch+1)..$#titles]; | |
455 | } | |
1aff8c62 | 456 | return; |
6a9e55b0 PV |
457 | } |
458 | ||
459 | sub get_mw_mediafile_for_page_revision { | |
460 | # Name of the file on Wiki, with the prefix. | |
6df7e0df | 461 | my $filename = shift; |
6a9e55b0 PV |
462 | my $timestamp = shift; |
463 | my %mediafile; | |
464 | ||
6df7e0df MM |
465 | # Search if on a media file with given timestamp exists on |
466 | # MediaWiki. In that case download the file. | |
6a9e55b0 PV |
467 | my $query = { |
468 | action => 'query', | |
469 | prop => 'imageinfo', | |
86e95ef2 | 470 | titles => "File:${filename}", |
6a9e55b0 PV |
471 | iistart => $timestamp, |
472 | iiend => $timestamp, | |
473 | iiprop => 'timestamp|archivename|url', | |
474 | iilimit => 1 | |
475 | }; | |
476 | my $result = $mediawiki->api($query); | |
477 | ||
6df7e0df | 478 | my ($fileid, $file) = each( %{$result->{query}->{pages}} ); |
6a9e55b0 PV |
479 | # If not defined it means there is no revision of the file for |
480 | # given timestamp. | |
481 | if (defined($file->{imageinfo})) { | |
6a9e55b0 PV |
482 | $mediafile{title} = $filename; |
483 | ||
484 | my $fileinfo = pop(@{$file->{imageinfo}}); | |
485 | $mediafile{timestamp} = $fileinfo->{timestamp}; | |
6df7e0df MM |
486 | # Mediawiki::API's download function doesn't support https URLs |
487 | # and can't download old versions of files. | |
e83d36b6 | 488 | print {*STDERR} "\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n"; |
6df7e0df | 489 | $mediafile{content} = download_mw_mediafile($fileinfo->{url}); |
6a9e55b0 PV |
490 | } |
491 | return %mediafile; | |
492 | } | |
493 | ||
6df7e0df | 494 | sub download_mw_mediafile { |
8f04f7dd | 495 | my $download_url = shift; |
6a9e55b0 | 496 | |
8f04f7dd | 497 | my $response = $mediawiki->{ua}->get($download_url); |
fed56c06 | 498 | if ($response->code == HTTP_CODE_OK) { |
9742fb7e MM |
499 | # It is tempting to return |
500 | # $response->decoded_content({charset => "none"}), but | |
501 | # when doing so, utf8::downgrade($content) fails with | |
502 | # "Wide character in subroutine entry". | |
503 | $response->decode(); | |
504 | return $response->content(); | |
6a9e55b0 | 505 | } else { |
e83d36b6 CM |
506 | print {*STDERR} "Error downloading mediafile from :\n"; |
507 | print {*STDERR} "URL: ${download_url}\n"; | |
508 | print {*STDERR} 'Server response: ' . $response->code . q{ } . $response->message . "\n"; | |
6a9e55b0 | 509 | exit 1; |
6a9e55b0 PV |
510 | } |
511 | } | |
512 | ||
428c995c JN |
513 | sub get_last_local_revision { |
514 | # Get note regarding last mediawiki revision | |
86e95ef2 | 515 | my $note = run_git("notes --ref=${remotename}/mediawiki show refs/mediawiki/${remotename}/master 2>/dev/null"); |
428c995c JN |
516 | my @note_info = split(/ /, $note); |
517 | ||
518 | my $lastrevision_number; | |
86e95ef2 | 519 | if (!(defined($note_info[0]) && $note_info[0] eq 'mediawiki_revision:')) { |
e83d36b6 | 520 | print {*STDERR} 'No previous mediawiki revision found'; |
428c995c JN |
521 | $lastrevision_number = 0; |
522 | } else { | |
523 | # Notes are formatted : mediawiki_revision: #number | |
524 | $lastrevision_number = $note_info[1]; | |
525 | chomp($lastrevision_number); | |
e83d36b6 | 526 | print {*STDERR} "Last local mediawiki revision found is ${lastrevision_number}"; |
428c995c JN |
527 | } |
528 | return $lastrevision_number; | |
529 | } | |
530 | ||
5a29217d MM |
531 | # Get the last remote revision without taking in account which pages are |
532 | # tracked or not. This function makes a single request to the wiki thus | |
533 | # avoid a loop onto all tracked pages. This is useful for the fetch-by-rev | |
534 | # option. | |
535 | sub get_last_global_remote_rev { | |
192f7a08 | 536 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
5a29217d MM |
537 | |
538 | my $query = { | |
539 | action => 'query', | |
540 | list => 'recentchanges', | |
541 | prop => 'revisions', | |
542 | rclimit => '1', | |
543 | rcdir => 'older', | |
544 | }; | |
545 | my $result = $mediawiki->api($query); | |
546 | return $result->{query}->{recentchanges}[0]->{revid}; | |
547 | } | |
548 | ||
549 | # Get the last remote revision concerning the tracked pages and the tracked | |
550 | # categories. | |
428c995c | 551 | sub get_last_remote_revision { |
192f7a08 | 552 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
428c995c | 553 | |
b1ede9a9 MM |
554 | my %pages_hash = get_mw_pages(); |
555 | my @pages = values(%pages_hash); | |
428c995c JN |
556 | |
557 | my $max_rev_num = 0; | |
558 | ||
e83d36b6 | 559 | print {*STDERR} "Getting last revision id on tracked pages...\n"; |
a393f488 | 560 | |
428c995c JN |
561 | foreach my $page (@pages) { |
562 | my $id = $page->{pageid}; | |
563 | ||
564 | my $query = { | |
565 | action => 'query', | |
566 | prop => 'revisions', | |
3c1ed90e | 567 | rvprop => 'ids|timestamp', |
428c995c JN |
568 | pageids => $id, |
569 | }; | |
570 | ||
571 | my $result = $mediawiki->api($query); | |
572 | ||
573 | my $lastrev = pop(@{$result->{query}->{pages}->{$id}->{revisions}}); | |
574 | ||
3c1ed90e MM |
575 | $basetimestamps{$lastrev->{revid}} = $lastrev->{timestamp}; |
576 | ||
428c995c JN |
577 | $max_rev_num = ($lastrev->{revid} > $max_rev_num ? $lastrev->{revid} : $max_rev_num); |
578 | } | |
579 | ||
e83d36b6 | 580 | print {*STDERR} "Last remote revision found is $max_rev_num.\n"; |
428c995c JN |
581 | return $max_rev_num; |
582 | } | |
583 | ||
584 | # Clean content before sending it to MediaWiki | |
585 | sub mediawiki_clean { | |
586 | my $string = shift; | |
587 | my $page_created = shift; | |
588 | # Mediawiki does not allow blank space at the end of a page and ends with a single \n. | |
589 | # This function right trims a string and adds a \n at the end to follow this rule | |
590 | $string =~ s/\s+$//; | |
aeb95eea | 591 | if ($string eq EMPTY && $page_created) { |
428c995c JN |
592 | # Creating empty pages is forbidden. |
593 | $string = EMPTY_CONTENT; | |
594 | } | |
595 | return $string."\n"; | |
596 | } | |
597 | ||
598 | # Filter applied on MediaWiki data before adding them to Git | |
599 | sub mediawiki_smudge { | |
600 | my $string = shift; | |
601 | if ($string eq EMPTY_CONTENT) { | |
aeb95eea | 602 | $string = EMPTY; |
428c995c JN |
603 | } |
604 | # This \n is important. This is due to mediawiki's way to handle end of files. | |
86e95ef2 | 605 | return "${string}\n"; |
428c995c JN |
606 | } |
607 | ||
428c995c JN |
608 | sub literal_data { |
609 | my ($content) = @_; | |
e83d36b6 | 610 | print {*STDOUT} 'data ', bytes::length($content), "\n", $content; |
1aff8c62 | 611 | return; |
428c995c JN |
612 | } |
613 | ||
6a9e55b0 PV |
614 | sub literal_data_raw { |
615 | # Output possibly binary content. | |
616 | my ($content) = @_; | |
617 | # Avoid confusion between size in bytes and in characters | |
618 | utf8::downgrade($content); | |
0e7c41c0 | 619 | binmode STDOUT, ':raw'; |
e83d36b6 | 620 | print {*STDOUT} 'data ', bytes::length($content), "\n", $content; |
0e7c41c0 | 621 | binmode STDOUT, ':encoding(UTF-8)'; |
1aff8c62 | 622 | return; |
6a9e55b0 PV |
623 | } |
624 | ||
428c995c JN |
625 | sub mw_capabilities { |
626 | # Revisions are imported to the private namespace | |
627 | # refs/mediawiki/$remotename/ by the helper and fetched into | |
628 | # refs/remotes/$remotename later by fetch. | |
e83d36b6 CM |
629 | print {*STDOUT} "refspec refs/heads/*:refs/mediawiki/${remotename}/*\n"; |
630 | print {*STDOUT} "import\n"; | |
631 | print {*STDOUT} "list\n"; | |
632 | print {*STDOUT} "push\n"; | |
aa38dc68 MM |
633 | if ($dumb_push) { |
634 | print {*STDOUT} "no-private-update\n"; | |
635 | } | |
e83d36b6 | 636 | print {*STDOUT} "\n"; |
1aff8c62 | 637 | return; |
428c995c JN |
638 | } |
639 | ||
640 | sub mw_list { | |
641 | # MediaWiki do not have branches, we consider one branch arbitrarily | |
642 | # called master, and HEAD pointing to it. | |
e83d36b6 CM |
643 | print {*STDOUT} "? refs/heads/master\n"; |
644 | print {*STDOUT} "\@refs/heads/master HEAD\n"; | |
645 | print {*STDOUT} "\n"; | |
1aff8c62 | 646 | return; |
428c995c JN |
647 | } |
648 | ||
649 | sub mw_option { | |
e83d36b6 CM |
650 | print {*STDERR} "remote-helper command 'option $_[0]' not yet implemented\n"; |
651 | print {*STDOUT} "unsupported\n"; | |
1aff8c62 | 652 | return; |
428c995c JN |
653 | } |
654 | ||
655 | sub fetch_mw_revisions_for_page { | |
656 | my $page = shift; | |
657 | my $id = shift; | |
658 | my $fetch_from = shift; | |
659 | my @page_revs = (); | |
660 | my $query = { | |
661 | action => 'query', | |
662 | prop => 'revisions', | |
663 | rvprop => 'ids', | |
664 | rvdir => 'newer', | |
665 | rvstartid => $fetch_from, | |
666 | rvlimit => 500, | |
667 | pageids => $id, | |
1d905f74 BP |
668 | |
669 | # Let MediaWiki know that we support the latest API. | |
670 | continue => '', | |
428c995c JN |
671 | }; |
672 | ||
673 | my $revnum = 0; | |
674 | # Get 500 revisions at a time due to the mediawiki api limit | |
675 | while (1) { | |
676 | my $result = $mediawiki->api($query); | |
677 | ||
678 | # Parse each of those 500 revisions | |
679 | foreach my $revision (@{$result->{query}->{pages}->{$id}->{revisions}}) { | |
680 | my $page_rev_ids; | |
681 | $page_rev_ids->{pageid} = $page->{pageid}; | |
682 | $page_rev_ids->{revid} = $revision->{revid}; | |
683 | push(@page_revs, $page_rev_ids); | |
684 | $revnum++; | |
685 | } | |
1d905f74 BP |
686 | |
687 | if ($result->{'query-continue'}) { # For legacy APIs | |
688 | $query->{rvstartid} = $result->{'query-continue'}->{revisions}->{rvstartid}; | |
689 | } elsif ($result->{continue}) { # For newer APIs | |
690 | $query->{rvstartid} = $result->{continue}->{rvcontinue}; | |
691 | $query->{continue} = $result->{continue}->{continue}; | |
692 | } else { | |
693 | last; | |
694 | } | |
428c995c JN |
695 | } |
696 | if ($shallow_import && @page_revs) { | |
e83d36b6 | 697 | print {*STDERR} " Found 1 revision (shallow import).\n"; |
428c995c JN |
698 | @page_revs = sort {$b->{revid} <=> $a->{revid}} (@page_revs); |
699 | return $page_revs[0]; | |
700 | } | |
e83d36b6 | 701 | print {*STDERR} " Found ${revnum} revision(s).\n"; |
428c995c JN |
702 | return @page_revs; |
703 | } | |
704 | ||
705 | sub fetch_mw_revisions { | |
706 | my $pages = shift; my @pages = @{$pages}; | |
707 | my $fetch_from = shift; | |
708 | ||
709 | my @revisions = (); | |
710 | my $n = 1; | |
711 | foreach my $page (@pages) { | |
712 | my $id = $page->{pageid}; | |
e83d36b6 | 713 | print {*STDERR} "page ${n}/", scalar(@pages), ': ', $page->{title}, "\n"; |
428c995c JN |
714 | $n++; |
715 | my @page_revs = fetch_mw_revisions_for_page($page, $id, $fetch_from); | |
716 | @revisions = (@page_revs, @revisions); | |
717 | } | |
718 | ||
719 | return ($n, @revisions); | |
720 | } | |
721 | ||
462d97da MM |
722 | sub fe_escape_path { |
723 | my $path = shift; | |
724 | $path =~ s/\\/\\\\/g; | |
725 | $path =~ s/"/\\"/g; | |
726 | $path =~ s/\n/\\n/g; | |
86e95ef2 | 727 | return qq("${path}"); |
462d97da MM |
728 | } |
729 | ||
428c995c JN |
730 | sub import_file_revision { |
731 | my $commit = shift; | |
732 | my %commit = %{$commit}; | |
733 | my $full_import = shift; | |
734 | my $n = shift; | |
6a9e55b0 PV |
735 | my $mediafile = shift; |
736 | my %mediafile; | |
737 | if ($mediafile) { | |
738 | %mediafile = %{$mediafile}; | |
739 | } | |
428c995c JN |
740 | |
741 | my $title = $commit{title}; | |
742 | my $comment = $commit{comment}; | |
743 | my $content = $commit{content}; | |
744 | my $author = $commit{author}; | |
745 | my $date = $commit{date}; | |
746 | ||
e83d36b6 CM |
747 | print {*STDOUT} "commit refs/mediawiki/${remotename}/master\n"; |
748 | print {*STDOUT} "mark :${n}\n"; | |
749 | print {*STDOUT} "committer ${author} <${author}\@${wiki_name}> " . $date->epoch . " +0000\n"; | |
428c995c JN |
750 | literal_data($comment); |
751 | ||
752 | # If it's not a clone, we need to know where to start from | |
753 | if (!$full_import && $n == 1) { | |
e83d36b6 | 754 | print {*STDOUT} "from refs/mediawiki/${remotename}/master^0\n"; |
428c995c JN |
755 | } |
756 | if ($content ne DELETED_CONTENT) { | |
e83d36b6 | 757 | print {*STDOUT} 'M 644 inline ' . |
86e95ef2 | 758 | fe_escape_path("${title}.mw") . "\n"; |
428c995c | 759 | literal_data($content); |
6a9e55b0 | 760 | if (%mediafile) { |
e83d36b6 | 761 | print {*STDOUT} 'M 644 inline ' |
462d97da | 762 | . fe_escape_path($mediafile{title}) . "\n"; |
6a9e55b0 PV |
763 | literal_data_raw($mediafile{content}); |
764 | } | |
e83d36b6 | 765 | print {*STDOUT} "\n\n"; |
428c995c | 766 | } else { |
e83d36b6 | 767 | print {*STDOUT} 'D ' . fe_escape_path("${title}.mw") . "\n"; |
428c995c JN |
768 | } |
769 | ||
770 | # mediawiki revision number in the git note | |
771 | if ($full_import && $n == 1) { | |
e83d36b6 | 772 | print {*STDOUT} "reset refs/notes/${remotename}/mediawiki\n"; |
428c995c | 773 | } |
e83d36b6 CM |
774 | print {*STDOUT} "commit refs/notes/${remotename}/mediawiki\n"; |
775 | print {*STDOUT} "committer ${author} <${author}\@${wiki_name}> " . $date->epoch . " +0000\n"; | |
86e95ef2 | 776 | literal_data('Note added by git-mediawiki during import'); |
428c995c | 777 | if (!$full_import && $n == 1) { |
e83d36b6 | 778 | print {*STDOUT} "from refs/notes/${remotename}/mediawiki^0\n"; |
428c995c | 779 | } |
e83d36b6 | 780 | print {*STDOUT} "N inline :${n}\n"; |
86e95ef2 | 781 | literal_data("mediawiki_revision: $commit{mw_revision}"); |
e83d36b6 | 782 | print {*STDOUT} "\n\n"; |
1aff8c62 | 783 | return; |
428c995c JN |
784 | } |
785 | ||
786 | # parse a sequence of | |
787 | # <cmd> <arg1> | |
788 | # <cmd> <arg2> | |
789 | # \n | |
790 | # (like batch sequence of import and sequence of push statements) | |
791 | sub get_more_refs { | |
792 | my $cmd = shift; | |
793 | my @refs; | |
794 | while (1) { | |
795 | my $line = <STDIN>; | |
eb96b750 | 796 | if ($line =~ /^$cmd (.*)$/) { |
428c995c JN |
797 | push(@refs, $1); |
798 | } elsif ($line eq "\n") { | |
799 | return @refs; | |
800 | } else { | |
8a43b36a | 801 | die("Invalid command in a '$cmd' batch: $_\n"); |
428c995c JN |
802 | } |
803 | } | |
1aff8c62 | 804 | return; |
428c995c JN |
805 | } |
806 | ||
807 | sub mw_import { | |
808 | # multiple import commands can follow each other. | |
86e95ef2 | 809 | my @refs = (shift, get_more_refs('import')); |
428c995c JN |
810 | foreach my $ref (@refs) { |
811 | mw_import_ref($ref); | |
812 | } | |
e83d36b6 | 813 | print {*STDOUT} "done\n"; |
1aff8c62 | 814 | return; |
428c995c JN |
815 | } |
816 | ||
817 | sub mw_import_ref { | |
818 | my $ref = shift; | |
819 | # The remote helper will call "import HEAD" and | |
820 | # "import refs/heads/master". | |
821 | # Since HEAD is a symbolic ref to master (by convention, | |
822 | # followed by the output of the command "list" that we gave), | |
823 | # we don't need to do anything in this case. | |
86e95ef2 | 824 | if ($ref eq 'HEAD') { |
428c995c JN |
825 | return; |
826 | } | |
827 | ||
192f7a08 | 828 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
428c995c | 829 | |
e83d36b6 | 830 | print {*STDERR} "Searching revisions...\n"; |
428c995c JN |
831 | my $last_local = get_last_local_revision(); |
832 | my $fetch_from = $last_local + 1; | |
833 | if ($fetch_from == 1) { | |
e83d36b6 | 834 | print {*STDERR} ", fetching from beginning.\n"; |
428c995c | 835 | } else { |
e83d36b6 | 836 | print {*STDERR} ", fetching from here.\n"; |
428c995c | 837 | } |
428c995c | 838 | |
5a29217d | 839 | my $n = 0; |
86e95ef2 | 840 | if ($fetch_strategy eq 'by_rev') { |
e83d36b6 | 841 | print {*STDERR} "Fetching & writing export data by revs...\n"; |
5a29217d | 842 | $n = mw_import_ref_by_revs($fetch_from); |
86e95ef2 | 843 | } elsif ($fetch_strategy eq 'by_page') { |
e83d36b6 | 844 | print {*STDERR} "Fetching & writing export data by pages...\n"; |
5a29217d MM |
845 | $n = mw_import_ref_by_pages($fetch_from); |
846 | } else { | |
e83d36b6 CM |
847 | print {*STDERR} qq(fatal: invalid fetch strategy "${fetch_strategy}".\n); |
848 | print {*STDERR} "Check your configuration variables remote.${remotename}.fetchStrategy and mediawiki.fetchStrategy\n"; | |
5a29217d MM |
849 | exit 1; |
850 | } | |
851 | ||
852 | if ($fetch_from == 1 && $n == 0) { | |
e83d36b6 | 853 | print {*STDERR} "You appear to have cloned an empty MediaWiki.\n"; |
5a29217d | 854 | # Something has to be done remote-helper side. If nothing is done, an error is |
2582ab18 | 855 | # thrown saying that HEAD is referring to unknown object 0000000000000000000 |
5a29217d MM |
856 | # and the clone fails. |
857 | } | |
1aff8c62 | 858 | return; |
5a29217d MM |
859 | } |
860 | ||
861 | sub mw_import_ref_by_pages { | |
862 | ||
863 | my $fetch_from = shift; | |
864 | my %pages_hash = get_mw_pages(); | |
865 | my @pages = values(%pages_hash); | |
866 | ||
428c995c JN |
867 | my ($n, @revisions) = fetch_mw_revisions(\@pages, $fetch_from); |
868 | ||
4465b6d6 | 869 | @revisions = sort {$a->{revid} <=> $b->{revid}} @revisions; |
81f6a7a4 | 870 | my @revision_ids = map { $_->{revid} } @revisions; |
4465b6d6 | 871 | |
312fa9aa MM |
872 | return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash); |
873 | } | |
874 | ||
5a29217d MM |
875 | sub mw_import_ref_by_revs { |
876 | ||
877 | my $fetch_from = shift; | |
878 | my %pages_hash = get_mw_pages(); | |
428c995c | 879 | |
5a29217d MM |
880 | my $last_remote = get_last_global_remote_rev(); |
881 | my @revision_ids = $fetch_from..$last_remote; | |
882 | return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash); | |
883 | } | |
884 | ||
885 | # Import revisions given in second argument (array of integers). | |
886 | # Only pages appearing in the third argument (hash indexed by page titles) | |
887 | # will be imported. | |
312fa9aa MM |
888 | sub mw_import_revids { |
889 | my $fetch_from = shift; | |
890 | my $revision_ids = shift; | |
891 | my $pages = shift; | |
892 | ||
893 | my $n = 0; | |
5a29217d | 894 | my $n_actual = 0; |
64127575 | 895 | my $last_timestamp = 0; # Placeholder in case $rev->timestamp is undefined |
428c995c | 896 | |
d49a0384 | 897 | foreach my $pagerevid (@{$revision_ids}) { |
ebd5fe1c MM |
898 | # Count page even if we skip it, since we display |
899 | # $n/$total and $total includes skipped pages. | |
900 | $n++; | |
901 | ||
428c995c JN |
902 | # fetch the content of the pages |
903 | my $query = { | |
904 | action => 'query', | |
905 | prop => 'revisions', | |
906 | rvprop => 'content|timestamp|comment|user|ids', | |
4465b6d6 | 907 | revids => $pagerevid, |
428c995c JN |
908 | }; |
909 | ||
910 | my $result = $mediawiki->api($query); | |
911 | ||
ff0bfd75 | 912 | if (!$result) { |
8a43b36a | 913 | die "Failed to retrieve modified page for revision $pagerevid\n"; |
ff0bfd75 MM |
914 | } |
915 | ||
ebd5fe1c MM |
916 | if (defined($result->{query}->{badrevids}->{$pagerevid})) { |
917 | # The revision id does not exist on the remote wiki. | |
918 | next; | |
919 | } | |
920 | ||
ff0bfd75 | 921 | if (!defined($result->{query}->{pages})) { |
86e95ef2 | 922 | die "Invalid revision ${pagerevid}.\n"; |
ff0bfd75 MM |
923 | } |
924 | ||
4465b6d6 MM |
925 | my @result_pages = values(%{$result->{query}->{pages}}); |
926 | my $result_page = $result_pages[0]; | |
927 | my $rev = $result_pages[0]->{revisions}->[0]; | |
428c995c | 928 | |
4465b6d6 | 929 | my $page_title = $result_page->{title}; |
5a29217d MM |
930 | |
931 | if (!exists($pages->{$page_title})) { | |
d49a0384 | 932 | print {*STDERR} "${n}/", scalar(@{$revision_ids}), |
86e95ef2 | 933 | ": Skipping revision #$rev->{revid} of ${page_title}\n"; |
5a29217d MM |
934 | next; |
935 | } | |
936 | ||
937 | $n_actual++; | |
938 | ||
428c995c JN |
939 | my %commit; |
940 | $commit{author} = $rev->{user} || 'Anonymous'; | |
dcb1ea62 | 941 | $commit{comment} = $rev->{comment} || EMPTY_MESSAGE; |
192f7a08 | 942 | $commit{title} = smudge_filename($page_title); |
4465b6d6 | 943 | $commit{mw_revision} = $rev->{revid}; |
428c995c JN |
944 | $commit{content} = mediawiki_smudge($rev->{'*'}); |
945 | ||
946 | if (!defined($rev->{timestamp})) { | |
947 | $last_timestamp++; | |
948 | } else { | |
949 | $last_timestamp = $rev->{timestamp}; | |
950 | } | |
951 | $commit{date} = DateTime::Format::ISO8601->parse_datetime($last_timestamp); | |
952 | ||
6a9e55b0 | 953 | # Differentiates classic pages and media files. |
6df7e0df | 954 | my ($namespace, $filename) = $page_title =~ /^([^:]*):(.*)$/; |
6a9e55b0 | 955 | my %mediafile; |
2272400c MM |
956 | if ($namespace) { |
957 | my $id = get_mw_namespace_id($namespace); | |
86e95ef2 | 958 | if ($id && $id == get_mw_namespace_id('File')) { |
2272400c MM |
959 | %mediafile = get_mw_mediafile_for_page_revision($filename, $rev->{timestamp}); |
960 | } | |
6a9e55b0 PV |
961 | } |
962 | # If this is a revision of the media page for new version | |
963 | # of a file do one common commit for both file and media page. | |
964 | # Else do commit only for that page. | |
d49a0384 | 965 | print {*STDERR} "${n}/", scalar(@{$revision_ids}), ": Revision #$rev->{revid} of $commit{title}\n"; |
5a29217d | 966 | import_file_revision(\%commit, ($fetch_from == 1), $n_actual, \%mediafile); |
428c995c JN |
967 | } |
968 | ||
5a29217d | 969 | return $n_actual; |
428c995c JN |
970 | } |
971 | ||
972 | sub error_non_fast_forward { | |
86e95ef2 | 973 | my $advice = run_git('config --bool advice.pushNonFastForward'); |
fd47d7b9 | 974 | chomp($advice); |
86e95ef2 | 975 | if ($advice ne 'false') { |
fd47d7b9 MM |
976 | # Native git-push would show this after the summary. |
977 | # We can't ask it to display it cleanly, so print it | |
978 | # ourselves before. | |
e83d36b6 CM |
979 | print {*STDERR} "To prevent you from losing history, non-fast-forward updates were rejected\n"; |
980 | print {*STDERR} "Merge the remote changes (e.g. 'git pull') before pushing again. See the\n"; | |
981 | print {*STDERR} "'Note about fast-forwards' section of 'git push --help' for details.\n"; | |
fd47d7b9 | 982 | } |
e83d36b6 | 983 | print {*STDOUT} qq(error $_[0] "non-fast-forward"\n); |
428c995c JN |
984 | return 0; |
985 | } | |
986 | ||
b3d98595 NKT |
987 | sub mw_upload_file { |
988 | my $complete_file_name = shift; | |
989 | my $new_sha1 = shift; | |
990 | my $extension = shift; | |
991 | my $file_deleted = shift; | |
992 | my $summary = shift; | |
993 | my $newrevid; | |
86e95ef2 | 994 | my $path = "File:${complete_file_name}"; |
b3d98595 NKT |
995 | my %hashFiles = get_allowed_file_extensions(); |
996 | if (!exists($hashFiles{$extension})) { | |
e83d36b6 CM |
997 | print {*STDERR} "${complete_file_name} is not a permitted file on this wiki.\n"; |
998 | print {*STDERR} "Check the configuration of file uploads in your mediawiki.\n"; | |
b3d98595 NKT |
999 | return $newrevid; |
1000 | } | |
2e3a16b2 | 1001 | # Deleting and uploading a file requires a privileged user |
b3d98595 | 1002 | if ($file_deleted) { |
192f7a08 | 1003 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
b3d98595 NKT |
1004 | my $query = { |
1005 | action => 'delete', | |
1006 | title => $path, | |
1007 | reason => $summary | |
1008 | }; | |
1009 | if (!$mediawiki->edit($query)) { | |
e83d36b6 CM |
1010 | print {*STDERR} "Failed to delete file on remote wiki\n"; |
1011 | print {*STDERR} "Check your permissions on the remote site. Error code:\n"; | |
1012 | print {*STDERR} $mediawiki->{error}->{code} . ':' . $mediawiki->{error}->{details}; | |
b3d98595 NKT |
1013 | exit 1; |
1014 | } | |
1015 | } else { | |
1016 | # Don't let perl try to interpret file content as UTF-8 => use "raw" | |
86e95ef2 | 1017 | my $content = run_git("cat-file blob ${new_sha1}", 'raw'); |
aeb95eea | 1018 | if ($content ne EMPTY) { |
192f7a08 | 1019 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
b3d98595 | 1020 | $mediawiki->{config}->{upload_url} = |
86e95ef2 | 1021 | "${url}/index.php/Special:Upload"; |
b3d98595 NKT |
1022 | $mediawiki->edit({ |
1023 | action => 'upload', | |
1024 | filename => $complete_file_name, | |
1025 | comment => $summary, | |
1026 | file => [undef, | |
1027 | $complete_file_name, | |
1028 | Content => $content], | |
1029 | ignorewarnings => 1, | |
1030 | }, { | |
1031 | skip_encoding => 1 | |
1032 | } ) || die $mediawiki->{error}->{code} . ':' | |
8a43b36a | 1033 | . $mediawiki->{error}->{details} . "\n"; |
b3d98595 NKT |
1034 | my $last_file_page = $mediawiki->get_page({title => $path}); |
1035 | $newrevid = $last_file_page->{revid}; | |
e83d36b6 | 1036 | print {*STDERR} "Pushed file: ${new_sha1} - ${complete_file_name}.\n"; |
b3d98595 | 1037 | } else { |
e83d36b6 | 1038 | print {*STDERR} "Empty file ${complete_file_name} not pushed.\n"; |
b3d98595 NKT |
1039 | } |
1040 | } | |
1041 | return $newrevid; | |
1042 | } | |
1043 | ||
428c995c JN |
1044 | sub mw_push_file { |
1045 | my $diff_info = shift; | |
1046 | # $diff_info contains a string in this format: | |
1047 | # 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status> | |
1048 | my @diff_info_split = split(/[ \t]/, $diff_info); | |
1049 | ||
1050 | # Filename, including .mw extension | |
1051 | my $complete_file_name = shift; | |
1052 | # Commit message | |
1053 | my $summary = shift; | |
93f0d338 MM |
1054 | # MediaWiki revision number. Keep the previous one by default, |
1055 | # in case there's no edit to perform. | |
b3d98595 NKT |
1056 | my $oldrevid = shift; |
1057 | my $newrevid; | |
428c995c | 1058 | |
dcb1ea62 | 1059 | if ($summary eq EMPTY_MESSAGE) { |
aeb95eea | 1060 | $summary = EMPTY; |
dcb1ea62 MM |
1061 | } |
1062 | ||
428c995c JN |
1063 | my $new_sha1 = $diff_info_split[3]; |
1064 | my $old_sha1 = $diff_info_split[2]; | |
1065 | my $page_created = ($old_sha1 eq NULL_SHA1); | |
1066 | my $page_deleted = ($new_sha1 eq NULL_SHA1); | |
192f7a08 | 1067 | $complete_file_name = clean_filename($complete_file_name); |
428c995c | 1068 | |
b3d98595 NKT |
1069 | my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/; |
1070 | if (!defined($extension)) { | |
aeb95eea | 1071 | $extension = EMPTY; |
b3d98595 | 1072 | } |
86e95ef2 | 1073 | if ($extension eq 'mw') { |
2045e293 | 1074 | my $ns = get_mw_namespace_id_for_page($complete_file_name); |
86e95ef2 | 1075 | if ($ns && $ns == get_mw_namespace_id('File') && (!$export_media)) { |
e83d36b6 | 1076 | print {*STDERR} "Ignoring media file related page: ${complete_file_name}\n"; |
86e95ef2 | 1077 | return ($oldrevid, 'ok'); |
2045e293 | 1078 | } |
428c995c JN |
1079 | my $file_content; |
1080 | if ($page_deleted) { | |
1081 | # Deleting a page usually requires | |
2582ab18 | 1082 | # special privileges. A common |
428c995c JN |
1083 | # convention is to replace the page |
1084 | # with this content instead: | |
1085 | $file_content = DELETED_CONTENT; | |
1086 | } else { | |
86e95ef2 | 1087 | $file_content = run_git("cat-file blob ${new_sha1}"); |
428c995c JN |
1088 | } |
1089 | ||
192f7a08 | 1090 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
428c995c JN |
1091 | |
1092 | my $result = $mediawiki->edit( { | |
1093 | action => 'edit', | |
1094 | summary => $summary, | |
1095 | title => $title, | |
b3d98595 | 1096 | basetimestamp => $basetimestamps{$oldrevid}, |
428c995c JN |
1097 | text => mediawiki_clean($file_content, $page_created), |
1098 | }, { | |
1099 | skip_encoding => 1 # Helps with names with accentuated characters | |
3c1ed90e MM |
1100 | }); |
1101 | if (!$result) { | |
1102 | if ($mediawiki->{error}->{code} == 3) { | |
1103 | # edit conflicts, considered as non-fast-forward | |
e83d36b6 | 1104 | print {*STDERR} 'Warning: Error ' . |
3c1ed90e | 1105 | $mediawiki->{error}->{code} . |
7c475837 | 1106 | ' from mediawiki: ' . $mediawiki->{error}->{details} . |
3c1ed90e | 1107 | ".\n"; |
86e95ef2 | 1108 | return ($oldrevid, 'non-fast-forward'); |
3c1ed90e MM |
1109 | } else { |
1110 | # Other errors. Shouldn't happen => just die() | |
1111 | die 'Fatal: Error ' . | |
1112 | $mediawiki->{error}->{code} . | |
7c475837 | 1113 | ' from mediawiki: ' . $mediawiki->{error}->{details} . "\n"; |
3c1ed90e MM |
1114 | } |
1115 | } | |
93f0d338 | 1116 | $newrevid = $result->{edit}->{newrevid}; |
e83d36b6 | 1117 | print {*STDERR} "Pushed file: ${new_sha1} - ${title}\n"; |
2045e293 | 1118 | } elsif ($export_media) { |
b3d98595 NKT |
1119 | $newrevid = mw_upload_file($complete_file_name, $new_sha1, |
1120 | $extension, $page_deleted, | |
1121 | $summary); | |
2045e293 | 1122 | } else { |
e83d36b6 | 1123 | print {*STDERR} "Ignoring media file ${title}\n"; |
428c995c | 1124 | } |
b3d98595 | 1125 | $newrevid = ($newrevid or $oldrevid); |
86e95ef2 | 1126 | return ($newrevid, 'ok'); |
428c995c JN |
1127 | } |
1128 | ||
1129 | sub mw_push { | |
1130 | # multiple push statements can follow each other | |
86e95ef2 | 1131 | my @refsspecs = (shift, get_more_refs('push')); |
428c995c JN |
1132 | my $pushed; |
1133 | for my $refspec (@refsspecs) { | |
1134 | my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/ | |
8a43b36a | 1135 | or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>\n"); |
428c995c | 1136 | if ($force) { |
e83d36b6 | 1137 | print {*STDERR} "Warning: forced push not allowed on a MediaWiki.\n"; |
428c995c | 1138 | } |
aeb95eea | 1139 | if ($local eq EMPTY) { |
e83d36b6 CM |
1140 | print {*STDERR} "Cannot delete remote branch on a MediaWiki\n"; |
1141 | print {*STDOUT} "error ${remote} cannot delete\n"; | |
428c995c JN |
1142 | next; |
1143 | } | |
86e95ef2 | 1144 | if ($remote ne 'refs/heads/master') { |
e83d36b6 CM |
1145 | print {*STDERR} "Only push to the branch 'master' is supported on a MediaWiki\n"; |
1146 | print {*STDOUT} "error ${remote} only master allowed\n"; | |
428c995c JN |
1147 | next; |
1148 | } | |
1149 | if (mw_push_revision($local, $remote)) { | |
1150 | $pushed = 1; | |
1151 | } | |
1152 | } | |
1153 | ||
1154 | # Notify Git that the push is done | |
e83d36b6 | 1155 | print {*STDOUT} "\n"; |
428c995c | 1156 | |
93f0d338 | 1157 | if ($pushed && $dumb_push) { |
e83d36b6 CM |
1158 | print {*STDERR} "Just pushed some revisions to MediaWiki.\n"; |
1159 | print {*STDERR} "The pushed revisions now have to be re-imported, and your current branch\n"; | |
1160 | print {*STDERR} "needs to be updated with these re-imported commits. You can do this with\n"; | |
1161 | print {*STDERR} "\n"; | |
1162 | print {*STDERR} " git pull --rebase\n"; | |
1163 | print {*STDERR} "\n"; | |
428c995c | 1164 | } |
1aff8c62 | 1165 | return; |
428c995c JN |
1166 | } |
1167 | ||
1168 | sub mw_push_revision { | |
1169 | my $local = shift; | |
1170 | my $remote = shift; # actually, this has to be "refs/heads/master" at this point. | |
1171 | my $last_local_revid = get_last_local_revision(); | |
e83d36b6 | 1172 | print {*STDERR} ".\n"; # Finish sentence started by get_last_local_revision() |
428c995c | 1173 | my $last_remote_revid = get_last_remote_revision(); |
93f0d338 | 1174 | my $mw_revision = $last_remote_revid; |
428c995c JN |
1175 | |
1176 | # Get sha1 of commit pointed by local HEAD | |
86e95ef2 CM |
1177 | my $HEAD_sha1 = run_git("rev-parse ${local} 2>/dev/null"); |
1178 | chomp($HEAD_sha1); | |
428c995c | 1179 | # Get sha1 of commit pointed by remotes/$remotename/master |
86e95ef2 | 1180 | my $remoteorigin_sha1 = run_git("rev-parse refs/remotes/${remotename}/master 2>/dev/null"); |
428c995c JN |
1181 | chomp($remoteorigin_sha1); |
1182 | ||
1183 | if ($last_local_revid > 0 && | |
1184 | $last_local_revid < $last_remote_revid) { | |
1185 | return error_non_fast_forward($remote); | |
1186 | } | |
1187 | ||
1188 | if ($HEAD_sha1 eq $remoteorigin_sha1) { | |
1189 | # nothing to push | |
1190 | return 0; | |
1191 | } | |
1192 | ||
1193 | # Get every commit in between HEAD and refs/remotes/origin/master, | |
1194 | # including HEAD and refs/remotes/origin/master | |
1195 | my @commit_pairs = (); | |
1196 | if ($last_local_revid > 0) { | |
1197 | my $parsed_sha1 = $remoteorigin_sha1; | |
1198 | # Find a path from last MediaWiki commit to pushed commit | |
e83d36b6 | 1199 | print {*STDERR} "Computing path from local to remote ...\n"; |
86e95ef2 | 1200 | my @local_ancestry = split(/\n/, run_git("rev-list --boundary --parents ${local} ^${parsed_sha1}")); |
93e92d4d MM |
1201 | my %local_ancestry; |
1202 | foreach my $line (@local_ancestry) { | |
eb96b750 | 1203 | if (my ($child, $parents) = $line =~ /^-?([a-f0-9]+) ([a-f0-9 ]+)/) { |
11499573 | 1204 | foreach my $parent (split(/ /, $parents)) { |
93e92d4d MM |
1205 | $local_ancestry{$parent} = $child; |
1206 | } | |
eb96b750 | 1207 | } elsif (!$line =~ /^([a-f0-9]+)/) { |
86e95ef2 | 1208 | die "Unexpected output from git rev-list: ${line}\n"; |
93e92d4d MM |
1209 | } |
1210 | } | |
428c995c | 1211 | while ($parsed_sha1 ne $HEAD_sha1) { |
93e92d4d MM |
1212 | my $child = $local_ancestry{$parsed_sha1}; |
1213 | if (!$child) { | |
e83d36b6 | 1214 | print {*STDERR} "Cannot find a path in history from remote commit to last commit\n"; |
428c995c JN |
1215 | return error_non_fast_forward($remote); |
1216 | } | |
93e92d4d MM |
1217 | push(@commit_pairs, [$parsed_sha1, $child]); |
1218 | $parsed_sha1 = $child; | |
428c995c JN |
1219 | } |
1220 | } else { | |
1221 | # No remote mediawiki revision. Export the whole | |
1222 | # history (linearized with --first-parent) | |
e83d36b6 | 1223 | print {*STDERR} "Warning: no common ancestor, pushing complete history\n"; |
86e95ef2 | 1224 | my $history = run_git("rev-list --first-parent --children ${local}"); |
05d4c7b1 | 1225 | my @history = split(/\n/, $history); |
428c995c JN |
1226 | @history = @history[1..$#history]; |
1227 | foreach my $line (reverse @history) { | |
6b825a46 | 1228 | my @commit_info_split = split(/[ \n]/, $line); |
428c995c JN |
1229 | push(@commit_pairs, \@commit_info_split); |
1230 | } | |
1231 | } | |
1232 | ||
1233 | foreach my $commit_info_split (@commit_pairs) { | |
1234 | my $sha1_child = @{$commit_info_split}[0]; | |
1235 | my $sha1_commit = @{$commit_info_split}[1]; | |
86e95ef2 | 1236 | my $diff_infos = run_git("diff-tree -r --raw -z ${sha1_child} ${sha1_commit}"); |
428c995c JN |
1237 | # TODO: we could detect rename, and encode them with a #redirect on the wiki. |
1238 | # TODO: for now, it's just a delete+add | |
1239 | my @diff_info_list = split(/\0/, $diff_infos); | |
28c24bd7 | 1240 | # Keep the subject line of the commit message as mediawiki comment for the revision |
86e95ef2 | 1241 | my $commit_msg = run_git(qq(log --no-walk --format="%s" ${sha1_commit})); |
428c995c JN |
1242 | chomp($commit_msg); |
1243 | # Push every blob | |
1244 | while (@diff_info_list) { | |
3c1ed90e | 1245 | my $status; |
428c995c JN |
1246 | # git diff-tree -z gives an output like |
1247 | # <metadata>\0<filename1>\0 | |
1248 | # <metadata>\0<filename2>\0 | |
1249 | # and we've split on \0. | |
1250 | my $info = shift(@diff_info_list); | |
1251 | my $file = shift(@diff_info_list); | |
3c1ed90e | 1252 | ($mw_revision, $status) = mw_push_file($info, $file, $commit_msg, $mw_revision); |
86e95ef2 | 1253 | if ($status eq 'non-fast-forward') { |
3c1ed90e MM |
1254 | # we may already have sent part of the |
1255 | # commit to MediaWiki, but it's too | |
1256 | # late to cancel it. Stop the push in | |
1257 | # the middle, but still give an | |
1258 | # accurate error message. | |
1259 | return error_non_fast_forward($remote); | |
1260 | } | |
86e95ef2 | 1261 | if ($status ne 'ok') { |
8a43b36a | 1262 | die("Unknown error from mw_push_file()\n"); |
3c1ed90e | 1263 | } |
93f0d338 | 1264 | } |
b8b4e1b3 | 1265 | if (!$dumb_push) { |
86e95ef2 | 1266 | run_git(qq(notes --ref=${remotename}/mediawiki add -f -m "mediawiki_revision: ${mw_revision}" ${sha1_commit})); |
428c995c JN |
1267 | } |
1268 | } | |
1269 | ||
e83d36b6 | 1270 | print {*STDOUT} "ok ${remote}\n"; |
428c995c JN |
1271 | return 1; |
1272 | } | |
b3d98595 NKT |
1273 | |
1274 | sub get_allowed_file_extensions { | |
192f7a08 | 1275 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
b3d98595 NKT |
1276 | |
1277 | my $query = { | |
1278 | action => 'query', | |
1279 | meta => 'siteinfo', | |
1280 | siprop => 'fileextensions' | |
1281 | }; | |
1282 | my $result = $mediawiki->api($query); | |
81f6a7a4 CM |
1283 | my @file_extensions = map { $_->{ext}} @{$result->{query}->{fileextensions}}; |
1284 | my %hashFile = map { $_ => 1 } @file_extensions; | |
b3d98595 NKT |
1285 | |
1286 | return %hashFile; | |
1287 | } | |
6a9e55b0 | 1288 | |
6df7e0df MM |
1289 | # In memory cache for MediaWiki namespace ids. |
1290 | my %namespace_id; | |
1291 | ||
1292 | # Namespaces whose id is cached in the configuration file | |
1293 | # (to avoid duplicates) | |
1294 | my %cached_mw_namespace_id; | |
1295 | ||
6a9e55b0 PV |
1296 | # Return MediaWiki id for a canonical namespace name. |
1297 | # Ex.: "File", "Project". | |
6a9e55b0 | 1298 | sub get_mw_namespace_id { |
192f7a08 | 1299 | $mediawiki = connect_maybe($mediawiki, $remotename, $url); |
6a9e55b0 PV |
1300 | my $name = shift; |
1301 | ||
1302 | if (!exists $namespace_id{$name}) { | |
1303 | # Look at configuration file, if the record for that namespace is | |
6df7e0df | 1304 | # already cached. Namespaces are stored in form: |
6a9e55b0 | 1305 | # "Name_of_namespace:Id_namespace", ex.: "File:6". |
86e95ef2 CM |
1306 | my @temp = split(/\n/, |
1307 | run_git("config --get-all remote.${remotename}.namespaceCache")); | |
6a9e55b0 PV |
1308 | chomp(@temp); |
1309 | foreach my $ns (@temp) { | |
6df7e0df | 1310 | my ($n, $id) = split(/:/, $ns); |
2272400c MM |
1311 | if ($id eq 'notANameSpace') { |
1312 | $namespace_id{$n} = {is_namespace => 0}; | |
1313 | } else { | |
1314 | $namespace_id{$n} = {is_namespace => 1, id => $id}; | |
1315 | } | |
6df7e0df | 1316 | $cached_mw_namespace_id{$n} = 1; |
6a9e55b0 PV |
1317 | } |
1318 | } | |
1319 | ||
1320 | if (!exists $namespace_id{$name}) { | |
e83d36b6 | 1321 | print {*STDERR} "Namespace ${name} not found in cache, querying the wiki ...\n"; |
6a9e55b0 PV |
1322 | # NS not found => get namespace id from MW and store it in |
1323 | # configuration file. | |
1324 | my $query = { | |
1325 | action => 'query', | |
1326 | meta => 'siteinfo', | |
1327 | siprop => 'namespaces' | |
1328 | }; | |
1329 | my $result = $mediawiki->api($query); | |
1330 | ||
1331 | while (my ($id, $ns) = each(%{$result->{query}->{namespaces}})) { | |
6df7e0df | 1332 | if (defined($ns->{id}) && defined($ns->{canonical})) { |
2272400c | 1333 | $namespace_id{$ns->{canonical}} = {is_namespace => 1, id => $ns->{id}}; |
6df7e0df MM |
1334 | if ($ns->{'*'}) { |
1335 | # alias (e.g. french Fichier: as alias for canonical File:) | |
2272400c | 1336 | $namespace_id{$ns->{'*'}} = {is_namespace => 1, id => $ns->{id}}; |
6df7e0df MM |
1337 | } |
1338 | } | |
6a9e55b0 PV |
1339 | } |
1340 | } | |
1341 | ||
2272400c MM |
1342 | my $ns = $namespace_id{$name}; |
1343 | my $id; | |
6df7e0df | 1344 | |
b8b4e1b3 | 1345 | if (!defined $ns) { |
09eebbad AB |
1346 | my @namespaces = map { s/ /_/g; $_; } sort keys %namespace_id; |
1347 | print {*STDERR} "No such namespace ${name} on MediaWiki, known namespaces: @namespaces\n"; | |
2272400c MM |
1348 | $ns = {is_namespace => 0}; |
1349 | $namespace_id{$name} = $ns; | |
1350 | } | |
1351 | ||
1352 | if ($ns->{is_namespace}) { | |
1353 | $id = $ns->{id}; | |
1354 | } | |
1355 | ||
1356 | # Store "notANameSpace" as special value for inexisting namespaces | |
1357 | my $store_id = ($id || 'notANameSpace'); | |
1358 | ||
68840cb5 | 1359 | # Store explicitly requested namespaces on disk |
2272400c | 1360 | if (!exists $cached_mw_namespace_id{$name}) { |
86e95ef2 | 1361 | run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}")); |
2272400c | 1362 | $cached_mw_namespace_id{$name} = 1; |
6a9e55b0 | 1363 | } |
2272400c | 1364 | return $id; |
6a9e55b0 | 1365 | } |
2045e293 MM |
1366 | |
1367 | sub get_mw_namespace_id_for_page { | |
6c2fbe25 CM |
1368 | my $namespace = shift; |
1369 | if ($namespace =~ /^([^:]*):/) { | |
2045e293 MM |
1370 | return get_mw_namespace_id($namespace); |
1371 | } else { | |
1372 | return; | |
1373 | } | |
1374 | } |