contrib/mw-to-git/git-remote-mediawiki.perl

   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.
  12 # Documentation & bugtracker: https://github.com/moy/Git-Mediawiki/
  13
  14 use strict;
  15 use MediaWiki::API;
  16 use Git;
  17 use DateTime::Format::ISO8601;
  18 use warnings;
  19
  20 # By default, use UTF-8 to communicate with Git and the user
  21 binmode STDERR, ":utf8";
  22 binmode STDOUT, ":utf8";
  23
  24 use URI::Escape;
  25 use IPC::Open2;
  26
  27 # Mediawiki filenames can contain forward slashes. This variable decides by which pattern they should be replaced
  28 use constant SLASH_REPLACEMENT => "%2F";
  29
  30 # It's not always possible to delete pages (may require some
  31 # privileges). Deleted pages are replaced with this content.
  32 use constant DELETED_CONTENT => "[[Category:Deleted]]\n";
  33
  34 # It's not possible to create empty pages. New empty files in Git are
  35 # sent with this content instead.
  36 use constant EMPTY_CONTENT => "<!-- empty page -->\n";
  37
  38 # used to reflect file creation or deletion in diff.
  39 use constant NULL_SHA1 => "0000000000000000000000000000000000000000";
  40
  41 # Used on Git's side to reflect empty edit messages on the wiki
  42 use constant EMPTY_MESSAGE => '*Empty MediaWiki Message*';
  43
  44 my $remotename = $ARGV[0];
  45 my $url = $ARGV[1];
  46
  47 # Accept both space-separated and multiple keys in config file.
  48 # Spaces should be written as _ anyway because we'll use chomp.
  49 my @tracked_pages = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".pages"));
  50 chomp(@tracked_pages);
  51
  52 # Just like @tracked_pages, but for MediaWiki categories.
  53 my @tracked_categories = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".categories"));
  54 chomp(@tracked_categories);
  55
  56 # Import media files on pull
  57 my $import_media = run_git("config --get --bool remote.". $remotename .".mediaimport");
  58 chomp($import_media);
  59 $import_media = ($import_media eq "true");
  60
  61 # Export media files on push
  62 my $export_media = run_git("config --get --bool remote.". $remotename .".mediaexport");
  63 chomp($export_media);
  64 $export_media = !($export_media eq "false");
  65
  66 my $wiki_login = run_git("config --get remote.". $remotename .".mwLogin");
  67 # Note: mwPassword is discourraged. Use the credential system instead.
  68 my $wiki_passwd = run_git("config --get remote.". $remotename .".mwPassword");
  69 my $wiki_domain = run_git("config --get remote.". $remotename .".mwDomain");
  70 chomp($wiki_login);
  71 chomp($wiki_passwd);
  72 chomp($wiki_domain);
  73
  74 # Import only last revisions (both for clone and fetch)
  75 my $shallow_import = run_git("config --get --bool remote.". $remotename .".shallow");
  76 chomp($shallow_import);
  77 $shallow_import = ($shallow_import eq "true");
  78
  79 # Fetch (clone and pull) by revisions instead of by pages. This behavior
  80 # is more efficient when we have a wiki with lots of pages and we fetch
  81 # the revisions quite often so that they concern only few pages.
  82 # Possible values:
  83 # - by_rev: perform one query per new revision on the remote wiki
  84 # - by_page: query each tracked page for new revision
  85 my $fetch_strategy = run_git("config --get remote.$remotename.fetchStrategy");
  86 unless ($fetch_strategy) {
  87         $fetch_strategy = run_git("config --get mediawiki.fetchStrategy");
  88 }
  89 chomp($fetch_strategy);
  90 unless ($fetch_strategy) {
  91         $fetch_strategy = "by_page";
  92 }
  93
  94 # Dumb push: don't update notes and mediawiki ref to reflect the last push.
  95 #
  96 # Configurable with mediawiki.dumbPush, or per-remote with
  97 # remote.<remotename>.dumbPush.
  98 #
  99 # This means the user will have to re-import the just-pushed
 100 # revisions. On the other hand, this means that the Git revisions
 101 # corresponding to MediaWiki revisions are all imported from the wiki,
 102 # regardless of whether they were initially created in Git or from the
 103 # web interface, hence all users will get the same history (i.e. if
 104 # the push from Git to MediaWiki loses some information, everybody
 105 # will get the history with information lost). If the import is
 106 # deterministic, this means everybody gets the same sha1 for each
 107 # MediaWiki revision.
 108 my $dumb_push = run_git("config --get --bool remote.$remotename.dumbPush");
 109 unless ($dumb_push) {
 110         $dumb_push = run_git("config --get --bool mediawiki.dumbPush");
 111 }
 112 chomp($dumb_push);
 113 $dumb_push = ($dumb_push eq "true");
 114
 115 my $wiki_name = $url;
 116 $wiki_name =~ s/[^\/]*:\/\///;
 117 # If URL is like http://user:password@example.com/, we clearly don't
 118 # want the password in $wiki_name. While we're there, also remove user
 119 # and '@' sign, to avoid author like MWUser@HTTPUser@host.com
 120 $wiki_name =~ s/^.*@//;
 121
 122 # Commands parser
 123 my $entry;
 124 my @cmd;
 125 while (<STDIN>) {
 126         chomp;
 127         @cmd = split(/ /);
 128         if (defined($cmd[0])) {
 129                 # Line not blank
 130                 if ($cmd[0] eq "capabilities") {
 131                         die("Too many arguments for capabilities") unless (!defined($cmd[1]));
 132                         mw_capabilities();
 133                 } elsif ($cmd[0] eq "list") {
 134                         die("Too many arguments for list") unless (!defined($cmd[2]));
 135                         mw_list($cmd[1]);
 136                 } elsif ($cmd[0] eq "import") {
 137                         die("Invalid arguments for import") unless ($cmd[1] ne "" && !defined($cmd[2]));
 138                         mw_import($cmd[1]);
 139                 } elsif ($cmd[0] eq "option") {
 140                         die("Too many arguments for option") unless ($cmd[1] ne "" && $cmd[2] ne "" && !defined($cmd[3]));
 141                         mw_option($cmd[1],$cmd[2]);
 142                 } elsif ($cmd[0] eq "push") {
 143                         mw_push($cmd[1]);
 144                 } else {
 145                         print STDERR "Unknown command. Aborting...\n";
 146                         last;
 147                 }
 148         } else {
 149                 # blank line: we should terminate
 150                 last;
 151         }
 152
 153         BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
 154                          # command is fully processed.
 155 }
 156
 157 ########################## Functions ##############################
 158
 159 # MediaWiki API instance, created lazily.
 160 my $mediawiki;
 161
 162 sub mw_connect_maybe {
 163         if ($mediawiki) {
 164                 return;
 165         }
 166         $mediawiki = MediaWiki::API->new;
 167         $mediawiki->{config}->{api_url} = "$url/api.php";
 168         if ($wiki_login) {
 169                 my %credential = (
 170                         'url' => $url,
 171                         'username' => $wiki_login,
 172                         'password' => $wiki_passwd
 173                 );
 174                 Git::credential(\%credential);
 175                 my $request = {lgname => $credential{username},
 176                                lgpassword => $credential{password},
 177                                lgdomain => $wiki_domain};
 178                 if ($mediawiki->login($request)) {
 179                         Git::credential(\%credential, 'approve');
 180                         print STDERR "Logged in mediawiki user \"$credential{username}\".\n";
 181                 } else {
 182                         print STDERR "Failed to log in mediawiki user \"$credential{username}\" on $url\n";
 183                         print STDERR "  (error " .
 184                                 $mediawiki->{error}->{code} . ': ' .
 185                                 $mediawiki->{error}->{details} . ")\n";
 186                         Git::credential(\%credential, 'reject');
 187                         exit 1;
 188                 }
 189         }
 190 }
 191
 192 sub fatal_mw_error {
 193         my $action = shift;
 194         print STDERR "fatal: could not $action.\n";
 195         print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 196         if ($url =~ /^https/) {
 197                 print STDERR "fatal: make sure '$url/api.php' is a valid page\n";
 198                 print STDERR "fatal: and the SSL certificate is correct.\n";
 199         } else {
 200                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 201         }
 202         print STDERR "fatal: (error " .
 203             $mediawiki->{error}->{code} . ': ' .
 204             $mediawiki->{error}->{details} . ")\n";
 205         exit 1;
 206 }
 207
 208 ## Functions for listing pages on the remote wiki
 209 sub get_mw_tracked_pages {
 210         my $pages = shift;
 211         get_mw_page_list(\@tracked_pages, $pages);
 212 }
 213
 214 sub get_mw_page_list {
 215         my $page_list = shift;
 216         my $pages = shift;
 217         my @some_pages = @$page_list;
 218         while (@some_pages) {
 219                 my $last = 50;
 220                 if ($#some_pages < $last) {
 221                         $last = $#some_pages;
 222                 }
 223                 my @slice = @some_pages[0..$last];
 224                 get_mw_first_pages(\@slice, $pages);
 225                 @some_pages = @some_pages[51..$#some_pages];
 226         }
 227 }
 228
 229 sub get_mw_tracked_categories {
 230         my $pages = shift;
 231         foreach my $category (@tracked_categories) {
 232                 if (index($category, ':') < 0) {
 233                         # Mediawiki requires the Category
 234                         # prefix, but let's not force the user
 235                         # to specify it.
 236                         $category = "Category:" . $category;
 237                 }
 238                 my $mw_pages = $mediawiki->list( {
 239                         action => 'query',
 240                         list => 'categorymembers',
 241                         cmtitle => $category,
 242                         cmlimit => 'max' } )
 243                         || die $mediawiki->{error}->{code} . ': '
 244                                 . $mediawiki->{error}->{details};
 245                 foreach my $page (@{$mw_pages}) {
 246                         $pages->{$page->{title}} = $page;
 247                 }
 248         }
 249 }
 250
 251 sub get_mw_all_pages {
 252         my $pages = shift;
 253         # No user-provided list, get the list of pages from the API.
 254         my $mw_pages = $mediawiki->list({
 255                 action => 'query',
 256                 list => 'allpages',
 257                 aplimit => 'max'
 258         });
 259         if (!defined($mw_pages)) {
 260                 fatal_mw_error("get the list of wiki pages");
 261         }
 262         foreach my $page (@{$mw_pages}) {
 263                 $pages->{$page->{title}} = $page;
 264         }
 265 }
 266
 267 # queries the wiki for a set of pages. Meant to be used within a loop
 268 # querying the wiki for slices of page list.
 269 sub get_mw_first_pages {
 270         my $some_pages = shift;
 271         my @some_pages = @{$some_pages};
 272
 273         my $pages = shift;
 274
 275         # pattern 'page1|page2|...' required by the API
 276         my $titles = join('|', @some_pages);
 277
 278         my $mw_pages = $mediawiki->api({
 279                 action => 'query',
 280                 titles => $titles,
 281         });
 282         if (!defined($mw_pages)) {
 283                 fatal_mw_error("query the list of wiki pages");
 284         }
 285         while (my ($id, $page) = each(%{$mw_pages->{query}->{pages}})) {
 286                 if ($id < 0) {
 287                         print STDERR "Warning: page $page->{title} not found on wiki\n";
 288                 } else {
 289                         $pages->{$page->{title}} = $page;
 290                 }
 291         }
 292 }
 293
 294 # Get the list of pages to be fetched according to configuration.
 295 sub get_mw_pages {
 296         mw_connect_maybe();
 297
 298         print STDERR "Listing pages on remote wiki...\n";
 299
 300         my %pages; # hash on page titles to avoid duplicates
 301         my $user_defined;
 302         if (@tracked_pages) {
 303                 $user_defined = 1;
 304                 # The user provided a list of pages titles, but we
 305                 # still need to query the API to get the page IDs.
 306                 get_mw_tracked_pages(\%pages);
 307         }
 308         if (@tracked_categories) {
 309                 $user_defined = 1;
 310                 get_mw_tracked_categories(\%pages);
 311         }
 312         if (!$user_defined) {
 313                 get_mw_all_pages(\%pages);
 314         }
 315         if ($import_media) {
 316                 print STDERR "Getting media files for selected pages...\n";
 317                 if ($user_defined) {
 318                         get_linked_mediafiles(\%pages);
 319                 } else {
 320                         get_all_mediafiles(\%pages);
 321                 }
 322         }
 323         print STDERR (scalar keys %pages) . " pages found.\n";
 324         return %pages;
 325 }
 326
 327 # usage: $out = run_git("command args");
 328 #        $out = run_git("command args", "raw"); # don't interpret output as UTF-8.
 329 sub run_git {
 330         my $args = shift;
 331         my $encoding = (shift || "encoding(UTF-8)");
 332         open(my $git, "-|:$encoding", "git " . $args);
 333         my $res = do { local $/; <$git> };
 334         close($git);
 335
 336         return $res;
 337 }
 338
 339
 340 sub get_all_mediafiles {
 341         my $pages = shift;
 342         # Attach list of all pages for media files from the API,
 343         # they are in a different namespace, only one namespace
 344         # can be queried at the same moment
 345         my $mw_pages = $mediawiki->list({
 346                 action => 'query',
 347                 list => 'allpages',
 348                 apnamespace => get_mw_namespace_id("File"),
 349                 aplimit => 'max'
 350         });
 351         if (!defined($mw_pages)) {
 352                 print STDERR "fatal: could not get the list of pages for media files.\n";
 353                 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 354                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 355                 exit 1;
 356         }
 357         foreach my $page (@{$mw_pages}) {
 358                 $pages->{$page->{title}} = $page;
 359         }
 360 }
 361
 362 sub get_linked_mediafiles {
 363         my $pages = shift;
 364         my @titles = map $_->{title}, values(%{$pages});
 365
 366         # The query is split in small batches because of the MW API limit of
 367         # the number of links to be returned (500 links max).
 368         my $batch = 10;
 369         while (@titles) {
 370                 if ($#titles < $batch) {
 371                         $batch = $#titles;
 372                 }
 373                 my @slice = @titles[0..$batch];
 374
 375                 # pattern 'page1|page2|...' required by the API
 376                 my $mw_titles = join('|', @slice);
 377
 378                 # Media files could be included or linked from
 379                 # a page, get all related
 380                 my $query = {
 381                         action => 'query',
 382                         prop => 'links|images',
 383                         titles => $mw_titles,
 384                         plnamespace => get_mw_namespace_id("File"),
 385                         pllimit => 'max'
 386                 };
 387                 my $result = $mediawiki->api($query);
 388
 389                 while (my ($id, $page) = each(%{$result->{query}->{pages}})) {
 390                         my @media_titles;
 391                         if (defined($page->{links})) {
 392                                 my @link_titles = map $_->{title}, @{$page->{links}};
 393                                 push(@media_titles, @link_titles);
 394                         }
 395                         if (defined($page->{images})) {
 396                                 my @image_titles = map $_->{title}, @{$page->{images}};
 397                                 push(@media_titles, @image_titles);
 398                         }
 399                         if (@media_titles) {
 400                                 get_mw_page_list(\@media_titles, $pages);
 401                         }
 402                 }
 403
 404                 @titles = @titles[($batch+1)..$#titles];
 405         }
 406 }
 407
 408 sub get_mw_mediafile_for_page_revision {
 409         # Name of the file on Wiki, with the prefix.
 410         my $filename = shift;
 411         my $timestamp = shift;
 412         my %mediafile;
 413
 414         # Search if on a media file with given timestamp exists on
 415         # MediaWiki. In that case download the file.
 416         my $query = {
 417                 action => 'query',
 418                 prop => 'imageinfo',
 419                 titles => "File:" . $filename,
 420                 iistart => $timestamp,
 421                 iiend => $timestamp,
 422                 iiprop => 'timestamp|archivename|url',
 423                 iilimit => 1
 424         };
 425         my $result = $mediawiki->api($query);
 426
 427         my ($fileid, $file) = each( %{$result->{query}->{pages}} );
 428         # If not defined it means there is no revision of the file for
 429         # given timestamp.
 430         if (defined($file->{imageinfo})) {
 431                 $mediafile{title} = $filename;
 432
 433                 my $fileinfo = pop(@{$file->{imageinfo}});
 434                 $mediafile{timestamp} = $fileinfo->{timestamp};
 435                 # Mediawiki::API's download function doesn't support https URLs
 436                 # and can't download old versions of files.
 437                 print STDERR "\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
 438                 $mediafile{content} = download_mw_mediafile($fileinfo->{url});
 439         }
 440         return %mediafile;
 441 }
 442
 443 sub download_mw_mediafile {
 444         my $url = shift;
 445
 446         my $response = $mediawiki->{ua}->get($url);
 447         if ($response->code == 200) {
 448                 return $response->decoded_content;
 449         } else {
 450                 print STDERR "Error downloading mediafile from :\n";
 451                 print STDERR "URL: $url\n";
 452                 print STDERR "Server response: " . $response->code . " " . $response->message . "\n";
 453                 exit 1;
 454         }
 455 }
 456
 457 sub get_last_local_revision {
 458         # Get note regarding last mediawiki revision
 459         my $note = run_git("notes --ref=$remotename/mediawiki show refs/mediawiki/$remotename/master 2>/dev/null");
 460         my @note_info = split(/ /, $note);
 461
 462         my $lastrevision_number;
 463         if (!(defined($note_info[0]) && $note_info[0] eq "mediawiki_revision:")) {
 464                 print STDERR "No previous mediawiki revision found";
 465                 $lastrevision_number = 0;
 466         } else {
 467                 # Notes are formatted : mediawiki_revision: #number
 468                 $lastrevision_number = $note_info[1];
 469                 chomp($lastrevision_number);
 470                 print STDERR "Last local mediawiki revision found is $lastrevision_number";
 471         }
 472         return $lastrevision_number;
 473 }
 474
 475 # Remember the timestamp corresponding to a revision id.
 476 my %basetimestamps;
 477
 478 # Get the last remote revision without taking in account which pages are
 479 # tracked or not. This function makes a single request to the wiki thus
 480 # avoid a loop onto all tracked pages. This is useful for the fetch-by-rev
 481 # option.
 482 sub get_last_global_remote_rev {
 483         mw_connect_maybe();
 484
 485         my $query = {
 486                 action => 'query',
 487                 list => 'recentchanges',
 488                 prop => 'revisions',
 489                 rclimit => '1',
 490                 rcdir => 'older',
 491         };
 492         my $result = $mediawiki->api($query);
 493         return $result->{query}->{recentchanges}[0]->{revid};
 494 }
 495
 496 # Get the last remote revision concerning the tracked pages and the tracked
 497 # categories.
 498 sub get_last_remote_revision {
 499         mw_connect_maybe();
 500
 501         my %pages_hash = get_mw_pages();
 502         my @pages = values(%pages_hash);
 503
 504         my $max_rev_num = 0;
 505
 506         print STDERR "Getting last revision id on tracked pages...\n";
 507
 508         foreach my $page (@pages) {
 509                 my $id = $page->{pageid};
 510
 511                 my $query = {
 512                         action => 'query',
 513                         prop => 'revisions',
 514                         rvprop => 'ids|timestamp',
 515                         pageids => $id,
 516                 };
 517
 518                 my $result = $mediawiki->api($query);
 519
 520                 my $lastrev = pop(@{$result->{query}->{pages}->{$id}->{revisions}});
 521
 522                 $basetimestamps{$lastrev->{revid}} = $lastrev->{timestamp};
 523
 524                 $max_rev_num = ($lastrev->{revid} > $max_rev_num ? $lastrev->{revid} : $max_rev_num);
 525         }
 526
 527         print STDERR "Last remote revision found is $max_rev_num.\n";
 528         return $max_rev_num;
 529 }
 530
 531 # Clean content before sending it to MediaWiki
 532 sub mediawiki_clean {
 533         my $string = shift;
 534         my $page_created = shift;
 535         # Mediawiki does not allow blank space at the end of a page and ends with a single \n.
 536         # This function right trims a string and adds a \n at the end to follow this rule
 537         $string =~ s/\s+$//;
 538         if ($string eq "" && $page_created) {
 539                 # Creating empty pages is forbidden.
 540                 $string = EMPTY_CONTENT;
 541         }
 542         return $string."\n";
 543 }
 544
 545 # Filter applied on MediaWiki data before adding them to Git
 546 sub mediawiki_smudge {
 547         my $string = shift;
 548         if ($string eq EMPTY_CONTENT) {
 549                 $string = "";
 550         }
 551         # This \n is important. This is due to mediawiki's way to handle end of files.
 552         return $string."\n";
 553 }
 554
 555 sub mediawiki_clean_filename {
 556         my $filename = shift;
 557         $filename =~ s/@{[SLASH_REPLACEMENT]}/\//g;
 558         # [, ], |, {, and } are forbidden by MediaWiki, even URL-encoded.
 559         # Do a variant of URL-encoding, i.e. looks like URL-encoding,
 560         # but with _ added to prevent MediaWiki from thinking this is
 561         # an actual special character.
 562         $filename =~ s/[\[\]\{\}\|]/sprintf("_%%_%x", ord($&))/ge;
 563         # If we use the uri escape before
 564         # we should unescape here, before anything
 565
 566         return $filename;
 567 }
 568
 569 sub mediawiki_smudge_filename {
 570         my $filename = shift;
 571         $filename =~ s/\//@{[SLASH_REPLACEMENT]}/g;
 572         $filename =~ s/ /_/g;
 573         # Decode forbidden characters encoded in mediawiki_clean_filename
 574         $filename =~ s/_%_([0-9a-fA-F][0-9a-fA-F])/sprintf("%c", hex($1))/ge;
 575         return $filename;
 576 }
 577
 578 sub literal_data {
 579         my ($content) = @_;
 580         print STDOUT "data ", bytes::length($content), "\n", $content;
 581 }
 582
 583 sub literal_data_raw {
 584         # Output possibly binary content.
 585         my ($content) = @_;
 586         # Avoid confusion between size in bytes and in characters
 587         utf8::downgrade($content);
 588         binmode STDOUT, ":raw";
 589         print STDOUT "data ", bytes::length($content), "\n", $content;
 590         binmode STDOUT, ":utf8";
 591 }
 592
 593 sub mw_capabilities {
 594         # Revisions are imported to the private namespace
 595         # refs/mediawiki/$remotename/ by the helper and fetched into
 596         # refs/remotes/$remotename later by fetch.
 597         print STDOUT "refspec refs/heads/*:refs/mediawiki/$remotename/*\n";
 598         print STDOUT "import\n";
 599         print STDOUT "list\n";
 600         print STDOUT "push\n";
 601         print STDOUT "\n";
 602 }
 603
 604 sub mw_list {
 605         # MediaWiki do not have branches, we consider one branch arbitrarily
 606         # called master, and HEAD pointing to it.
 607         print STDOUT "? refs/heads/master\n";
 608         print STDOUT "\@refs/heads/master HEAD\n";
 609         print STDOUT "\n";
 610 }
 611
 612 sub mw_option {
 613         print STDERR "remote-helper command 'option $_[0]' not yet implemented\n";
 614         print STDOUT "unsupported\n";
 615 }
 616
 617 sub fetch_mw_revisions_for_page {
 618         my $page = shift;
 619         my $id = shift;
 620         my $fetch_from = shift;
 621         my @page_revs = ();
 622         my $query = {
 623                 action => 'query',
 624                 prop => 'revisions',
 625                 rvprop => 'ids',
 626                 rvdir => 'newer',
 627                 rvstartid => $fetch_from,
 628                 rvlimit => 500,
 629                 pageids => $id,
 630         };
 631
 632         my $revnum = 0;
 633         # Get 500 revisions at a time due to the mediawiki api limit
 634         while (1) {
 635                 my $result = $mediawiki->api($query);
 636
 637                 # Parse each of those 500 revisions
 638                 foreach my $revision (@{$result->{query}->{pages}->{$id}->{revisions}}) {
 639                         my $page_rev_ids;
 640                         $page_rev_ids->{pageid} = $page->{pageid};
 641                         $page_rev_ids->{revid} = $revision->{revid};
 642                         push(@page_revs, $page_rev_ids);
 643                         $revnum++;
 644                 }
 645                 last unless $result->{'query-continue'};
 646                 $query->{rvstartid} = $result->{'query-continue'}->{revisions}->{rvstartid};
 647         }
 648         if ($shallow_import && @page_revs) {
 649                 print STDERR "  Found 1 revision (shallow import).\n";
 650                 @page_revs = sort {$b->{revid} <=> $a->{revid}} (@page_revs);
 651                 return $page_revs[0];
 652         }
 653         print STDERR "  Found ", $revnum, " revision(s).\n";
 654         return @page_revs;
 655 }
 656
 657 sub fetch_mw_revisions {
 658         my $pages = shift; my @pages = @{$pages};
 659         my $fetch_from = shift;
 660
 661         my @revisions = ();
 662         my $n = 1;
 663         foreach my $page (@pages) {
 664                 my $id = $page->{pageid};
 665
 666                 print STDERR "page $n/", scalar(@pages), ": ". $page->{title} ."\n";
 667                 $n++;
 668                 my @page_revs = fetch_mw_revisions_for_page($page, $id, $fetch_from);
 669                 @revisions = (@page_revs, @revisions);
 670         }
 671
 672         return ($n, @revisions);
 673 }
 674
 675 sub fe_escape_path {
 676     my $path = shift;
 677     $path =~ s/\\/\\\\/g;
 678     $path =~ s/"/\\"/g;
 679     $path =~ s/\n/\\n/g;
 680     return '"' . $path . '"';
 681 }
 682
 683 sub import_file_revision {
 684         my $commit = shift;
 685         my %commit = %{$commit};
 686         my $full_import = shift;
 687         my $n = shift;
 688         my $mediafile = shift;
 689         my %mediafile;
 690         if ($mediafile) {
 691                 %mediafile = %{$mediafile};
 692         }
 693
 694         my $title = $commit{title};
 695         my $comment = $commit{comment};
 696         my $content = $commit{content};
 697         my $author = $commit{author};
 698         my $date = $commit{date};
 699
 700         print STDOUT "commit refs/mediawiki/$remotename/master\n";
 701         print STDOUT "mark :$n\n";
 702         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 703         literal_data($comment);
 704
 705         # If it's not a clone, we need to know where to start from
 706         if (!$full_import && $n == 1) {
 707                 print STDOUT "from refs/mediawiki/$remotename/master^0\n";
 708         }
 709         if ($content ne DELETED_CONTENT) {
 710                 print STDOUT "M 644 inline " .
 711                     fe_escape_path($title . ".mw") . "\n";
 712                 literal_data($content);
 713                 if (%mediafile) {
 714                         print STDOUT "M 644 inline "
 715                             . fe_escape_path($mediafile{title}) . "\n";
 716                         literal_data_raw($mediafile{content});
 717                 }
 718                 print STDOUT "\n\n";
 719         } else {
 720                 print STDOUT "D " . fe_escape_path($title . ".mw") . "\n";
 721         }
 722
 723         # mediawiki revision number in the git note
 724         if ($full_import && $n == 1) {
 725                 print STDOUT "reset refs/notes/$remotename/mediawiki\n";
 726         }
 727         print STDOUT "commit refs/notes/$remotename/mediawiki\n";
 728         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 729         literal_data("Note added by git-mediawiki during import");
 730         if (!$full_import && $n == 1) {
 731                 print STDOUT "from refs/notes/$remotename/mediawiki^0\n";
 732         }
 733         print STDOUT "N inline :$n\n";
 734         literal_data("mediawiki_revision: " . $commit{mw_revision});
 735         print STDOUT "\n\n";
 736 }
 737
 738 # parse a sequence of
 739 # <cmd> <arg1>
 740 # <cmd> <arg2>
 741 # \n
 742 # (like batch sequence of import and sequence of push statements)
 743 sub get_more_refs {
 744         my $cmd = shift;
 745         my @refs;
 746         while (1) {
 747                 my $line = <STDIN>;
 748                 if ($line =~ m/^$cmd (.*)$/) {
 749                         push(@refs, $1);
 750                 } elsif ($line eq "\n") {
 751                         return @refs;
 752                 } else {
 753                         die("Invalid command in a '$cmd' batch: ". $_);
 754                 }
 755         }
 756 }
 757
 758 sub mw_import {
 759         # multiple import commands can follow each other.
 760         my @refs = (shift, get_more_refs("import"));
 761         foreach my $ref (@refs) {
 762                 mw_import_ref($ref);
 763         }
 764         print STDOUT "done\n";
 765 }
 766
 767 sub mw_import_ref {
 768         my $ref = shift;
 769         # The remote helper will call "import HEAD" and
 770         # "import refs/heads/master".
 771         # Since HEAD is a symbolic ref to master (by convention,
 772         # followed by the output of the command "list" that we gave),
 773         # we don't need to do anything in this case.
 774         if ($ref eq "HEAD") {
 775                 return;
 776         }
 777
 778         mw_connect_maybe();
 779
 780         print STDERR "Searching revisions...\n";
 781         my $last_local = get_last_local_revision();
 782         my $fetch_from = $last_local + 1;
 783         if ($fetch_from == 1) {
 784                 print STDERR ", fetching from beginning.\n";
 785         } else {
 786                 print STDERR ", fetching from here.\n";
 787         }
 788
 789         my $n = 0;
 790         if ($fetch_strategy eq "by_rev") {
 791                 print STDERR "Fetching & writing export data by revs...\n";
 792                 $n = mw_import_ref_by_revs($fetch_from);
 793         } elsif ($fetch_strategy eq "by_page") {
 794                 print STDERR "Fetching & writing export data by pages...\n";
 795                 $n = mw_import_ref_by_pages($fetch_from);
 796         } else {
 797                 print STDERR "fatal: invalid fetch strategy \"$fetch_strategy\".\n";
 798                 print STDERR "Check your configuration variables remote.$remotename.fetchStrategy and mediawiki.fetchStrategy\n";
 799                 exit 1;
 800         }
 801
 802         if ($fetch_from == 1 && $n == 0) {
 803                 print STDERR "You appear to have cloned an empty MediaWiki.\n";
 804                 # Something has to be done remote-helper side. If nothing is done, an error is
 805                 # thrown saying that HEAD is referring to unknown object 0000000000000000000
 806                 # and the clone fails.
 807         }
 808 }
 809
 810 sub mw_import_ref_by_pages {
 811
 812         my $fetch_from = shift;
 813         my %pages_hash = get_mw_pages();
 814         my @pages = values(%pages_hash);
 815
 816         my ($n, @revisions) = fetch_mw_revisions(\@pages, $fetch_from);
 817
 818         @revisions = sort {$a->{revid} <=> $b->{revid}} @revisions;
 819         my @revision_ids = map $_->{revid}, @revisions;
 820
 821         return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
 822 }
 823
 824 sub mw_import_ref_by_revs {
 825
 826         my $fetch_from = shift;
 827         my %pages_hash = get_mw_pages();
 828
 829         my $last_remote = get_last_global_remote_rev();
 830         my @revision_ids = $fetch_from..$last_remote;
 831         return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
 832 }
 833
 834 # Import revisions given in second argument (array of integers).
 835 # Only pages appearing in the third argument (hash indexed by page titles)
 836 # will be imported.
 837 sub mw_import_revids {
 838         my $fetch_from = shift;
 839         my $revision_ids = shift;
 840         my $pages = shift;
 841
 842         my $n = 0;
 843         my $n_actual = 0;
 844         my $last_timestamp = 0; # Placeholer in case $rev->timestamp is undefined
 845
 846         foreach my $pagerevid (@$revision_ids) {
 847                 # Count page even if we skip it, since we display
 848                 # $n/$total and $total includes skipped pages.
 849                 $n++;
 850
 851                 # fetch the content of the pages
 852                 my $query = {
 853                         action => 'query',
 854                         prop => 'revisions',
 855                         rvprop => 'content|timestamp|comment|user|ids',
 856                         revids => $pagerevid,
 857                 };
 858
 859                 my $result = $mediawiki->api($query);
 860
 861                 if (!$result) {
 862                         die "Failed to retrieve modified page for revision $pagerevid";
 863                 }
 864
 865                 if (defined($result->{query}->{badrevids}->{$pagerevid})) {
 866                         # The revision id does not exist on the remote wiki.
 867                         next;
 868                 }
 869
 870                 if (!defined($result->{query}->{pages})) {
 871                         die "Invalid revision $pagerevid.";
 872                 }
 873
 874                 my @result_pages = values(%{$result->{query}->{pages}});
 875                 my $result_page = $result_pages[0];
 876                 my $rev = $result_pages[0]->{revisions}->[0];
 877
 878                 my $page_title = $result_page->{title};
 879
 880                 if (!exists($pages->{$page_title})) {
 881                         print STDERR "$n/", scalar(@$revision_ids),
 882                                 ": Skipping revision #$rev->{revid} of $page_title\n";
 883                         next;
 884                 }
 885
 886                 $n_actual++;
 887
 888                 my %commit;
 889                 $commit{author} = $rev->{user} || 'Anonymous';
 890                 $commit{comment} = $rev->{comment} || EMPTY_MESSAGE;
 891                 $commit{title} = mediawiki_smudge_filename($page_title);
 892                 $commit{mw_revision} = $rev->{revid};
 893                 $commit{content} = mediawiki_smudge($rev->{'*'});
 894
 895                 if (!defined($rev->{timestamp})) {
 896                         $last_timestamp++;
 897                 } else {
 898                         $last_timestamp = $rev->{timestamp};
 899                 }
 900                 $commit{date} = DateTime::Format::ISO8601->parse_datetime($last_timestamp);
 901
 902                 # Differentiates classic pages and media files.
 903                 my ($namespace, $filename) = $page_title =~ /^([^:]*):(.*)$/;
 904                 my %mediafile;
 905                 if ($namespace) {
 906                         my $id = get_mw_namespace_id($namespace);
 907                         if ($id && $id == get_mw_namespace_id("File")) {
 908                                 %mediafile = get_mw_mediafile_for_page_revision($filename, $rev->{timestamp});
 909                         }
 910                 }
 911                 # If this is a revision of the media page for new version
 912                 # of a file do one common commit for both file and media page.
 913                 # Else do commit only for that page.
 914                 print STDERR "$n/", scalar(@$revision_ids), ": Revision #$rev->{revid} of $commit{title}\n";
 915                 import_file_revision(\%commit, ($fetch_from == 1), $n_actual, \%mediafile);
 916         }
 917
 918         return $n_actual;
 919 }
 920
 921 sub error_non_fast_forward {
 922         my $advice = run_git("config --bool advice.pushNonFastForward");
 923         chomp($advice);
 924         if ($advice ne "false") {
 925                 # Native git-push would show this after the summary.
 926                 # We can't ask it to display it cleanly, so print it
 927                 # ourselves before.
 928                 print STDERR "To prevent you from losing history, non-fast-forward updates were rejected\n";
 929                 print STDERR "Merge the remote changes (e.g. 'git pull') before pushing again. See the\n";
 930                 print STDERR "'Note about fast-forwards' section of 'git push --help' for details.\n";
 931         }
 932         print STDOUT "error $_[0] \"non-fast-forward\"\n";
 933         return 0;
 934 }
 935
 936 sub mw_upload_file {
 937         my $complete_file_name = shift;
 938         my $new_sha1 = shift;
 939         my $extension = shift;
 940         my $file_deleted = shift;
 941         my $summary = shift;
 942         my $newrevid;
 943         my $path = "File:" . $complete_file_name;
 944         my %hashFiles = get_allowed_file_extensions();
 945         if (!exists($hashFiles{$extension})) {
 946                 print STDERR "$complete_file_name is not a permitted file on this wiki.\n";
 947                 print STDERR "Check the configuration of file uploads in your mediawiki.\n";
 948                 return $newrevid;
 949         }
 950         # Deleting and uploading a file requires a priviledged user
 951         if ($file_deleted) {
 952                 mw_connect_maybe();
 953                 my $query = {
 954                         action => 'delete',
 955                         title => $path,
 956                         reason => $summary
 957                 };
 958                 if (!$mediawiki->edit($query)) {
 959                         print STDERR "Failed to delete file on remote wiki\n";
 960                         print STDERR "Check your permissions on the remote site. Error code:\n";
 961                         print STDERR $mediawiki->{error}->{code} . ':' . $mediawiki->{error}->{details};
 962                         exit 1;
 963                 }
 964         } else {
 965                 # Don't let perl try to interpret file content as UTF-8 => use "raw"
 966                 my $content = run_git("cat-file blob $new_sha1", "raw");
 967                 if ($content ne "") {
 968                         mw_connect_maybe();
 969                         $mediawiki->{config}->{upload_url} =
 970                                 "$url/index.php/Special:Upload";
 971                         $mediawiki->edit({
 972                                 action => 'upload',
 973                                 filename => $complete_file_name,
 974                                 comment => $summary,
 975                                 file => [undef,
 976                                          $complete_file_name,
 977                                          Content => $content],
 978                                 ignorewarnings => 1,
 979                         }, {
 980                                 skip_encoding => 1
 981                         } ) || die $mediawiki->{error}->{code} . ':'
 982                                  . $mediawiki->{error}->{details};
 983                         my $last_file_page = $mediawiki->get_page({title => $path});
 984                         $newrevid = $last_file_page->{revid};
 985                         print STDERR "Pushed file: $new_sha1 - $complete_file_name.\n";
 986                 } else {
 987                         print STDERR "Empty file $complete_file_name not pushed.\n";
 988                 }
 989         }
 990         return $newrevid;
 991 }
 992
 993 sub mw_push_file {
 994         my $diff_info = shift;
 995         # $diff_info contains a string in this format:
 996         # 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status>
 997         my @diff_info_split = split(/[ \t]/, $diff_info);
 998
 999         # Filename, including .mw extension
1000         my $complete_file_name = shift;
1001         # Commit message
1002         my $summary = shift;
1003         # MediaWiki revision number. Keep the previous one by default,
1004         # in case there's no edit to perform.
1005         my $oldrevid = shift;
1006         my $newrevid;
1007
1008         if ($summary eq EMPTY_MESSAGE) {
1009                 $summary = '';
1010         }
1011
1012         my $new_sha1 = $diff_info_split[3];
1013         my $old_sha1 = $diff_info_split[2];
1014         my $page_created = ($old_sha1 eq NULL_SHA1);
1015         my $page_deleted = ($new_sha1 eq NULL_SHA1);
1016         $complete_file_name = mediawiki_clean_filename($complete_file_name);
1017
1018         my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/;
1019         if (!defined($extension)) {
1020                 $extension = "";
1021         }
1022         if ($extension eq "mw") {
1023                 my $ns = get_mw_namespace_id_for_page($complete_file_name);
1024                 if ($ns && $ns == get_mw_namespace_id("File") && (!$export_media)) {
1025                         print STDERR "Ignoring media file related page: $complete_file_name\n";
1026                         return ($oldrevid, "ok");
1027                 }
1028                 my $file_content;
1029                 if ($page_deleted) {
1030                         # Deleting a page usually requires
1031                         # special privileges. A common
1032                         # convention is to replace the page
1033                         # with this content instead:
1034                         $file_content = DELETED_CONTENT;
1035                 } else {
1036                         $file_content = run_git("cat-file blob $new_sha1");
1037                 }
1038
1039                 mw_connect_maybe();
1040
1041                 my $result = $mediawiki->edit( {
1042                         action => 'edit',
1043                         summary => $summary,
1044                         title => $title,
1045                         basetimestamp => $basetimestamps{$oldrevid},
1046                         text => mediawiki_clean($file_content, $page_created),
1047                                   }, {
1048                                           skip_encoding => 1 # Helps with names with accentuated characters
1049                                   });
1050                 if (!$result) {
1051                         if ($mediawiki->{error}->{code} == 3) {
1052                                 # edit conflicts, considered as non-fast-forward
1053                                 print STDERR 'Warning: Error ' .
1054                                     $mediawiki->{error}->{code} .
1055                                     ' from mediwiki: ' . $mediawiki->{error}->{details} .
1056                                     ".\n";
1057                                 return ($oldrevid, "non-fast-forward");
1058                         } else {
1059                                 # Other errors. Shouldn't happen => just die()
1060                                 die 'Fatal: Error ' .
1061                                     $mediawiki->{error}->{code} .
1062                                     ' from mediwiki: ' . $mediawiki->{error}->{details};
1063                         }
1064                 }
1065                 $newrevid = $result->{edit}->{newrevid};
1066                 print STDERR "Pushed file: $new_sha1 - $title\n";
1067         } elsif ($export_media) {
1068                 $newrevid = mw_upload_file($complete_file_name, $new_sha1,
1069                                            $extension, $page_deleted,
1070                                            $summary);
1071         } else {
1072                 print STDERR "Ignoring media file $title\n";
1073         }
1074         $newrevid = ($newrevid or $oldrevid);
1075         return ($newrevid, "ok");
1076 }
1077
1078 sub mw_push {
1079         # multiple push statements can follow each other
1080         my @refsspecs = (shift, get_more_refs("push"));
1081         my $pushed;
1082         for my $refspec (@refsspecs) {
1083                 my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/
1084                     or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>");
1085                 if ($force) {
1086                         print STDERR "Warning: forced push not allowed on a MediaWiki.\n";
1087                 }
1088                 if ($local eq "") {
1089                         print STDERR "Cannot delete remote branch on a MediaWiki\n";
1090                         print STDOUT "error $remote cannot delete\n";
1091                         next;
1092                 }
1093                 if ($remote ne "refs/heads/master") {
1094                         print STDERR "Only push to the branch 'master' is supported on a MediaWiki\n";
1095                         print STDOUT "error $remote only master allowed\n";
1096                         next;
1097                 }
1098                 if (mw_push_revision($local, $remote)) {
1099                         $pushed = 1;
1100                 }
1101         }
1102
1103         # Notify Git that the push is done
1104         print STDOUT "\n";
1105
1106         if ($pushed && $dumb_push) {
1107                 print STDERR "Just pushed some revisions to MediaWiki.\n";
1108                 print STDERR "The pushed revisions now have to be re-imported, and your current branch\n";
1109                 print STDERR "needs to be updated with these re-imported commits. You can do this with\n";
1110                 print STDERR "\n";
1111                 print STDERR "  git pull --rebase\n";
1112                 print STDERR "\n";
1113         }
1114 }
1115
1116 sub mw_push_revision {
1117         my $local = shift;
1118         my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
1119         my $last_local_revid = get_last_local_revision();
1120         print STDERR ".\n"; # Finish sentence started by get_last_local_revision()
1121         my $last_remote_revid = get_last_remote_revision();
1122         my $mw_revision = $last_remote_revid;
1123
1124         # Get sha1 of commit pointed by local HEAD
1125         my $HEAD_sha1 = run_git("rev-parse $local 2>/dev/null"); chomp($HEAD_sha1);
1126         # Get sha1 of commit pointed by remotes/$remotename/master
1127         my $remoteorigin_sha1 = run_git("rev-parse refs/remotes/$remotename/master 2>/dev/null");
1128         chomp($remoteorigin_sha1);
1129
1130         if ($last_local_revid > 0 &&
1131             $last_local_revid < $last_remote_revid) {
1132                 return error_non_fast_forward($remote);
1133         }
1134
1135         if ($HEAD_sha1 eq $remoteorigin_sha1) {
1136                 # nothing to push
1137                 return 0;
1138         }
1139
1140         # Get every commit in between HEAD and refs/remotes/origin/master,
1141         # including HEAD and refs/remotes/origin/master
1142         my @commit_pairs = ();
1143         if ($last_local_revid > 0) {
1144                 my $parsed_sha1 = $remoteorigin_sha1;
1145                 # Find a path from last MediaWiki commit to pushed commit
1146                 print STDERR "Computing path from local to remote ...\n";
1147                 my @local_ancestry = split(/\n/, run_git("rev-list --boundary --parents $local ^$parsed_sha1"));
1148                 my %local_ancestry;
1149                 foreach my $line (@local_ancestry) {
1150                         if (my ($child, $parents) = $line =~ m/^-?([a-f0-9]+) ([a-f0-9 ]+)/) {
1151                                 foreach my $parent (split(' ', $parents)) {
1152                                         $local_ancestry{$parent} = $child;
1153                                 }
1154                         } elsif (!$line =~ m/^([a-f0-9]+)/) {
1155                                 die "Unexpected output from git rev-list: $line";
1156                         }
1157                 }
1158                 while ($parsed_sha1 ne $HEAD_sha1) {
1159                         my $child = $local_ancestry{$parsed_sha1};
1160                         if (!$child) {
1161                                 printf STDERR "Cannot find a path in history from remote commit to last commit\n";
1162                                 return error_non_fast_forward($remote);
1163                         }
1164                         push(@commit_pairs, [$parsed_sha1, $child]);
1165                         $parsed_sha1 = $child;
1166                 }
1167         } else {
1168                 # No remote mediawiki revision. Export the whole
1169                 # history (linearized with --first-parent)
1170                 print STDERR "Warning: no common ancestor, pushing complete history\n";
1171                 my $history = run_git("rev-list --first-parent --children $local");
1172                 my @history = split(/\n/, $history);
1173                 @history = @history[1..$#history];
1174                 foreach my $line (reverse @history) {
1175                         my @commit_info_split = split(/ |\n/, $line);
1176                         push(@commit_pairs, \@commit_info_split);
1177                 }
1178         }
1179
1180         foreach my $commit_info_split (@commit_pairs) {
1181                 my $sha1_child = @{$commit_info_split}[0];
1182                 my $sha1_commit = @{$commit_info_split}[1];
1183                 my $diff_infos = run_git("diff-tree -r --raw -z $sha1_child $sha1_commit");
1184                 # TODO: we could detect rename, and encode them with a #redirect on the wiki.
1185                 # TODO: for now, it's just a delete+add
1186                 my @diff_info_list = split(/\0/, $diff_infos);
1187                 # Keep the subject line of the commit message as mediawiki comment for the revision
1188                 my $commit_msg = run_git("log --no-walk --format=\"%s\" $sha1_commit");
1189                 chomp($commit_msg);
1190                 # Push every blob
1191                 while (@diff_info_list) {
1192                         my $status;
1193                         # git diff-tree -z gives an output like
1194                         # <metadata>\0<filename1>\0
1195                         # <metadata>\0<filename2>\0
1196                         # and we've split on \0.
1197                         my $info = shift(@diff_info_list);
1198                         my $file = shift(@diff_info_list);
1199                         ($mw_revision, $status) = mw_push_file($info, $file, $commit_msg, $mw_revision);
1200                         if ($status eq "non-fast-forward") {
1201                                 # we may already have sent part of the
1202                                 # commit to MediaWiki, but it's too
1203                                 # late to cancel it. Stop the push in
1204                                 # the middle, but still give an
1205                                 # accurate error message.
1206                                 return error_non_fast_forward($remote);
1207                         }
1208                         if ($status ne "ok") {
1209                                 die("Unknown error from mw_push_file()");
1210                         }
1211                 }
1212                 unless ($dumb_push) {
1213                         run_git("notes --ref=$remotename/mediawiki add -f -m \"mediawiki_revision: $mw_revision\" $sha1_commit");
1214                         run_git("update-ref -m \"Git-MediaWiki push\" refs/mediawiki/$remotename/master $sha1_commit $sha1_child");
1215                 }
1216         }
1217
1218         print STDOUT "ok $remote\n";
1219         return 1;
1220 }
1221
1222 sub get_allowed_file_extensions {
1223         mw_connect_maybe();
1224
1225         my $query = {
1226                 action => 'query',
1227                 meta => 'siteinfo',
1228                 siprop => 'fileextensions'
1229                 };
1230         my $result = $mediawiki->api($query);
1231         my @file_extensions= map $_->{ext},@{$result->{query}->{fileextensions}};
1232         my %hashFile = map {$_ => 1}@file_extensions;
1233
1234         return %hashFile;
1235 }
1236
1237 # In memory cache for MediaWiki namespace ids.
1238 my %namespace_id;
1239
1240 # Namespaces whose id is cached in the configuration file
1241 # (to avoid duplicates)
1242 my %cached_mw_namespace_id;
1243
1244 # Return MediaWiki id for a canonical namespace name.
1245 # Ex.: "File", "Project".
1246 sub get_mw_namespace_id {
1247         mw_connect_maybe();
1248         my $name = shift;
1249
1250         if (!exists $namespace_id{$name}) {
1251                 # Look at configuration file, if the record for that namespace is
1252                 # already cached. Namespaces are stored in form:
1253                 # "Name_of_namespace:Id_namespace", ex.: "File:6".
1254                 my @temp = split(/[\n]/, run_git("config --get-all remote."
1255                                                 . $remotename .".namespaceCache"));
1256                 chomp(@temp);
1257                 foreach my $ns (@temp) {
1258                         my ($n, $id) = split(/:/, $ns);
1259                         if ($id eq 'notANameSpace') {
1260                                 $namespace_id{$n} = {is_namespace => 0};
1261                         } else {
1262                                 $namespace_id{$n} = {is_namespace => 1, id => $id};
1263                         }
1264                         $cached_mw_namespace_id{$n} = 1;
1265                 }
1266         }
1267
1268         if (!exists $namespace_id{$name}) {
1269                 print STDERR "Namespace $name not found in cache, querying the wiki ...\n";
1270                 # NS not found => get namespace id from MW and store it in
1271                 # configuration file.
1272                 my $query = {
1273                         action => 'query',
1274                         meta => 'siteinfo',
1275                         siprop => 'namespaces'
1276                 };
1277                 my $result = $mediawiki->api($query);
1278
1279                 while (my ($id, $ns) = each(%{$result->{query}->{namespaces}})) {
1280                         if (defined($ns->{id}) && defined($ns->{canonical})) {
1281                                 $namespace_id{$ns->{canonical}} = {is_namespace => 1, id => $ns->{id}};
1282                                 if ($ns->{'*'}) {
1283                                         # alias (e.g. french Fichier: as alias for canonical File:)
1284                                         $namespace_id{$ns->{'*'}} = {is_namespace => 1, id => $ns->{id}};
1285                                 }
1286                         }
1287                 }
1288         }
1289
1290         my $ns = $namespace_id{$name};
1291         my $id;
1292
1293         unless (defined $ns) {
1294                 print STDERR "No such namespace $name on MediaWiki.\n";
1295                 $ns = {is_namespace => 0};
1296                 $namespace_id{$name} = $ns;
1297         }
1298
1299         if ($ns->{is_namespace}) {
1300                 $id = $ns->{id};
1301         }
1302
1303         # Store "notANameSpace" as special value for inexisting namespaces
1304         my $store_id = ($id || 'notANameSpace');
1305
1306         # Store explicitely requested namespaces on disk
1307         if (!exists $cached_mw_namespace_id{$name}) {
1308                 run_git("config --add remote.". $remotename
1309                         .".namespaceCache \"". $name .":". $store_id ."\"");
1310                 $cached_mw_namespace_id{$name} = 1;
1311         }
1312         return $id;
1313 }
1314
1315 sub get_mw_namespace_id_for_page {
1316         if (my ($namespace) = $_[0] =~ /^([^:]*):/) {
1317                 return get_mw_namespace_id($namespace);
1318         } else {
1319                 return;
1320         }
1321 }