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
  19 # By default, use UTF-8 to communicate with Git and the user
  20 binmode STDERR, ":utf8";
  21 binmode STDOUT, ":utf8";
  22
  23 use URI::Escape;
  24 use IPC::Open2;
  25
  26 use warnings;
  27
  28 # Mediawiki filenames can contain forward slashes. This variable decides by which pattern they should be replaced
  29 use constant SLASH_REPLACEMENT => "%2F";
  30
  31 # It's not always possible to delete pages (may require some
  32 # privileges). Deleted pages are replaced with this content.
  33 use constant DELETED_CONTENT => "[[Category:Deleted]]\n";
  34
  35 # It's not possible to create empty pages. New empty files in Git are
  36 # sent with this content instead.
  37 use constant EMPTY_CONTENT => "<!-- empty page -->\n";
  38
  39 # used to reflect file creation or deletion in diff.
  40 use constant NULL_SHA1 => "0000000000000000000000000000000000000000";
  41
  42 # Used on Git's side to reflect empty edit messages on the wiki
  43 use constant EMPTY_MESSAGE => '*Empty MediaWiki Message*';
  44
  45 my $remotename = $ARGV[0];
  46 my $url = $ARGV[1];
  47
  48 # Accept both space-separated and multiple keys in config file.
  49 # Spaces should be written as _ anyway because we'll use chomp.
  50 my @tracked_pages = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".pages"));
  51 chomp(@tracked_pages);
  52
  53 # Just like @tracked_pages, but for MediaWiki categories.
  54 my @tracked_categories = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".categories"));
  55 chomp(@tracked_categories);
  56
  57 # Import media files on pull
  58 my $import_media = run_git("config --get --bool remote.". $remotename .".mediaimport");
  59 chomp($import_media);
  60 $import_media = ($import_media eq "true");
  61
  62 # Export media files on push
  63 my $export_media = run_git("config --get --bool remote.". $remotename .".mediaexport");
  64 chomp($export_media);
  65 $export_media = !($export_media eq "false");
  66
  67 my $wiki_login = run_git("config --get remote.". $remotename .".mwLogin");
  68 # Note: mwPassword is discourraged. Use the credential system instead.
  69 my $wiki_passwd = run_git("config --get remote.". $remotename .".mwPassword");
  70 my $wiki_domain = run_git("config --get remote.". $remotename .".mwDomain");
  71 chomp($wiki_login);
  72 chomp($wiki_passwd);
  73 chomp($wiki_domain);
  74
  75 # Import only last revisions (both for clone and fetch)
  76 my $shallow_import = run_git("config --get --bool remote.". $remotename .".shallow");
  77 chomp($shallow_import);
  78 $shallow_import = ($shallow_import eq "true");
  79
  80 # Fetch (clone and pull) by revisions instead of by pages. This behavior
  81 # is more efficient when we have a wiki with lots of pages and we fetch
  82 # the revisions quite often so that they concern only few pages.
  83 # Possible values:
  84 # - by_rev: perform one query per new revision on the remote wiki
  85 # - by_page: query each tracked page for new revision
  86 my $fetch_strategy = run_git("config --get remote.$remotename.fetchStrategy");
  87 unless ($fetch_strategy) {
  88         $fetch_strategy = run_git("config --get mediawiki.fetchStrategy");
  89 }
  90 chomp($fetch_strategy);
  91 unless ($fetch_strategy) {
  92         $fetch_strategy = "by_page";
  93 }
  94
  95 # Dumb push: don't update notes and mediawiki ref to reflect the last push.
  96 #
  97 # Configurable with mediawiki.dumbPush, or per-remote with
  98 # remote.<remotename>.dumbPush.
  99 #
 100 # This means the user will have to re-import the just-pushed
 101 # revisions. On the other hand, this means that the Git revisions
 102 # corresponding to MediaWiki revisions are all imported from the wiki,
 103 # regardless of whether they were initially created in Git or from the
 104 # web interface, hence all users will get the same history (i.e. if
 105 # the push from Git to MediaWiki loses some information, everybody
 106 # will get the history with information lost). If the import is
 107 # deterministic, this means everybody gets the same sha1 for each
 108 # MediaWiki revision.
 109 my $dumb_push = run_git("config --get --bool remote.$remotename.dumbPush");
 110 unless ($dumb_push) {
 111         $dumb_push = run_git("config --get --bool mediawiki.dumbPush");
 112 }
 113 chomp($dumb_push);
 114 $dumb_push = ($dumb_push eq "true");
 115
 116 my $wiki_name = $url;
 117 $wiki_name =~ s/[^\/]*:\/\///;
 118 # If URL is like http://user:password@example.com/, we clearly don't
 119 # want the password in $wiki_name. While we're there, also remove user
 120 # and '@' sign, to avoid author like MWUser@HTTPUser@host.com
 121 $wiki_name =~ s/^.*@//;
 122
 123 # Commands parser
 124 my $entry;
 125 my @cmd;
 126 while (<STDIN>) {
 127         chomp;
 128         @cmd = split(/ /);
 129         if (defined($cmd[0])) {
 130                 # Line not blank
 131                 if ($cmd[0] eq "capabilities") {
 132                         die("Too many arguments for capabilities") unless (!defined($cmd[1]));
 133                         mw_capabilities();
 134                 } elsif ($cmd[0] eq "list") {
 135                         die("Too many arguments for list") unless (!defined($cmd[2]));
 136                         mw_list($cmd[1]);
 137                 } elsif ($cmd[0] eq "import") {
 138                         die("Invalid arguments for import") unless ($cmd[1] ne "" && !defined($cmd[2]));
 139                         mw_import($cmd[1]);
 140                 } elsif ($cmd[0] eq "option") {
 141                         die("Too many arguments for option") unless ($cmd[1] ne "" && $cmd[2] ne "" && !defined($cmd[3]));
 142                         mw_option($cmd[1],$cmd[2]);
 143                 } elsif ($cmd[0] eq "push") {
 144                         mw_push($cmd[1]);
 145                 } else {
 146                         print STDERR "Unknown command. Aborting...\n";
 147                         last;
 148                 }
 149         } else {
 150                 # blank line: we should terminate
 151                 last;
 152         }
 153
 154         BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
 155                          # command is fully processed.
 156 }
 157
 158 ########################## Functions ##############################
 159
 160 # MediaWiki API instance, created lazily.
 161 my $mediawiki;
 162
 163 sub mw_connect_maybe {
 164         if ($mediawiki) {
 165                 return;
 166         }
 167         $mediawiki = MediaWiki::API->new;
 168         $mediawiki->{config}->{api_url} = "$url/api.php";
 169         if ($wiki_login) {
 170                 my %credential = (
 171                         'url' => $url,
 172                         'username' => $wiki_login,
 173                         'password' => $wiki_passwd
 174                 );
 175                 Git::credential(\%credential);
 176                 my $request = {lgname => $credential{username},
 177                                lgpassword => $credential{password},
 178                                lgdomain => $wiki_domain};
 179                 if ($mediawiki->login($request)) {
 180                         Git::credential(\%credential, 'approve');
 181                         print STDERR "Logged in mediawiki user \"$credential{username}\".\n";
 182                 } else {
 183                         print STDERR "Failed to log in mediawiki user \"$credential{username}\" on $url\n";
 184                         print STDERR "  (error " .
 185                                 $mediawiki->{error}->{code} . ': ' .
 186                                 $mediawiki->{error}->{details} . ")\n";
 187                         Git::credential(\%credential, 'reject');
 188                         exit 1;
 189                 }
 190         }
 191 }
 192
 193 sub fatal_mw_error {
 194         my $action = shift;
 195         print STDERR "fatal: could not $action.\n";
 196         print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 197         if ($url =~ /^https/) {
 198                 print STDERR "fatal: make sure '$url/api.php' is a valid page\n";
 199                 print STDERR "fatal: and the SSL certificate is correct.\n";
 200         } else {
 201                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 202         }
 203         print STDERR "fatal: (error " .
 204             $mediawiki->{error}->{code} . ': ' .
 205             $mediawiki->{error}->{details} . ")\n";
 206         exit 1;
 207 }
 208
 209 ## Functions for listing pages on the remote wiki
 210 sub get_mw_tracked_pages {
 211         my $pages = shift;
 212         get_mw_page_list(\@tracked_pages, $pages);
 213 }
 214
 215 sub get_mw_page_list {
 216         my $page_list = shift;
 217         my $pages = shift;
 218         my @some_pages = @$page_list;
 219         while (@some_pages) {
 220                 my $last = 50;
 221                 if ($#some_pages < $last) {
 222                         $last = $#some_pages;
 223                 }
 224                 my @slice = @some_pages[0..$last];
 225                 get_mw_first_pages(\@slice, $pages);
 226                 @some_pages = @some_pages[51..$#some_pages];
 227         }
 228 }
 229
 230 sub get_mw_tracked_categories {
 231         my $pages = shift;
 232         foreach my $category (@tracked_categories) {
 233                 if (index($category, ':') < 0) {
 234                         # Mediawiki requires the Category
 235                         # prefix, but let's not force the user
 236                         # to specify it.
 237                         $category = "Category:" . $category;
 238                 }
 239                 my $mw_pages = $mediawiki->list( {
 240                         action => 'query',
 241                         list => 'categorymembers',
 242                         cmtitle => $category,
 243                         cmlimit => 'max' } )
 244                         || die $mediawiki->{error}->{code} . ': '
 245                                 . $mediawiki->{error}->{details};
 246                 foreach my $page (@{$mw_pages}) {
 247                         $pages->{$page->{title}} = $page;
 248                 }
 249         }
 250 }
 251
 252 sub get_mw_all_pages {
 253         my $pages = shift;
 254         # No user-provided list, get the list of pages from the API.
 255         my $mw_pages = $mediawiki->list({
 256                 action => 'query',
 257                 list => 'allpages',
 258                 aplimit => 'max'
 259         });
 260         if (!defined($mw_pages)) {
 261                 fatal_mw_error("get the list of wiki pages");
 262         }
 263         foreach my $page (@{$mw_pages}) {
 264                 $pages->{$page->{title}} = $page;
 265         }
 266 }
 267
 268 # queries the wiki for a set of pages. Meant to be used within a loop
 269 # querying the wiki for slices of page list.
 270 sub get_mw_first_pages {
 271         my $some_pages = shift;
 272         my @some_pages = @{$some_pages};
 273
 274         my $pages = shift;
 275
 276         # pattern 'page1|page2|...' required by the API
 277         my $titles = join('|', @some_pages);
 278
 279         my $mw_pages = $mediawiki->api({
 280                 action => 'query',
 281                 titles => $titles,
 282         });
 283         if (!defined($mw_pages)) {
 284                 fatal_mw_error("query the list of wiki pages");
 285         }
 286         while (my ($id, $page) = each(%{$mw_pages->{query}->{pages}})) {
 287                 if ($id < 0) {
 288                         print STDERR "Warning: page $page->{title} not found on wiki\n";
 289                 } else {
 290                         $pages->{$page->{title}} = $page;
 291                 }
 292         }
 293 }
 294
 295 # Get the list of pages to be fetched according to configuration.
 296 sub get_mw_pages {
 297         mw_connect_maybe();
 298
 299         print STDERR "Listing pages on remote wiki...\n";
 300
 301         my %pages; # hash on page titles to avoid duplicates
 302         my $user_defined;
 303         if (@tracked_pages) {
 304                 $user_defined = 1;
 305                 # The user provided a list of pages titles, but we
 306                 # still need to query the API to get the page IDs.
 307                 get_mw_tracked_pages(\%pages);
 308         }
 309         if (@tracked_categories) {
 310                 $user_defined = 1;
 311                 get_mw_tracked_categories(\%pages);
 312         }
 313         if (!$user_defined) {
 314                 get_mw_all_pages(\%pages);
 315         }
 316         if ($import_media) {
 317                 print STDERR "Getting media files for selected pages...\n";
 318                 if ($user_defined) {
 319                         get_linked_mediafiles(\%pages);
 320                 } else {
 321                         get_all_mediafiles(\%pages);
 322                 }
 323         }
 324         print STDERR (scalar keys %pages) . " pages found.\n";
 325         return %pages;
 326 }
 327
 328 # usage: $out = run_git("command args");
 329 #        $out = run_git("command args", "raw"); # don't interpret output as UTF-8.
 330 sub run_git {
 331         my $args = shift;
 332         my $encoding = (shift || "encoding(UTF-8)");
 333         open(my $git, "-|:$encoding", "git " . $args);
 334         my $res = do { local $/; <$git> };
 335         close($git);
 336
 337         return $res;
 338 }
 339
 340
 341 sub get_all_mediafiles {
 342         my $pages = shift;
 343         # Attach list of all pages for media files from the API,
 344         # they are in a different namespace, only one namespace
 345         # can be queried at the same moment
 346         my $mw_pages = $mediawiki->list({
 347                 action => 'query',
 348                 list => 'allpages',
 349                 apnamespace => get_mw_namespace_id("File"),
 350                 aplimit => 'max'
 351         });
 352         if (!defined($mw_pages)) {
 353                 print STDERR "fatal: could not get the list of pages for media files.\n";
 354                 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 355                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 356                 exit 1;
 357         }
 358         foreach my $page (@{$mw_pages}) {
 359                 $pages->{$page->{title}} = $page;
 360         }
 361 }
 362
 363 sub get_linked_mediafiles {
 364         my $pages = shift;
 365         my @titles = map $_->{title}, values(%{$pages});
 366
 367         # The query is split in small batches because of the MW API limit of
 368         # the number of links to be returned (500 links max).
 369         my $batch = 10;
 370         while (@titles) {
 371                 if ($#titles < $batch) {
 372                         $batch = $#titles;
 373                 }
 374                 my @slice = @titles[0..$batch];
 375
 376                 # pattern 'page1|page2|...' required by the API
 377                 my $mw_titles = join('|', @slice);
 378
 379                 # Media files could be included or linked from
 380                 # a page, get all related
 381                 my $query = {
 382                         action => 'query',
 383                         prop => 'links|images',
 384                         titles => $mw_titles,
 385                         plnamespace => get_mw_namespace_id("File"),
 386                         pllimit => 'max'
 387                 };
 388                 my $result = $mediawiki->api($query);
 389
 390                 while (my ($id, $page) = each(%{$result->{query}->{pages}})) {
 391                         my @media_titles;
 392                         if (defined($page->{links})) {
 393                                 my @link_titles = map $_->{title}, @{$page->{links}};
 394                                 push(@media_titles, @link_titles);
 395                         }
 396                         if (defined($page->{images})) {
 397                                 my @image_titles = map $_->{title}, @{$page->{images}};
 398                                 push(@media_titles, @image_titles);
 399                         }
 400                         if (@media_titles) {
 401                                 get_mw_page_list(\@media_titles, $pages);
 402                         }
 403                 }
 404
 405                 @titles = @titles[($batch+1)..$#titles];
 406         }
 407 }
 408
 409 sub get_mw_mediafile_for_page_revision {
 410         # Name of the file on Wiki, with the prefix.
 411         my $filename = shift;
 412         my $timestamp = shift;
 413         my %mediafile;
 414
 415         # Search if on a media file with given timestamp exists on
 416         # MediaWiki. In that case download the file.
 417         my $query = {
 418                 action => 'query',
 419                 prop => 'imageinfo',
 420                 titles => "File:" . $filename,
 421                 iistart => $timestamp,
 422                 iiend => $timestamp,
 423                 iiprop => 'timestamp|archivename|url',
 424                 iilimit => 1
 425         };
 426         my $result = $mediawiki->api($query);
 427
 428         my ($fileid, $file) = each( %{$result->{query}->{pages}} );
 429         # If not defined it means there is no revision of the file for
 430         # given timestamp.
 431         if (defined($file->{imageinfo})) {
 432                 $mediafile{title} = $filename;
 433
 434                 my $fileinfo = pop(@{$file->{imageinfo}});
 435                 $mediafile{timestamp} = $fileinfo->{timestamp};
 436                 # Mediawiki::API's download function doesn't support https URLs
 437                 # and can't download old versions of files.
 438                 print STDERR "\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
 439                 $mediafile{content} = download_mw_mediafile($fileinfo->{url});
 440         }
 441         return %mediafile;
 442 }
 443
 444 sub download_mw_mediafile {
 445         my $url = shift;
 446
 447         my $response = $mediawiki->{ua}->get($url);
 448         if ($response->code == 200) {
 449                 return $response->decoded_content;
 450         } else {
 451                 print STDERR "Error downloading mediafile from :\n";
 452                 print STDERR "URL: $url\n";
 453                 print STDERR "Server response: " . $response->code . " " . $response->message . "\n";
 454                 exit 1;
 455         }
 456 }
 457
 458 sub get_last_local_revision {
 459         # Get note regarding last mediawiki revision
 460         my $note = run_git("notes --ref=$remotename/mediawiki show refs/mediawiki/$remotename/master 2>/dev/null");
 461         my @note_info = split(/ /, $note);
 462
 463         my $lastrevision_number;
 464         if (!(defined($note_info[0]) && $note_info[0] eq "mediawiki_revision:")) {
 465                 print STDERR "No previous mediawiki revision found";
 466                 $lastrevision_number = 0;
 467         } else {
 468                 # Notes are formatted : mediawiki_revision: #number
 469                 $lastrevision_number = $note_info[1];
 470                 chomp($lastrevision_number);
 471                 print STDERR "Last local mediawiki revision found is $lastrevision_number";
 472         }
 473         return $lastrevision_number;
 474 }
 475
 476 # Remember the timestamp corresponding to a revision id.
 477 my %basetimestamps;
 478
 479 # Get the last remote revision without taking in account which pages are
 480 # tracked or not. This function makes a single request to the wiki thus
 481 # avoid a loop onto all tracked pages. This is useful for the fetch-by-rev
 482 # option.
 483 sub get_last_global_remote_rev {
 484         mw_connect_maybe();
 485
 486         my $query = {
 487                 action => 'query',
 488                 list => 'recentchanges',
 489                 prop => 'revisions',
 490                 rclimit => '1',
 491                 rcdir => 'older',
 492         };
 493         my $result = $mediawiki->api($query);
 494         return $result->{query}->{recentchanges}[0]->{revid};
 495 }
 496
 497 # Get the last remote revision concerning the tracked pages and the tracked
 498 # categories.
 499 sub get_last_remote_revision {
 500         mw_connect_maybe();
 501
 502         my %pages_hash = get_mw_pages();
 503         my @pages = values(%pages_hash);
 504
 505         my $max_rev_num = 0;
 506
 507         print STDERR "Getting last revision id on tracked pages...\n";
 508
 509         foreach my $page (@pages) {
 510                 my $id = $page->{pageid};
 511
 512                 my $query = {
 513                         action => 'query',
 514                         prop => 'revisions',
 515                         rvprop => 'ids|timestamp',
 516                         pageids => $id,
 517                 };
 518
 519                 my $result = $mediawiki->api($query);
 520
 521                 my $lastrev = pop(@{$result->{query}->{pages}->{$id}->{revisions}});
 522
 523                 $basetimestamps{$lastrev->{revid}} = $lastrev->{timestamp};
 524
 525                 $max_rev_num = ($lastrev->{revid} > $max_rev_num ? $lastrev->{revid} : $max_rev_num);
 526         }
 527
 528         print STDERR "Last remote revision found is $max_rev_num.\n";
 529         return $max_rev_num;
 530 }
 531
 532 # Clean content before sending it to MediaWiki
 533 sub mediawiki_clean {
 534         my $string = shift;
 535         my $page_created = shift;
 536         # Mediawiki does not allow blank space at the end of a page and ends with a single \n.
 537         # This function right trims a string and adds a \n at the end to follow this rule
 538         $string =~ s/\s+$//;
 539         if ($string eq "" && $page_created) {
 540                 # Creating empty pages is forbidden.
 541                 $string = EMPTY_CONTENT;
 542         }
 543         return $string."\n";
 544 }
 545
 546 # Filter applied on MediaWiki data before adding them to Git
 547 sub mediawiki_smudge {
 548         my $string = shift;
 549         if ($string eq EMPTY_CONTENT) {
 550                 $string = "";
 551         }
 552         # This \n is important. This is due to mediawiki's way to handle end of files.
 553         return $string."\n";
 554 }
 555
 556 sub mediawiki_clean_filename {
 557         my $filename = shift;
 558         $filename =~ s/@{[SLASH_REPLACEMENT]}/\//g;
 559         # [, ], |, {, and } are forbidden by MediaWiki, even URL-encoded.
 560         # Do a variant of URL-encoding, i.e. looks like URL-encoding,
 561         # but with _ added to prevent MediaWiki from thinking this is
 562         # an actual special character.
 563         $filename =~ s/[\[\]\{\}\|]/sprintf("_%%_%x", ord($&))/ge;
 564         # If we use the uri escape before
 565         # we should unescape here, before anything
 566
 567         return $filename;
 568 }
 569
 570 sub mediawiki_smudge_filename {
 571         my $filename = shift;
 572         $filename =~ s/\//@{[SLASH_REPLACEMENT]}/g;
 573         $filename =~ s/ /_/g;
 574         # Decode forbidden characters encoded in mediawiki_clean_filename
 575         $filename =~ s/_%_([0-9a-fA-F][0-9a-fA-F])/sprintf("%c", hex($1))/ge;
 576         return $filename;
 577 }
 578
 579 sub literal_data {
 580         my ($content) = @_;
 581         print STDOUT "data ", bytes::length($content), "\n", $content;
 582 }
 583
 584 sub literal_data_raw {
 585         # Output possibly binary content.
 586         my ($content) = @_;
 587         # Avoid confusion between size in bytes and in characters
 588         utf8::downgrade($content);
 589         binmode STDOUT, ":raw";
 590         print STDOUT "data ", bytes::length($content), "\n", $content;
 591         binmode STDOUT, ":utf8";
 592 }
 593
 594 sub mw_capabilities {
 595         # Revisions are imported to the private namespace
 596         # refs/mediawiki/$remotename/ by the helper and fetched into
 597         # refs/remotes/$remotename later by fetch.
 598         print STDOUT "refspec refs/heads/*:refs/mediawiki/$remotename/*\n";
 599         print STDOUT "import\n";
 600         print STDOUT "list\n";
 601         print STDOUT "push\n";
 602         print STDOUT "\n";
 603 }
 604
 605 sub mw_list {
 606         # MediaWiki do not have branches, we consider one branch arbitrarily
 607         # called master, and HEAD pointing to it.
 608         print STDOUT "? refs/heads/master\n";
 609         print STDOUT "\@refs/heads/master HEAD\n";
 610         print STDOUT "\n";
 611 }
 612
 613 sub mw_option {
 614         print STDERR "remote-helper command 'option $_[0]' not yet implemented\n";
 615         print STDOUT "unsupported\n";
 616 }
 617
 618 sub fetch_mw_revisions_for_page {
 619         my $page = shift;
 620         my $id = shift;
 621         my $fetch_from = shift;
 622         my @page_revs = ();
 623         my $query = {
 624                 action => 'query',
 625                 prop => 'revisions',
 626                 rvprop => 'ids',
 627                 rvdir => 'newer',
 628                 rvstartid => $fetch_from,
 629                 rvlimit => 500,
 630                 pageids => $id,
 631         };
 632
 633         my $revnum = 0;
 634         # Get 500 revisions at a time due to the mediawiki api limit
 635         while (1) {
 636                 my $result = $mediawiki->api($query);
 637
 638                 # Parse each of those 500 revisions
 639                 foreach my $revision (@{$result->{query}->{pages}->{$id}->{revisions}}) {
 640                         my $page_rev_ids;
 641                         $page_rev_ids->{pageid} = $page->{pageid};
 642                         $page_rev_ids->{revid} = $revision->{revid};
 643                         push(@page_revs, $page_rev_ids);
 644                         $revnum++;
 645                 }
 646                 last unless $result->{'query-continue'};
 647                 $query->{rvstartid} = $result->{'query-continue'}->{revisions}->{rvstartid};
 648         }
 649         if ($shallow_import && @page_revs) {
 650                 print STDERR "  Found 1 revision (shallow import).\n";
 651                 @page_revs = sort {$b->{revid} <=> $a->{revid}} (@page_revs);
 652                 return $page_revs[0];
 653         }
 654         print STDERR "  Found ", $revnum, " revision(s).\n";
 655         return @page_revs;
 656 }
 657
 658 sub fetch_mw_revisions {
 659         my $pages = shift; my @pages = @{$pages};
 660         my $fetch_from = shift;
 661
 662         my @revisions = ();
 663         my $n = 1;
 664         foreach my $page (@pages) {
 665                 my $id = $page->{pageid};
 666
 667                 print STDERR "page $n/", scalar(@pages), ": ". $page->{title} ."\n";
 668                 $n++;
 669                 my @page_revs = fetch_mw_revisions_for_page($page, $id, $fetch_from);
 670                 @revisions = (@page_revs, @revisions);
 671         }
 672
 673         return ($n, @revisions);
 674 }
 675
 676 sub fe_escape_path {
 677     my $path = shift;
 678     $path =~ s/\\/\\\\/g;
 679     $path =~ s/"/\\"/g;
 680     $path =~ s/\n/\\n/g;
 681     return '"' . $path . '"';
 682 }
 683
 684 sub import_file_revision {
 685         my $commit = shift;
 686         my %commit = %{$commit};
 687         my $full_import = shift;
 688         my $n = shift;
 689         my $mediafile = shift;
 690         my %mediafile;
 691         if ($mediafile) {
 692                 %mediafile = %{$mediafile};
 693         }
 694
 695         my $title = $commit{title};
 696         my $comment = $commit{comment};
 697         my $content = $commit{content};
 698         my $author = $commit{author};
 699         my $date = $commit{date};
 700
 701         print STDOUT "commit refs/mediawiki/$remotename/master\n";
 702         print STDOUT "mark :$n\n";
 703         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 704         literal_data($comment);
 705
 706         # If it's not a clone, we need to know where to start from
 707         if (!$full_import && $n == 1) {
 708                 print STDOUT "from refs/mediawiki/$remotename/master^0\n";
 709         }
 710         if ($content ne DELETED_CONTENT) {
 711                 print STDOUT "M 644 inline " .
 712                     fe_escape_path($title . ".mw") . "\n";
 713                 literal_data($content);
 714                 if (%mediafile) {
 715                         print STDOUT "M 644 inline "
 716                             . fe_escape_path($mediafile{title}) . "\n";
 717                         literal_data_raw($mediafile{content});
 718                 }
 719                 print STDOUT "\n\n";
 720         } else {
 721                 print STDOUT "D " . fe_escape_path($title . ".mw") . "\n";
 722         }
 723
 724         # mediawiki revision number in the git note
 725         if ($full_import && $n == 1) {
 726                 print STDOUT "reset refs/notes/$remotename/mediawiki\n";
 727         }
 728         print STDOUT "commit refs/notes/$remotename/mediawiki\n";
 729         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 730         literal_data("Note added by git-mediawiki during import");
 731         if (!$full_import && $n == 1) {
 732                 print STDOUT "from refs/notes/$remotename/mediawiki^0\n";
 733         }
 734         print STDOUT "N inline :$n\n";
 735         literal_data("mediawiki_revision: " . $commit{mw_revision});
 736         print STDOUT "\n\n";
 737 }
 738
 739 # parse a sequence of
 740 # <cmd> <arg1>
 741 # <cmd> <arg2>
 742 # \n
 743 # (like batch sequence of import and sequence of push statements)
 744 sub get_more_refs {
 745         my $cmd = shift;
 746         my @refs;
 747         while (1) {
 748                 my $line = <STDIN>;
 749                 if ($line =~ m/^$cmd (.*)$/) {
 750                         push(@refs, $1);
 751                 } elsif ($line eq "\n") {
 752                         return @refs;
 753                 } else {
 754                         die("Invalid command in a '$cmd' batch: ". $_);
 755                 }
 756         }
 757 }
 758
 759 sub mw_import {
 760         # multiple import commands can follow each other.
 761         my @refs = (shift, get_more_refs("import"));
 762         foreach my $ref (@refs) {
 763                 mw_import_ref($ref);
 764         }
 765         print STDOUT "done\n";
 766 }
 767
 768 sub mw_import_ref {
 769         my $ref = shift;
 770         # The remote helper will call "import HEAD" and
 771         # "import refs/heads/master".
 772         # Since HEAD is a symbolic ref to master (by convention,
 773         # followed by the output of the command "list" that we gave),
 774         # we don't need to do anything in this case.
 775         if ($ref eq "HEAD") {
 776                 return;
 777         }
 778
 779         mw_connect_maybe();
 780
 781         print STDERR "Searching revisions...\n";
 782         my $last_local = get_last_local_revision();
 783         my $fetch_from = $last_local + 1;
 784         if ($fetch_from == 1) {
 785                 print STDERR ", fetching from beginning.\n";
 786         } else {
 787                 print STDERR ", fetching from here.\n";
 788         }
 789
 790         my $n = 0;
 791         if ($fetch_strategy eq "by_rev") {
 792                 print STDERR "Fetching & writing export data by revs...\n";
 793                 $n = mw_import_ref_by_revs($fetch_from);
 794         } elsif ($fetch_strategy eq "by_page") {
 795                 print STDERR "Fetching & writing export data by pages...\n";
 796                 $n = mw_import_ref_by_pages($fetch_from);
 797         } else {
 798                 print STDERR "fatal: invalid fetch strategy \"$fetch_strategy\".\n";
 799                 print STDERR "Check your configuration variables remote.$remotename.fetchStrategy and mediawiki.fetchStrategy\n";
 800                 exit 1;
 801         }
 802
 803         if ($fetch_from == 1 && $n == 0) {
 804                 print STDERR "You appear to have cloned an empty MediaWiki.\n";
 805                 # Something has to be done remote-helper side. If nothing is done, an error is
 806                 # thrown saying that HEAD is referring to unknown object 0000000000000000000
 807                 # and the clone fails.
 808         }
 809 }
 810
 811 sub mw_import_ref_by_pages {
 812
 813         my $fetch_from = shift;
 814         my %pages_hash = get_mw_pages();
 815         my @pages = values(%pages_hash);
 816
 817         my ($n, @revisions) = fetch_mw_revisions(\@pages, $fetch_from);
 818
 819         @revisions = sort {$a->{revid} <=> $b->{revid}} @revisions;
 820         my @revision_ids = map $_->{revid}, @revisions;
 821
 822         return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
 823 }
 824
 825 sub mw_import_ref_by_revs {
 826
 827         my $fetch_from = shift;
 828         my %pages_hash = get_mw_pages();
 829
 830         my $last_remote = get_last_global_remote_rev();
 831         my @revision_ids = $fetch_from..$last_remote;
 832         return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
 833 }
 834
 835 # Import revisions given in second argument (array of integers).
 836 # Only pages appearing in the third argument (hash indexed by page titles)
 837 # will be imported.
 838 sub mw_import_revids {
 839         my $fetch_from = shift;
 840         my $revision_ids = shift;
 841         my $pages = shift;
 842
 843         my $n = 0;
 844         my $n_actual = 0;
 845         my $last_timestamp = 0; # Placeholer in case $rev->timestamp is undefined
 846
 847         foreach my $pagerevid (@$revision_ids) {
 848                 # Count page even if we skip it, since we display
 849                 # $n/$total and $total includes skipped pages.
 850                 $n++;
 851
 852                 # fetch the content of the pages
 853                 my $query = {
 854                         action => 'query',
 855                         prop => 'revisions',
 856                         rvprop => 'content|timestamp|comment|user|ids',
 857                         revids => $pagerevid,
 858                 };
 859
 860                 my $result = $mediawiki->api($query);
 861
 862                 if (!$result) {
 863                         die "Failed to retrieve modified page for revision $pagerevid";
 864                 }
 865
 866                 if (defined($result->{query}->{badrevids}->{$pagerevid})) {
 867                         # The revision id does not exist on the remote wiki.
 868                         next;
 869                 }
 870
 871                 if (!defined($result->{query}->{pages})) {
 872                         die "Invalid revision $pagerevid.";
 873                 }
 874
 875                 my @result_pages = values(%{$result->{query}->{pages}});
 876                 my $result_page = $result_pages[0];
 877                 my $rev = $result_pages[0]->{revisions}->[0];
 878
 879                 my $page_title = $result_page->{title};
 880
 881                 if (!exists($pages->{$page_title})) {
 882                         print STDERR "$n/", scalar(@$revision_ids),
 883                                 ": Skipping revision #$rev->{revid} of $page_title\n";
 884                         next;
 885                 }
 886
 887                 $n_actual++;
 888
 889                 my %commit;
 890                 $commit{author} = $rev->{user} || 'Anonymous';
 891                 $commit{comment} = $rev->{comment} || EMPTY_MESSAGE;
 892                 $commit{title} = mediawiki_smudge_filename($page_title);
 893                 $commit{mw_revision} = $rev->{revid};
 894                 $commit{content} = mediawiki_smudge($rev->{'*'});
 895
 896                 if (!defined($rev->{timestamp})) {
 897                         $last_timestamp++;
 898                 } else {
 899                         $last_timestamp = $rev->{timestamp};
 900                 }
 901                 $commit{date} = DateTime::Format::ISO8601->parse_datetime($last_timestamp);
 902
 903                 # Differentiates classic pages and media files.
 904                 my ($namespace, $filename) = $page_title =~ /^([^:]*):(.*)$/;
 905                 my %mediafile;
 906                 if ($namespace) {
 907                         my $id = get_mw_namespace_id($namespace);
 908                         if ($id && $id == get_mw_namespace_id("File")) {
 909                                 %mediafile = get_mw_mediafile_for_page_revision($filename, $rev->{timestamp});
 910                         }
 911                 }
 912                 # If this is a revision of the media page for new version
 913                 # of a file do one common commit for both file and media page.
 914                 # Else do commit only for that page.
 915                 print STDERR "$n/", scalar(@$revision_ids), ": Revision #$rev->{revid} of $commit{title}\n";
 916                 import_file_revision(\%commit, ($fetch_from == 1), $n_actual, \%mediafile);
 917         }
 918
 919         return $n_actual;
 920 }
 921
 922 sub error_non_fast_forward {
 923         my $advice = run_git("config --bool advice.pushNonFastForward");
 924         chomp($advice);
 925         if ($advice ne "false") {
 926                 # Native git-push would show this after the summary.
 927                 # We can't ask it to display it cleanly, so print it
 928                 # ourselves before.
 929                 print STDERR "To prevent you from losing history, non-fast-forward updates were rejected\n";
 930                 print STDERR "Merge the remote changes (e.g. 'git pull') before pushing again. See the\n";
 931                 print STDERR "'Note about fast-forwards' section of 'git push --help' for details.\n";
 932         }
 933         print STDOUT "error $_[0] \"non-fast-forward\"\n";
 934         return 0;
 935 }
 936
 937 sub mw_upload_file {
 938         my $complete_file_name = shift;
 939         my $new_sha1 = shift;
 940         my $extension = shift;
 941         my $file_deleted = shift;
 942         my $summary = shift;
 943         my $newrevid;
 944         my $path = "File:" . $complete_file_name;
 945         my %hashFiles = get_allowed_file_extensions();
 946         if (!exists($hashFiles{$extension})) {
 947                 print STDERR "$complete_file_name is not a permitted file on this wiki.\n";
 948                 print STDERR "Check the configuration of file uploads in your mediawiki.\n";
 949                 return $newrevid;
 950         }
 951         # Deleting and uploading a file requires a priviledged user
 952         if ($file_deleted) {
 953                 mw_connect_maybe();
 954                 my $query = {
 955                         action => 'delete',
 956                         title => $path,
 957                         reason => $summary
 958                 };
 959                 if (!$mediawiki->edit($query)) {
 960                         print STDERR "Failed to delete file on remote wiki\n";
 961                         print STDERR "Check your permissions on the remote site. Error code:\n";
 962                         print STDERR $mediawiki->{error}->{code} . ':' . $mediawiki->{error}->{details};
 963                         exit 1;
 964                 }
 965         } else {
 966                 # Don't let perl try to interpret file content as UTF-8 => use "raw"
 967                 my $content = run_git("cat-file blob $new_sha1", "raw");
 968                 if ($content ne "") {
 969                         mw_connect_maybe();
 970                         $mediawiki->{config}->{upload_url} =
 971                                 "$url/index.php/Special:Upload";
 972                         $mediawiki->edit({
 973                                 action => 'upload',
 974                                 filename => $complete_file_name,
 975                                 comment => $summary,
 976                                 file => [undef,
 977                                          $complete_file_name,
 978                                          Content => $content],
 979                                 ignorewarnings => 1,
 980                         }, {
 981                                 skip_encoding => 1
 982                         } ) || die $mediawiki->{error}->{code} . ':'
 983                                  . $mediawiki->{error}->{details};
 984                         my $last_file_page = $mediawiki->get_page({title => $path});
 985                         $newrevid = $last_file_page->{revid};
 986                         print STDERR "Pushed file: $new_sha1 - $complete_file_name.\n";
 987                 } else {
 988                         print STDERR "Empty file $complete_file_name not pushed.\n";
 989                 }
 990         }
 991         return $newrevid;
 992 }
 993
 994 sub mw_push_file {
 995         my $diff_info = shift;
 996         # $diff_info contains a string in this format:
 997         # 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status>
 998         my @diff_info_split = split(/[ \t]/, $diff_info);
 999
1000         # Filename, including .mw extension
1001         my $complete_file_name = shift;
1002         # Commit message
1003         my $summary = shift;
1004         # MediaWiki revision number. Keep the previous one by default,
1005         # in case there's no edit to perform.
1006         my $oldrevid = shift;
1007         my $newrevid;
1008
1009         if ($summary eq EMPTY_MESSAGE) {
1010                 $summary = '';
1011         }
1012
1013         my $new_sha1 = $diff_info_split[3];
1014         my $old_sha1 = $diff_info_split[2];
1015         my $page_created = ($old_sha1 eq NULL_SHA1);
1016         my $page_deleted = ($new_sha1 eq NULL_SHA1);
1017         $complete_file_name = mediawiki_clean_filename($complete_file_name);
1018
1019         my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/;
1020         if (!defined($extension)) {
1021                 $extension = "";
1022         }
1023         if ($extension eq "mw") {
1024                 my $ns = get_mw_namespace_id_for_page($complete_file_name);
1025                 if ($ns && $ns == get_mw_namespace_id("File") && (!$export_media)) {
1026                         print STDERR "Ignoring media file related page: $complete_file_name\n";
1027                         return ($oldrevid, "ok");
1028                 }
1029                 my $file_content;
1030                 if ($page_deleted) {
1031                         # Deleting a page usually requires
1032                         # special privileges. A common
1033                         # convention is to replace the page
1034                         # with this content instead:
1035                         $file_content = DELETED_CONTENT;
1036                 } else {
1037                         $file_content = run_git("cat-file blob $new_sha1");
1038                 }
1039
1040                 mw_connect_maybe();
1041
1042                 my $result = $mediawiki->edit( {
1043                         action => 'edit',
1044                         summary => $summary,
1045                         title => $title,
1046                         basetimestamp => $basetimestamps{$oldrevid},
1047                         text => mediawiki_clean($file_content, $page_created),
1048                                   }, {
1049                                           skip_encoding => 1 # Helps with names with accentuated characters
1050                                   });
1051                 if (!$result) {
1052                         if ($mediawiki->{error}->{code} == 3) {
1053                                 # edit conflicts, considered as non-fast-forward
1054                                 print STDERR 'Warning: Error ' .
1055                                     $mediawiki->{error}->{code} .
1056                                     ' from mediwiki: ' . $mediawiki->{error}->{details} .
1057                                     ".\n";
1058                                 return ($oldrevid, "non-fast-forward");
1059                         } else {
1060                                 # Other errors. Shouldn't happen => just die()
1061                                 die 'Fatal: Error ' .
1062                                     $mediawiki->{error}->{code} .
1063                                     ' from mediwiki: ' . $mediawiki->{error}->{details};
1064                         }
1065                 }
1066                 $newrevid = $result->{edit}->{newrevid};
1067                 print STDERR "Pushed file: $new_sha1 - $title\n";
1068         } elsif ($export_media) {
1069                 $newrevid = mw_upload_file($complete_file_name, $new_sha1,
1070                                            $extension, $page_deleted,
1071                                            $summary);
1072         } else {
1073                 print STDERR "Ignoring media file $title\n";
1074         }
1075         $newrevid = ($newrevid or $oldrevid);
1076         return ($newrevid, "ok");
1077 }
1078
1079 sub mw_push {
1080         # multiple push statements can follow each other
1081         my @refsspecs = (shift, get_more_refs("push"));
1082         my $pushed;
1083         for my $refspec (@refsspecs) {
1084                 my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/
1085                     or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>");
1086                 if ($force) {
1087                         print STDERR "Warning: forced push not allowed on a MediaWiki.\n";
1088                 }
1089                 if ($local eq "") {
1090                         print STDERR "Cannot delete remote branch on a MediaWiki\n";
1091                         print STDOUT "error $remote cannot delete\n";
1092                         next;
1093                 }
1094                 if ($remote ne "refs/heads/master") {
1095                         print STDERR "Only push to the branch 'master' is supported on a MediaWiki\n";
1096                         print STDOUT "error $remote only master allowed\n";
1097                         next;
1098                 }
1099                 if (mw_push_revision($local, $remote)) {
1100                         $pushed = 1;
1101                 }
1102         }
1103
1104         # Notify Git that the push is done
1105         print STDOUT "\n";
1106
1107         if ($pushed && $dumb_push) {
1108                 print STDERR "Just pushed some revisions to MediaWiki.\n";
1109                 print STDERR "The pushed revisions now have to be re-imported, and your current branch\n";
1110                 print STDERR "needs to be updated with these re-imported commits. You can do this with\n";
1111                 print STDERR "\n";
1112                 print STDERR "  git pull --rebase\n";
1113                 print STDERR "\n";
1114         }
1115 }
1116
1117 sub mw_push_revision {
1118         my $local = shift;
1119         my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
1120         my $last_local_revid = get_last_local_revision();
1121         print STDERR ".\n"; # Finish sentence started by get_last_local_revision()
1122         my $last_remote_revid = get_last_remote_revision();
1123         my $mw_revision = $last_remote_revid;
1124
1125         # Get sha1 of commit pointed by local HEAD
1126         my $HEAD_sha1 = run_git("rev-parse $local 2>/dev/null"); chomp($HEAD_sha1);
1127         # Get sha1 of commit pointed by remotes/$remotename/master
1128         my $remoteorigin_sha1 = run_git("rev-parse refs/remotes/$remotename/master 2>/dev/null");
1129         chomp($remoteorigin_sha1);
1130
1131         if ($last_local_revid > 0 &&
1132             $last_local_revid < $last_remote_revid) {
1133                 return error_non_fast_forward($remote);
1134         }
1135
1136         if ($HEAD_sha1 eq $remoteorigin_sha1) {
1137                 # nothing to push
1138                 return 0;
1139         }
1140
1141         # Get every commit in between HEAD and refs/remotes/origin/master,
1142         # including HEAD and refs/remotes/origin/master
1143         my @commit_pairs = ();
1144         if ($last_local_revid > 0) {
1145                 my $parsed_sha1 = $remoteorigin_sha1;
1146                 # Find a path from last MediaWiki commit to pushed commit
1147                 print STDERR "Computing path from local to remote ...\n";
1148                 my @local_ancestry = split(/\n/, run_git("rev-list --boundary --parents $local ^$parsed_sha1"));
1149                 my %local_ancestry;
1150                 foreach my $line (@local_ancestry) {
1151                         if (my ($child, $parents) = $line =~ m/^-?([a-f0-9]+) ([a-f0-9 ]+)/) {
1152                                 foreach my $parent (split(' ', $parents)) {
1153                                         $local_ancestry{$parent} = $child;
1154                                 }
1155                         } elsif (!$line =~ m/^([a-f0-9]+)/) {
1156                                 die "Unexpected output from git rev-list: $line";
1157                         }
1158                 }
1159                 while ($parsed_sha1 ne $HEAD_sha1) {
1160                         my $child = $local_ancestry{$parsed_sha1};
1161                         if (!$child) {
1162                                 printf STDERR "Cannot find a path in history from remote commit to last commit\n";
1163                                 return error_non_fast_forward($remote);
1164                         }
1165                         push(@commit_pairs, [$parsed_sha1, $child]);
1166                         $parsed_sha1 = $child;
1167                 }
1168         } else {
1169                 # No remote mediawiki revision. Export the whole
1170                 # history (linearized with --first-parent)
1171                 print STDERR "Warning: no common ancestor, pushing complete history\n";
1172                 my $history = run_git("rev-list --first-parent --children $local");
1173                 my @history = split('\n', $history);
1174                 @history = @history[1..$#history];
1175                 foreach my $line (reverse @history) {
1176                         my @commit_info_split = split(/ |\n/, $line);
1177                         push(@commit_pairs, \@commit_info_split);
1178                 }
1179         }
1180
1181         foreach my $commit_info_split (@commit_pairs) {
1182                 my $sha1_child = @{$commit_info_split}[0];
1183                 my $sha1_commit = @{$commit_info_split}[1];
1184                 my $diff_infos = run_git("diff-tree -r --raw -z $sha1_child $sha1_commit");
1185                 # TODO: we could detect rename, and encode them with a #redirect on the wiki.
1186                 # TODO: for now, it's just a delete+add
1187                 my @diff_info_list = split(/\0/, $diff_infos);
1188                 # Keep the subject line of the commit message as mediawiki comment for the revision
1189                 my $commit_msg = run_git("log --no-walk --format=\"%s\" $sha1_commit");
1190                 chomp($commit_msg);
1191                 # Push every blob
1192                 while (@diff_info_list) {
1193                         my $status;
1194                         # git diff-tree -z gives an output like
1195                         # <metadata>\0<filename1>\0
1196                         # <metadata>\0<filename2>\0
1197                         # and we've split on \0.
1198                         my $info = shift(@diff_info_list);
1199                         my $file = shift(@diff_info_list);
1200                         ($mw_revision, $status) = mw_push_file($info, $file, $commit_msg, $mw_revision);
1201                         if ($status eq "non-fast-forward") {
1202                                 # we may already have sent part of the
1203                                 # commit to MediaWiki, but it's too
1204                                 # late to cancel it. Stop the push in
1205                                 # the middle, but still give an
1206                                 # accurate error message.
1207                                 return error_non_fast_forward($remote);
1208                         }
1209                         if ($status ne "ok") {
1210                                 die("Unknown error from mw_push_file()");
1211                         }
1212                 }
1213                 unless ($dumb_push) {
1214                         run_git("notes --ref=$remotename/mediawiki add -f -m \"mediawiki_revision: $mw_revision\" $sha1_commit");
1215                         run_git("update-ref -m \"Git-MediaWiki push\" refs/mediawiki/$remotename/master $sha1_commit $sha1_child");
1216                 }
1217         }
1218
1219         print STDOUT "ok $remote\n";
1220         return 1;
1221 }
1222
1223 sub get_allowed_file_extensions {
1224         mw_connect_maybe();
1225
1226         my $query = {
1227                 action => 'query',
1228                 meta => 'siteinfo',
1229                 siprop => 'fileextensions'
1230                 };
1231         my $result = $mediawiki->api($query);
1232         my @file_extensions= map $_->{ext},@{$result->{query}->{fileextensions}};
1233         my %hashFile = map {$_ => 1}@file_extensions;
1234
1235         return %hashFile;
1236 }
1237
1238 # In memory cache for MediaWiki namespace ids.
1239 my %namespace_id;
1240
1241 # Namespaces whose id is cached in the configuration file
1242 # (to avoid duplicates)
1243 my %cached_mw_namespace_id;
1244
1245 # Return MediaWiki id for a canonical namespace name.
1246 # Ex.: "File", "Project".
1247 sub get_mw_namespace_id {
1248         mw_connect_maybe();
1249         my $name = shift;
1250
1251         if (!exists $namespace_id{$name}) {
1252                 # Look at configuration file, if the record for that namespace is
1253                 # already cached. Namespaces are stored in form:
1254                 # "Name_of_namespace:Id_namespace", ex.: "File:6".
1255                 my @temp = split(/[\n]/, run_git("config --get-all remote."
1256                                                 . $remotename .".namespaceCache"));
1257                 chomp(@temp);
1258                 foreach my $ns (@temp) {
1259                         my ($n, $id) = split(/:/, $ns);
1260                         if ($id eq 'notANameSpace') {
1261                                 $namespace_id{$n} = {is_namespace => 0};
1262                         } else {
1263                                 $namespace_id{$n} = {is_namespace => 1, id => $id};
1264                         }
1265                         $cached_mw_namespace_id{$n} = 1;
1266                 }
1267         }
1268
1269         if (!exists $namespace_id{$name}) {
1270                 print STDERR "Namespace $name not found in cache, querying the wiki ...\n";
1271                 # NS not found => get namespace id from MW and store it in
1272                 # configuration file.
1273                 my $query = {
1274                         action => 'query',
1275                         meta => 'siteinfo',
1276                         siprop => 'namespaces'
1277                 };
1278                 my $result = $mediawiki->api($query);
1279
1280                 while (my ($id, $ns) = each(%{$result->{query}->{namespaces}})) {
1281                         if (defined($ns->{id}) && defined($ns->{canonical})) {
1282                                 $namespace_id{$ns->{canonical}} = {is_namespace => 1, id => $ns->{id}};
1283                                 if ($ns->{'*'}) {
1284                                         # alias (e.g. french Fichier: as alias for canonical File:)
1285                                         $namespace_id{$ns->{'*'}} = {is_namespace => 1, id => $ns->{id}};
1286                                 }
1287                         }
1288                 }
1289         }
1290
1291         my $ns = $namespace_id{$name};
1292         my $id;
1293
1294         unless (defined $ns) {
1295                 print STDERR "No such namespace $name on MediaWiki.\n";
1296                 $ns = {is_namespace => 0};
1297                 $namespace_id{$name} = $ns;
1298         }
1299
1300         if ($ns->{is_namespace}) {
1301                 $id = $ns->{id};
1302         }
1303
1304         # Store "notANameSpace" as special value for inexisting namespaces
1305         my $store_id = ($id || 'notANameSpace');
1306
1307         # Store explicitely requested namespaces on disk
1308         if (!exists $cached_mw_namespace_id{$name}) {
1309                 run_git("config --add remote.". $remotename
1310                         .".namespaceCache \"". $name .":". $store_id ."\"");
1311                 $cached_mw_namespace_id{$name} = 1;
1312         }
1313         return $id;
1314 }
1315
1316 sub get_mw_namespace_id_for_page {
1317         if (my ($namespace) = $_[0] =~ /^([^:]*):/) {
1318                 return get_mw_namespace_id($namespace);
1319         } else {
1320                 return;
1321         }
1322 }