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