t/inc/Sanitizer.t

   1 #!/usr/bin/env php
   2 <?php
   3
   4 require 't/Test.php';
   5
   6 plan( 13 );
   7
   8 define( 'MEDIAWIKI', 1 );
   9 require_ok( 'includes/Defines.php' );
  10 require_ok( 'includes/GlobalFunctions.php' );
  11 require_ok( 'includes/Sanitizer.php' );
  12 require_ok( 'includes/normal/UtfNormal.php' );
  13 require_ok( 'includes/ProfilerStub.php' ); # For removeHTMLtags
  14
  15
  16 #
  17 # decodeCharReferences
  18 #
  19
  20 cmp_ok(
  21         Sanitizer::decodeCharReferences( '&eacute;cole' ),
  22         '==',
  23         "\xc3\xa9cole",
  24         'decode named entities'
  25 );
  26
  27 cmp_ok(
  28         Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&#233;cole!" ),
  29         '==',
  30         "\xc4\x88io bonas dans l'\xc3\xa9cole!",
  31         'decode numeric entities'
  32 );
  33
  34 cmp_ok(
  35         Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&eacute;cole!" ),
  36         '==',
  37         "\xc4\x88io bonas dans l'\xc3\xa9cole!",
  38         'decode mixed numeric/named entities'
  39 );
  40
  41 cmp_ok(
  42         Sanitizer::decodeCharReferences(
  43                 "&#x108;io bonas dans l'&eacute;cole! (mais pas &amp;#x108;io dans l'&#38;eacute;cole)"
  44         ),
  45         '==',
  46         "\xc4\x88io bonas dans l'\xc3\xa9cole! (mais pas &#x108;io dans l'&eacute;cole)",
  47         'decode mixed complex entities'
  48 );
  49
  50 cmp_ok( Sanitizer::decodeCharReferences( 'a & b' ), '==', 'a & b', 'Invalid ampersand' );
  51
  52 cmp_ok( Sanitizer::decodeCharReferences( '&foo;' ), '==', '&foo;', 'Invalid named entity' );
  53
  54 cmp_ok( Sanitizer::decodeCharReferences( "&#88888888888888;" ), '==', UTF8_REPLACEMENT, 'Invalid numbered entity' );
  55
  56 $wgUseTidy = false;
  57 cmp_ok(
  58         Sanitizer::removeHTMLtags( '<div>Hello world</div />' ),
  59         '==',
  60         '<div>Hello world</div>',
  61         'Self-closing closing div'
  62 );
  63
  64 /* vim: set filetype=php: */