ILIAS  release_5-1 Revision 5.0.0-5477-g43f3e3fab5f
UtfNormalTest.php
Go to the documentation of this file.
1 <?php
2 # Copyright (C) 2004 Brion Vibber <brion@pobox.com>
3 # http://www.mediawiki.org/
4 #
5 # This program is free software; you can redistribute it and/or modify
6 # it under the terms of the GNU General Public License as published by
7 # the Free Software Foundation; either version 2 of the License, or
8 # (at your option) any later version.
9 #
10 # This program is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 # GNU General Public License for more details.
14 #
15 # You should have received a copy of the GNU General Public License along
16 # with this program; if not, write to the Free Software Foundation, Inc.,
17 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 # http://www.gnu.org/copyleft/gpl.html
19 
27 $verbose = true;
28 #define( 'PRETTY_UTF8', true );
29 
30 if( defined( 'PRETTY_UTF8' ) ) {
31  function pretty( $string ) {
32  return preg_replace_callback(
33  '/([\x00-\xff])/',
34  function($hit) {
35  return sprintf("%02X", ord($hit[1]));
36  },
37  $string
38  );
39  }
40 } else {
44  function pretty( $string ) {
45  return trim( preg_replace_callback(
46  '/(.)/us',
47  function($hit) {
48  return sprintf("%04X ", utf8ToCodepoint($hit[1]));
49  },
50  $string
51  ));
52  }
53 }
54 
55 if( isset( $_SERVER['argv'] ) && in_array( '--icu', $_SERVER['argv'] ) ) {
56  dl( 'php_utfnormal.so' );
57 }
58 
59 require_once 'UtfNormalUtil.php';
60 require_once 'UtfNormal.php';
61 
62 if( php_sapi_name() != 'cli' ) {
63  die( "Run me from the command line please.\n" );
64 }
65 
66 $in = fopen("NormalizationTest.txt", "rt");
67 if( !$in ) {
68  print "Couldn't open NormalizationTest.txt -- can't run tests.\n";
69  print "If necessary, manually download this file. It can be obtained at\n";
70  print "http://www.unicode.org/Public/UNIDATA/NormalizationTest.txt";
71  exit(-1);
72 }
73 
75 
76 $total = 0;
79 $ok = true;
80 $testedChars = array();
81 while( false !== ( $line = fgets( $in ) ) ) {
82  list( $data, $comment ) = explode( '#', $line );
83  if( $data === '' ) continue;
84  $matches = array();
85  if( preg_match( '/@Part([\d])/', $data, $matches ) ) {
86  if( $matches[1] > 0 ) {
88  }
89  print "Part {$matches[1]}: $comment";
90  continue;
91  }
92 
93  $columns = array_map( "hexSequenceToUtf8", explode( ";", $data ) );
94  array_unshift( $columns, '' );
95 
96  $testedChars[$columns[1]] = true;
97  $total++;
98  if( testNormals( $normalizer, $columns, $comment ) ) {
99  $success++;
100  } else {
101  $failure++;
102  # print "FAILED: $comment";
103  }
104  if( $total % 100 == 0 ) print "$total ";
105 }
106 fclose( $in );
107 
109 
110 $in = fopen("UnicodeData.txt", "rt" );
111 if( !$in ) {
112  print "Can't open UnicodeData.txt for reading.\n";
113  print "If necessary, fetch this file from the internet:\n";
114  print "http://www.unicode.org/Public/UNIDATA/UnicodeData.txt\n";
115  exit(-1);
116 }
117 print "Now testing invariants...\n";
118 while( false !== ($line = fgets( $in ) ) ) {
119  $cols = explode( ';', $line );
120  $char = codepointToUtf8( hexdec( $cols[0] ) );
121  $desc = $cols[0] . ": " . $cols[1];
122  if( $char < "\x20" || $char >= UTF8_SURROGATE_FIRST && $char <= UTF8_SURROGATE_LAST ) {
123  # Can't check NULL with the ICU plugin, as null bytes fail in C land.
124  # Skip other control characters, as we strip them for XML safety.
125  # Surrogates are illegal on their own or in UTF-8, ignore.
126  continue;
127  }
128  if( empty( $testedChars[$char] ) ) {
129  $total++;
130  if( testInvariant( $normalizer, $char, $desc ) ) {
131  $success++;
132  } else {
133  $failure++;
134  }
135  if( $total % 100 == 0 ) print "$total ";
136  }
137 }
138 fclose( $in );
139 
141 
142 if( $ok ) {
143  print "TEST SUCCEEDED!\n";
144  exit(0);
145 } else {
146  print "TEST FAILED!\n";
147  exit(-1);
148 }
149 
150 ## ------
151 
153  $percSucc = intval( $success * 100 / $total );
154  $percFail = intval( $failure * 100 / $total );
155  print "\n";
156  print "$success tests successful ($percSucc%)\n";
157  print "$failure tests failed ($percFail%)\n\n";
158  $ok = ($success > 0 && $failure == 0);
159  $total = 0;
160  $success = 0;
161  $failure = 0;
162  return $ok;
163 }
164 
165 function testNormals( &$u, $c, $comment, $reportFailure = false ) {
166  $result = testNFC( $u, $c, $comment, $reportFailure );
167  $result = testNFD( $u, $c, $comment, $reportFailure ) && $result;
168  $result = testNFKC( $u, $c, $comment, $reportFailure ) && $result;
169  $result = testNFKD( $u, $c, $comment, $reportFailure ) && $result;
170  $result = testCleanUp( $u, $c, $comment, $reportFailure ) && $result;
171 
172  global $verbose;
173  if( $verbose && !$result && !$reportFailure ) {
174  print $comment;
175  testNormals( $u, $c, $comment, true );
176  }
177  return $result;
178 }
179 
180 function verbosify( $a, $b, $col, $form, $verbose ) {
181  #$result = ($a === $b);
182  $result = (strcmp( $a, $b ) == 0);
183  if( $verbose ) {
184  $aa = pretty( $a );
185  $bb = pretty( $b );
186  $ok = $result ? "succeed" : " failed";
187  $eq = $result ? "==" : "!=";
188  print " $ok $form c$col '$aa' $eq '$bb'\n";
189  }
190  return $result;
191 }
192 
193 function testNFC( &$u, $c, $comment, $verbose ) {
194  $result = verbosify( $c[2], $u->toNFC( $c[1] ), 1, 'NFC', $verbose );
195  $result = verbosify( $c[2], $u->toNFC( $c[2] ), 2, 'NFC', $verbose ) && $result;
196  $result = verbosify( $c[2], $u->toNFC( $c[3] ), 3, 'NFC', $verbose ) && $result;
197  $result = verbosify( $c[4], $u->toNFC( $c[4] ), 4, 'NFC', $verbose ) && $result;
198  $result = verbosify( $c[4], $u->toNFC( $c[5] ), 5, 'NFC', $verbose ) && $result;
199  return $result;
200 }
201 
202 function testCleanUp( &$u, $c, $comment, $verbose ) {
203  $x = $c[1];
204  $result = verbosify( $c[2], $u->cleanUp( $x ), 1, 'cleanUp', $verbose );
205  $x = $c[2];
206  $result = verbosify( $c[2], $u->cleanUp( $x ), 2, 'cleanUp', $verbose ) && $result;
207  $x = $c[3];
208  $result = verbosify( $c[2], $u->cleanUp( $x ), 3, 'cleanUp', $verbose ) && $result;
209  $x = $c[4];
210  $result = verbosify( $c[4], $u->cleanUp( $x ), 4, 'cleanUp', $verbose ) && $result;
211  $x = $c[5];
212  $result = verbosify( $c[4], $u->cleanUp( $x ), 5, 'cleanUp', $verbose ) && $result;
213  return $result;
214 }
215 
216 function testNFD( &$u, $c, $comment, $verbose ) {
217  $result = verbosify( $c[3], $u->toNFD( $c[1] ), 1, 'NFD', $verbose );
218  $result = verbosify( $c[3], $u->toNFD( $c[2] ), 2, 'NFD', $verbose ) && $result;
219  $result = verbosify( $c[3], $u->toNFD( $c[3] ), 3, 'NFD', $verbose ) && $result;
220  $result = verbosify( $c[5], $u->toNFD( $c[4] ), 4, 'NFD', $verbose ) && $result;
221  $result = verbosify( $c[5], $u->toNFD( $c[5] ), 5, 'NFD', $verbose ) && $result;
222  return $result;
223 }
224 
225 function testNFKC( &$u, $c, $comment, $verbose ) {
226  $result = verbosify( $c[4], $u->toNFKC( $c[1] ), 1, 'NFKC', $verbose );
227  $result = verbosify( $c[4], $u->toNFKC( $c[2] ), 2, 'NFKC', $verbose ) && $result;
228  $result = verbosify( $c[4], $u->toNFKC( $c[3] ), 3, 'NFKC', $verbose ) && $result;
229  $result = verbosify( $c[4], $u->toNFKC( $c[4] ), 4, 'NFKC', $verbose ) && $result;
230  $result = verbosify( $c[4], $u->toNFKC( $c[5] ), 5, 'NFKC', $verbose ) && $result;
231  return $result;
232 }
233 
234 function testNFKD( &$u, $c, $comment, $verbose ) {
235  $result = verbosify( $c[5], $u->toNFKD( $c[1] ), 1, 'NFKD', $verbose );
236  $result = verbosify( $c[5], $u->toNFKD( $c[2] ), 2, 'NFKD', $verbose ) && $result;
237  $result = verbosify( $c[5], $u->toNFKD( $c[3] ), 3, 'NFKD', $verbose ) && $result;
238  $result = verbosify( $c[5], $u->toNFKD( $c[4] ), 4, 'NFKD', $verbose ) && $result;
239  $result = verbosify( $c[5], $u->toNFKD( $c[5] ), 5, 'NFKD', $verbose ) && $result;
240  return $result;
241 }
242 
243 function testInvariant( &$u, $char, $desc, $reportFailure = false ) {
244  $result = verbosify( $char, $u->toNFC( $char ), 1, 'NFC', $reportFailure );
245  $result = verbosify( $char, $u->toNFD( $char ), 1, 'NFD', $reportFailure ) && $result;
246  $result = verbosify( $char, $u->toNFKC( $char ), 1, 'NFKC', $reportFailure ) && $result;
247  $result = verbosify( $char, $u->toNFKD( $char ), 1, 'NFKD', $reportFailure ) && $result;
248  $result = verbosify( $char, $u->cleanUp( $char ), 1, 'cleanUp', $reportFailure ) && $result;
249  global $verbose;
250  if( $verbose && !$result && !$reportFailure ) {
251  print $desc;
252  testInvariant( $u, $char, $desc, true );
253  }
254  return $result;
255 }
256 
257 ?>
Unicode normalization routines for working with UTF-8 strings.
testNFD(&$u, $c, $comment, $verbose)
$failure
exit
Definition: login.php:54
if((!isset($_SERVER['DOCUMENT_ROOT'])) OR(empty($_SERVER['DOCUMENT_ROOT']))) $_SERVER['DOCUMENT_ROOT']
utf8ToCodepoint( $char)
Determine the Unicode codepoint of a single-character UTF-8 sequence.
$result
$success
$x
Definition: example_009.php:98
testInvariant(&$u, $char, $desc, $reportFailure=false)
testNormals(&$u, $c, $comment, $reportFailure=false)
$verbose
testNFKC(&$u, $c, $comment, $verbose)
const UTF8_SURROGATE_LAST
Definition: UtfNormal.php:81
testNFC(&$u, $c, $comment, $verbose)
codepointToUtf8( $codepoint)
Return UTF-8 sequence for a given Unicode code point.
$total
if(! $in) $normalizer
$data
const UTF8_SURROGATE_FIRST
Definition: UtfNormal.php:80
$testedChars
verbosify( $a, $b, $col, $form, $verbose)
reportResults(&$total, &$success, &$failure)
$comment
Definition: buildRTE.php:83
testNFKD(&$u, $c, $comment, $verbose)
if(php_sapi_name() !='cli') $in
if(! $in) $columns
Definition: Utf8Test.php:46
testCleanUp(&$u, $c, $comment, $verbose)