Roll ICU to icu52

chrome/renderer/spellchecker/spellcheck_worditerator.cc

 // Copyright (c) 2012 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 // Implements a custom word iterator used for our spellchecker.
 
 #include "chrome/renderer/spellchecker/spellcheck_worditerator.h"
 
 #include <map>
 #include <string>
@@ skipping 48 matching lines @@
       "$LF           = [\\p{Word_Break = LF}];"
       "$Newline      = [\\p{Word_Break = Newline}];"
       "$Extend       = [\\p{Word_Break = Extend}];"
       "$Format       = [\\p{Word_Break = Format}];"
       "$Katakana     = [\\p{Word_Break = Katakana}];"
       // Not all the characters in a given script are ALetter.
       // For instance, U+05F4 is MidLetter. So, this may be
       // better, but it leads to an empty set error in Thai.
       // "$ALetter   = [[\\p{script=%s}] & [\\p{Word_Break = ALetter}]];"
       "$ALetter      = [\\p{script=%s}%s];"
-      "$MidNumLet    = [\\p{Word_Break = MidNumLet}];"
+      // U+0027 (single quote/apostrophe) is not in MidNumLet any more
+      // in UAX 29 rev 21 or later. For our purpose, U+0027
+      // has to be treated as MidNumLet.

[groby-ooo-7-16, 2014/06/10 20:18:11]

nit: Pointer to bug would be appreciated, so future generations (especially of
me ;) will know why.

[jungshik at Google, 2014/06/10 20:54:21]

Done.

+      "$MidNumLet    = [\\p{Word_Break = MidNumLet} \\u0027];"
       "$MidLetter    = [\\p{Word_Break = MidLetter}%s];"
       "$MidNum       = [\\p{Word_Break = MidNum}];"
       "$Numeric      = [\\p{Word_Break = Numeric}];"
       "$ExtendNumLet = [\\p{Word_Break = ExtendNumLet}];"
 
       "$Control        = [\\p{Grapheme_Cluster_Break = Control}]; "
       "%s"  // ALetterPlus
 
       "$KatakanaEx     = $Katakana     ($Extend |  $Format)*;"
       "$ALetterEx      = $ALetterPlus  ($Extend |  $Format)*;"
@@ skipping 65 matching lines @@
   // which don't need them.
   const char* aletter = uscript_getName(script_code_);
   if (!aletter)
     aletter = "Latin";
 
   const char kWithDictionary[] =
       "$dictionary   = [:LineBreak = Complex_Context:];"
       "$ALetterPlus  = [$ALetter [$dictionary-$Extend-$Control]];";
   const char kWithoutDictionary[] = "$ALetterPlus  = $ALetter;";
   const char* aletter_plus = kWithoutDictionary;
-  if (script_code_ == USCRIPT_HANGUL || script_code_ == USCRIPT_THAI)
+  if (script_code_ == USCRIPT_HANGUL || script_code_ == USCRIPT_THAI ||
+      script_code_ == USCRIPT_LAO || script_code_ == USCRIPT_KHMER)
     aletter_plus = kWithDictionary;
 
   // Treat numbers as word characters except for Arabic and Hebrew.
   const char* aletter_extra = " [0123456789]";
   if (script_code_ == USCRIPT_HEBREW || script_code_ == USCRIPT_ARABIC)
     aletter_extra = "";
 
   const char kMidLetterExtra[] = "";
   // For Hebrew, treat single/double quoation marks as MidLetter.
   const char kMidLetterExtraHebrew[] = "\"'";
@@ skipping 256 matching lines @@
   if (status != U_ZERO_ERROR && status != U_STRING_NOT_TERMINATED_WARNING)
     return false;
 
   // Copy the normalized text to the output.
   icu::StringCharacterIterator it(output);
   for (UChar c = it.first(); c != icu::CharacterIterator::DONE; c = it.next())
     attribute_->OutputChar(c, output_string);
 
   return !output_string->empty();
 }