Roll ICU to icu52

chrome/renderer/spellchecker/spellcheck_worditerator.cc

Index: chrome/renderer/spellchecker/spellcheck_worditerator.cc
===================================================================
--- chrome/renderer/spellchecker/spellcheck_worditerator.cc	(revision 271787)
+++ chrome/renderer/spellchecker/spellcheck_worditerator.cc	(working copy)
@@ -66,7 +66,10 @@
       // 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}];"
@@ -152,7 +155,8 @@
       "$ALetterPlus  = [$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.