Merges our hunspell change to hunspell 1.2.10....

Merges our hunspell change to hunspell 1.2.10.
This change refactors our hunspell change and merges it to hunspell 1.2.10. To compare with our old change to hunspell 1.2.8, this change changes the following points:
1. Use the original load_config() to parse the affix lines so we can use all hunspell options.
2. Enclose all our changes with '#ifdef HUNSPELL_CHROME_CLIENT' and '#endif'.
3. Remove unnecessary '#undef fclose'.

BUG=40909
TEST=unit_tests.exe --gtest_filter=SpellCheckTest*

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=50438

[Evan Stade, 2010-05-28 18:51:53]

great! this LG to my untrained eyes, but maybe you should wait for Brett to take
a look

http://codereview.chromium.org/2239005/diff/6001/7010
File third_party/hunspell/src/hunspell/hunspell.cxx (right):

http://codereview.chromium.org/2239005/diff/6001/7010#newcode600
third_party/hunspell/src/hunspell/hunspell.cxx:600: // We need to check the word
length if its valid to make coverity (Event
nit: We need to check if the word length is valid ...

http://codereview.chromium.org/2239005/diff/6001/7014
File third_party/hunspell/src/hunspell/suggestmgr.cxx (right):

http://codereview.chromium.org/2239005/diff/6001/7014#newcode17
third_party/hunspell/src/hunspell/suggestmgr.cxx:17: // A simmple class which
creates temporary hentry objects which can be
nit: simple

nit: s/can be/are

http://codereview.chromium.org/2239005/diff/6001/7014#newcode26
third_party/hunspell/src/hunspell/suggestmgr.cxx:26: //     hentry* scoped_copy
= factory.CreateHashEntry(source);
should be CreateScopedHashEntry ?

http://codereview.chromium.org/2239005/diff/6001/7014#newcode38
third_party/hunspell/src/hunspell/suggestmgr.cxx:38: // We need to improve it?
can we track this with a bug?

http://codereview.chromium.org/2239005/diff/6001/7014#newcode63
third_party/hunspell/src/hunspell/suggestmgr.cxx:63: //  
hash_item.entry->word[1] == hash_item->word[0].
nit: hash_item.word[0] instead of hash_item->word[0]

however, I don't really understand why this is necessary. Maybe you could
explain the purpose of astr?

[brettw, 2010-05-28 21:17:48]

LGTM

[Hironori Bono, 2010-06-22 09:25:42]

Thank you for your comments and sorry for my super-slow update. (I have been
salvaging this change from my corrupted disk.) I have updated the comments and
landed the updated one as r50438.

Regards,

Hironori Bono

http://codereview.chromium.org/2239005/diff/6001/7014
File third_party/hunspell/src/hunspell/suggestmgr.cxx (right):

http://codereview.chromium.org/2239005/diff/6001/7014#newcode17
third_party/hunspell/src/hunspell/suggestmgr.cxx:17: // A simmple class which
creates temporary hentry objects which can be
On 2010/05/28 18:51:53, Evan Stade wrote:
> nit: simple
> 
> nit: s/can be/are

Done.

http://codereview.chromium.org/2239005/diff/6001/7014#newcode26
third_party/hunspell/src/hunspell/suggestmgr.cxx:26: //     hentry* scoped_copy
= factory.CreateHashEntry(source);
On 2010/05/28 18:51:53, Evan Stade wrote:
> should be CreateScopedHashEntry ?

Done.

http://codereview.chromium.org/2239005/diff/6001/7014#newcode38
third_party/hunspell/src/hunspell/suggestmgr.cxx:38: // We need to improve it?
On 2010/05/28 18:51:53, Evan Stade wrote:
> can we track this with a bug?

Sorry for my obsolete comment. While fixing Issue 36523, I noticed hunspell 1.2
yielded more suggestions than hunspell 1.1 and it caused this performance
problem. (I should have written this comment after investigating it.) I have
removed this obsolete comment.

http://codereview.chromium.org/2239005/diff/6001/7014#newcode63
third_party/hunspell/src/hunspell/suggestmgr.cxx:63: //  
hash_item.entry->word[1] == hash_item->word[0].
On 2010/05/28 18:51:53, Evan Stade wrote:
> nit: hash_item.word[0] instead of hash_item->word[0]
> 
> however, I don't really understand why this is necessary. Maybe you could
> explain the purpose of astr?

Done. Sorry for my lazy comment. I have re-written the comment for this struct
to include description for 'astr'.