Fix some bugs in accessing details of the lastest regexp

Fix some bugs in accessing details of the lastest regexp
match.  Sometimes were were not updating it when we should
and sometimes we were leaving the lastMatchInfoOverride in
place when we should be using the updated regular last match
info.  Small optimization for zero length match in
String.prototype.replace.
Committed: https://code.google.com/p/v8/source/detail?r=11422

[ulan, 2012-04-23 16:59:23]

LGTM with comments.

https://chromiumcodereview.appspot.com/10184004/diff/1/src/macros.py
File src/macros.py (right):

https://chromiumcodereview.appspot.com/10184004/diff/1/src/macros.py#newcode199
src/macros.py:199: macro LAST_SUBJECT(array) = ((array)[1]);
We can avoid duplication:
macro LAST_SUBJECT(array) = ((array)[LAST_SUBJECT_INDEX]);

https://chromiumcodereview.appspot.com/10184004/diff/1/src/runtime.cc
File src/runtime.cc (right):

https://chromiumcodereview.appspot.com/10184004/diff/1/src/runtime.cc#newcode...
src/runtime.cc:2885: static void SetLastMatchInfoNoCaptures(Handle<String>
subject,
Why not just move the implementation here, instead of forward declaration?

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js
File src/string.js (right):

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode246
src/string.js:246: // We use this hack to detect whether
StringReplaceRegExpWithString found
Long line

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode249
src/string.js:249: lastMatchInfo[LAST_SUBJECT_INDEX] = 0;
This is sad, but I don't see a better way.

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode461
src/string.js:461: for ( ; i < len; i++) {
Can we make i local in the loop?

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode487
src/string.js:487: for ( ; i < len; i++) {
Can we make i local in the loop?

[Erik Corry, 2012-04-23 18:54:02]

https://chromiumcodereview.appspot.com/10184004/diff/1/src/macros.py
File src/macros.py (right):

https://chromiumcodereview.appspot.com/10184004/diff/1/src/macros.py#newcode199
src/macros.py:199: macro LAST_SUBJECT(array) = ((array)[1]);
On 2012/04/23 16:59:23, ulan wrote:
> We can avoid duplication:
> macro LAST_SUBJECT(array) = ((array)[LAST_SUBJECT_INDEX]);

Done.

https://chromiumcodereview.appspot.com/10184004/diff/1/src/runtime.cc
File src/runtime.cc (right):

https://chromiumcodereview.appspot.com/10184004/diff/1/src/runtime.cc#newcode...
src/runtime.cc:2885: static void SetLastMatchInfoNoCaptures(Handle<String>
subject,
On 2012/04/23 16:59:23, ulan wrote:
> Why not just move the implementation here, instead of forward declaration?

Because then it would be hard to review.

Done.

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js
File src/string.js (right):

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode246
src/string.js:246: // We use this hack to detect whether
StringReplaceRegExpWithString found
On 2012/04/23 16:59:23, ulan wrote:
> Long line

Done.

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode249
src/string.js:249: lastMatchInfo[LAST_SUBJECT_INDEX] = 0;
On 2012/04/23 16:59:23, ulan wrote:
> This is sad, but I don't see a better way.

Me neither.

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode461
src/string.js:461: for ( ; i < len; i++) {
On 2012/04/23 16:59:23, ulan wrote:
> Can we make i local in the loop?

Well not really.  This is JS, after all.  But we can pretend.

https://chromiumcodereview.appspot.com/10184004/diff/1/src/string.js#newcode487
src/string.js:487: for ( ; i < len; i++) {
On 2012/04/23 16:59:23, ulan wrote:
> Can we make i local in the loop?

As above.

[Erik Corry, 2012-04-23 19:04:06]

https://chromiumcodereview.appspot.com/10184004/diff/1/src/macros.py
File src/macros.py (right):

https://chromiumcodereview.appspot.com/10184004/diff/1/src/macros.py#newcode199
src/macros.py:199: macro LAST_SUBJECT(array) = ((array)[1]);
On 2012/04/23 18:54:02, Erik Corry wrote:
> On 2012/04/23 16:59:23, ulan wrote:
> > We can avoid duplication:
> > macro LAST_SUBJECT(array) = ((array)[LAST_SUBJECT_INDEX]);
> 
> Done.

Spoke too soon.  Our macro system is not advanced enough for this.