Pre-create properties on JSRegExp objects

Pre-create properties on JSRegExp objects
Initialize properties in single runtime call.

[Lasse Reichstein, 2010-03-26 13:18:58]

Medium size review.

[Erik Corry, 2010-03-26 13:51:34]

LGTM.

We could do the same for the strange array-with-two-extra-properties that is
returned by RegExp.prototype.exec.

http://codereview.chromium.org/1350003/diff/1/2
File src/bootstrapper.cc (right):

http://codereview.chromium.org/1350003/diff/1/2#newcode765
src/bootstrapper.cc:765: 3);
You already made a constant for this '3'.

http://codereview.chromium.org/1350003/diff/1/3
File src/factory.h (right):

http://codereview.chromium.org/1350003/diff/1/3#newcode44
src/factory.h:44: 
Inadvertent edit?

http://codereview.chromium.org/1350003/diff/1/7
File src/runtime.cc (right):

http://codereview.chromium.org/1350003/diff/1/7#newcode1263
src/runtime.cc:1263: // Map has changed, so use generic, but slower, method.
This is a strange situation.  Do we have a test of this?  Does JSC set the same
attributes on the properties in this case.

[Lasse Reichstein, 2010-03-26 14:19:09]

http://codereview.chromium.org/1350003/diff/1/2
File src/bootstrapper.cc (right):

http://codereview.chromium.org/1350003/diff/1/2#newcode765
src/bootstrapper.cc:765: 3);
Actually, not. This "3" is the enumeration index, not the field index. They just
happen to be the same value.

However, I'm not sure whether I need an index at all, since the value is not
enumerable.
I'll turn it into a counter. Easier to read that way, and the compiler should
constant fold it anyway.

http://codereview.chromium.org/1350003/diff/1/3
File src/factory.h (right):

http://codereview.chromium.org/1350003/diff/1/3#newcode44
src/factory.h:44: 
Yes, removed.

http://codereview.chromium.org/1350003/diff/1/7
File src/runtime.cc (right):

http://codereview.chromium.org/1350003/diff/1/7#newcode1263
src/runtime.cc:1263: // Map has changed, so use generic, but slower, method.
I guess the only way this can happen is when using RegExp.prototype.compile on a
regexp that has been fiddeled with. 
I'll add a test of this case.