Eliminate the code that handles fallback to JSCRE. The only way to get...

Eliminate the code that handles fallback to JSCRE.  The only way to get
JSCRE now is to use the --noirregexp flag.  Also add code to check that
we react sensibly to some very large regexps.

Committed: http://code.google.com/p/v8/source/detail?r=1166

[Lasse Reichstein, 2009-01-26 13:48:52]

LGTM

http://codereview.chromium.org/18587/diff/1/6
File src/jsregexp.cc (right):

http://codereview.chromium.org/18587/diff/1/6#newcode1279
Line 1279: while (!work_list.is_empty()) {
You could bail out of this loop if the regexp is too big (i.e., work list is
mepty or reg_exp_too_big_ is true).

http://codereview.chromium.org/18587/diff/1/6#newcode1483
Line 1483: return;
The return is superfluous.

http://codereview.chromium.org/18587/diff/1/2
File src/regexp-macro-assembler.h (right):

http://codereview.chromium.org/18587/diff/1/2#newcode41
Line 41: // The implmentation must be able to handle at least:
Typo "implementation".

I agree that any regexp failing these offsets is a pathological case, and they
are acceptable to the ia32-version (2^16 registers should fit on the system
stack).

[Erik Corry, 2009-01-26 19:59:20]

http://codereview.chromium.org/18587/diff/1/6
File src/jsregexp.cc (right):

http://codereview.chromium.org/18587/diff/1/6#newcode1279
Line 1279: while (!work_list.is_empty()) {
On 2009/01/26 13:48:53, Lasse Reichstein wrote:
> You could bail out of this loop if the regexp is too big (i.e., work list is
> mepty or reg_exp_too_big_ is true).

Unfortunately I can't because there are labels that are linked but won't get
bound in this case.  There's an assert in debug mode that will trigger in this
case.

http://codereview.chromium.org/18587/diff/1/6#newcode1483
Line 1483: return;
On 2009/01/26 13:48:53, Lasse Reichstein wrote:
> The return is superfluous.

I don't think so!

[Lasse Reichstein, 2009-01-27 06:54:28]

http://codereview.chromium.org/18587/diff/1/6
File src/jsregexp.cc (right):

http://codereview.chromium.org/18587/diff/1/6#newcode1483
Line 1483: return;
On 2009/01/26 19:59:21, Erik Corry wrote:
> On 2009/01/26 13:48:53, Lasse Reichstein wrote: > The return is superfluous. I
don't think so!

Argh, easily fooled by missing indentation. Ignore me.