Current custom call generators cannot cope with the case when receiver is not a JSArray.

Current custom call generators cannot cope with the case when receiver is not a JSArray.

Add a support for bailout from custom call generators (just return undefined).

BUG=684
Committed: http://code.google.com/p/v8/source/detail?r=4503

[antonm, 2010-04-23 16:54:27]

Lasse,

may you have a look?

[Lasse Reichstein, 2010-04-26 09:52:31]

LGTM

http://codereview.chromium.org/1699005/diff/1/2
File src/ia32/stub-cache-ia32.cc (right):

http://codereview.chromium.org/1699005/diff/1/2#newcode1492
src/ia32/stub-cache-ia32.cc:1492: return result;
Add an assert before the return to ensure that the result is a Code*.

http://codereview.chromium.org/1699005/diff/1/3
File test/mjsunit/array-pop.js (right):

http://codereview.chromium.org/1699005/diff/1/3#newcode63
test/mjsunit/array-pop.js:63: // Play with prototype chains.
Change description to say that it's using a non-JSArray as receiver for the pop,
and that it's a test for custom generators.

http://codereview.chromium.org/1699005/diff/1/4
File test/mjsunit/array-push.js (right):

http://codereview.chromium.org/1699005/diff/1/4#newcode107
test/mjsunit/array-push.js:107: // Play with prototype chains.
Same comment as for pop.

[antonm, 2010-04-26 14:42:00]

http://codereview.chromium.org/1699005/diff/1/2
File src/ia32/stub-cache-ia32.cc (right):

http://codereview.chromium.org/1699005/diff/1/2#newcode1492
src/ia32/stub-cache-ia32.cc:1492: return result;
On 2010/04/26 09:52:31, Lasse Reichstein wrote:
> Add an assert before the return to ensure that the result is a Code*.

Could it be a failure?

http://codereview.chromium.org/1699005/diff/1/3
File test/mjsunit/array-pop.js (right):

http://codereview.chromium.org/1699005/diff/1/3#newcode63
test/mjsunit/array-pop.js:63: // Play with prototype chains.
On 2010/04/26 09:52:31, Lasse Reichstein wrote:
> Change description to say that it's using a non-JSArray as receiver for the
pop,
> and that it's a test for custom generators.
> 

Done.

http://codereview.chromium.org/1699005/diff/1/4
File test/mjsunit/array-push.js (right):

http://codereview.chromium.org/1699005/diff/1/4#newcode107
test/mjsunit/array-push.js:107: // Play with prototype chains.
On 2010/04/26 09:52:31, Lasse Reichstein wrote:
> Same comment as for pop.
> 

Done.

[antonm, 2010-04-26 15:03:38]

Lasse,

as I came to late into the office and won't see till Wednesday, I am going to
submit this CL right now w/o assert, but would add one later when we clear it. 

On 2010/04/26 14:42:00, antonm wrote:
> http://codereview.chromium.org/1699005/diff/1/2
> File src/ia32/stub-cache-ia32.cc (right):
> 
> http://codereview.chromium.org/1699005/diff/1/2#newcode1492
> src/ia32/stub-cache-ia32.cc:1492: return result;
> On 2010/04/26 09:52:31, Lasse Reichstein wrote:
> > Add an assert before the return to ensure that the result is a Code*.
> 
> Could it be a failure?
> 
> http://codereview.chromium.org/1699005/diff/1/3
> File test/mjsunit/array-pop.js (right):
> 
> http://codereview.chromium.org/1699005/diff/1/3#newcode63
> test/mjsunit/array-pop.js:63: // Play with prototype chains.
> On 2010/04/26 09:52:31, Lasse Reichstein wrote:
> > Change description to say that it's using a non-JSArray as receiver for the
> pop,
> > and that it's a test for custom generators.
> > 
> 
> Done.
> 
> http://codereview.chromium.org/1699005/diff/1/4
> File test/mjsunit/array-push.js (right):
> 
> http://codereview.chromium.org/1699005/diff/1/4#newcode107
> test/mjsunit/array-push.js:107: // Play with prototype chains.
> On 2010/04/26 09:52:31, Lasse Reichstein wrote:
> > Same comment as for pop.
> > 
> 
> Done.