Handle insertion order for simple constructors

Handle insertion order for simple constructors

When functions only have simple assignments of the form this.x = ... the object is created in generated code without actually calling the constructor. In this case the initial map for the function already contains the properties assigned in the constructor. The field descriptors in this initial map now has an enumeration index assigned to make property enumeration order the insertion order. The insertion order here is the order of the this.x assignments in the code.

BUG=http://crbug.com/3867
TEST=test/mjsunit/regress/regress-crbug-3867.js


Committed: http://code.google.com/p/v8/source/detail?r=3768

[Kasper Lund, 2010-02-02 13:14:52]

LGTM, but:

http://codereview.chromium.org/566016/diff/3001/4001
File test/mjsunit/regress/regress-crbug-3867.js (right):

http://codereview.chromium.org/566016/diff/3001/4001#newcode1
test/mjsunit/regress/regress-crbug-3867.js:1: // Copyright 2010 the V8 project
authors. All rights reserved.
Make sure this isn't executable.

http://codereview.chromium.org/566016/diff/3001/4001#newcode41
test/mjsunit/regress/regress-crbug-3867.js:41: assertArrayEquals(["s1", "s2",
"s3"], props({s1: 0, s2: 0, s3: 0}));
It would be great with a test where the property names aren't sorted to make
sure that it's really the insertion order that matters for named properties.

[Kasper Lund, 2010-02-02 13:23:47]

Additional comment:

http://codereview.chromium.org/566016/diff/3001/4001
File test/mjsunit/regress/regress-crbug-3867.js (right):

http://codereview.chromium.org/566016/diff/3001/4001#newcode40
test/mjsunit/regress/regress-crbug-3867.js:40: assertArrayEquals(["s1", "s2",
"s3"], props(new A()));
I would be nice to also test that adding properties to an object created through
'new A()' keeps the right insertion order of the properties (basically test that
you've set the NextEnumerationIndex correctly).

[Søren Thygesen Gjesse, 2010-02-02 13:33:00]

http://codereview.chromium.org/566016/diff/3001/4001
File test/mjsunit/regress/regress-crbug-3867.js (right):

http://codereview.chromium.org/566016/diff/3001/4001#newcode1
test/mjsunit/regress/regress-crbug-3867.js:1: // Copyright 2010 the V8 project
authors. All rights reserved.
On 2010/02/02 13:14:52, Kasper Lund wrote:
> Make sure this isn't executable.

Done.

http://codereview.chromium.org/566016/diff/3001/4001#newcode40
test/mjsunit/regress/regress-crbug-3867.js:40: assertArrayEquals(["s1", "s2",
"s3"], props(new A()));
On 2010/02/02 13:23:47, Kasper Lund wrote:
> I would be nice to also test that adding properties to an object created
through
> 'new A()' keeps the right insertion order of the properties (basically test
that
> you've set the NextEnumerationIndex correctly).

Done.

http://codereview.chromium.org/566016/diff/3001/4001#newcode41
test/mjsunit/regress/regress-crbug-3867.js:41: assertArrayEquals(["s1", "s2",
"s3"], props({s1: 0, s2: 0, s3: 0}));
On 2010/02/02 13:14:52, Kasper Lund wrote:
> It would be great with a test where the property names aren't sorted to make
> sure that it's really the insertion order that matters for named properties.

Done.

[Kasper Lund, 2010-02-02 13:34:19]

Thank you for updating the test. Looks nice.