[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor

[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor

BUG=chromium:613796
LOG=N

Committed: https://crrev.com/2ca36cc33c7f96f3509716505ca7f35739a2d509
Cr-Commit-Position: refs/heads/master@{#36425}

[danno, 2016-05-21 09:09:35]

danno@chromium.org changed reviewers:
+ bmeurer@chromium.org

[danno, 2016-05-21 09:09:35]

PTAL

[Benedikt Meurer, 2016-05-21 09:12:22]

LGTM with comment.

https://codereview.chromium.org/2002813002/diff/1/src/code-stub-assembler.cc
File src/code-stub-assembler.cc (right):

https://codereview.chromium.org/2002813002/diff/1/src/code-stub-assembler.cc#...
src/code-stub-assembler.cc:704: // Don't use doubles to store the hole double,
since manipulating the
Mhm, yeah, we really need Float64 and Float32 helper classes to pass around
float64/float32 values in C++ while preserving the bit pattern. Maybe Enrico can
fix that, it's been on my list for long time.

In either case, please add TODO here, that this code can be removed once we have
Float64Constant operators using the Float64 class for preserving the precise bit
pattern.

[danno, 2016-05-21 09:17:17]

Description was changed from

==========
[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor
==========

to

==========
[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor

BUG=chromium:613796
LOG=N
==========

[danno, 2016-05-21 09:21:32]

The CQ bit was checked by danno@chromium.org

[danno, 2016-05-21 09:21:33]

The patchset sent to the CQ was uploaded after l-g-t-m from bmeurer@chromium.org
Link to the patchset: https://codereview.chromium.org/2002813002/#ps20001
(title: "Review feedback")

[commit-bot: I haz the power, 2016-05-21 09:21:43]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2002813002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/2002813002/20001

[danno, 2016-05-21 09:22:16]

Thanks for the weekend review! Feedback addressed, landing.

https://codereview.chromium.org/2002813002/diff/1/src/code-stub-assembler.cc
File src/code-stub-assembler.cc (right):

https://codereview.chromium.org/2002813002/diff/1/src/code-stub-assembler.cc#...
src/code-stub-assembler.cc:704: // Don't use doubles to store the hole double,
since manipulating the
On 2016/05/21 09:12:22, Benedikt Meurer wrote:
> Mhm, yeah, we really need Float64 and Float32 helper classes to pass around
> float64/float32 values in C++ while preserving the bit pattern. Maybe Enrico
can
> fix that, it's been on my list for long time.
> 
> In either case, please add TODO here, that this code can be removed once we
have
> Float64Constant operators using the Float64 class for preserving the precise
bit
> pattern.

Done.

[commit-bot: I haz the power, 2016-05-21 09:52:22]

Description was changed from

==========
[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor

BUG=chromium:613796
LOG=N
==========

to

==========
[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor

BUG=chromium:613796
LOG=N
==========

[commit-bot: I haz the power, 2016-05-21 09:52:23]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-05-21 09:54:46]

Description was changed from

==========
[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor

BUG=chromium:613796
LOG=N
==========

to

==========
[stubs] Fix hole-related double bug in ArrayNoArgumentConstructor

BUG=chromium:613796
LOG=N

Committed: https://crrev.com/2ca36cc33c7f96f3509716505ca7f35739a2d509
Cr-Commit-Position: refs/heads/master@{#36425}
==========

[commit-bot: I haz the power, 2016-05-21 09:54:47]

Patchset 2 (id:??) landed as
https://crrev.com/2ca36cc33c7f96f3509716505ca7f35739a2d509
Cr-Commit-Position: refs/heads/master@{#36425}