[interpreter] Remove special "prototype" load in class literals.

[interpreter] Remove special "prototype" load in class literals.

This allows us to remove the somewhat awkward BuildLoadObjectField
from the BytecodeGraphBuilder and also allows us to simplify the
bytecode stream for class literals.

R=oth@chromium.org

Committed: https://crrev.com/5fdf5c1e4d99feb773fa101da5c85d337e406fce
Cr-Commit-Position: refs/heads/master@{#33820}

[Michael Starzinger, 2016-02-08 14:23:42]

Orion: PTAL.
Benedikt: FYI.

[oth, 2016-02-08 15:00:04]

Nice! lgtm, thanks.

[rmcilroy, 2016-02-08 15:05:21]

rmcilroy@chromium.org changed reviewers:
+ rmcilroy@chromium.org

[rmcilroy, 2016-02-08 15:05:21]

https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
File test/cctest/interpreter/test-bytecode-generator.cc (right):

https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
test/cctest/interpreter/test-bytecode-generator.cc:8395:
DISABLED_TEST(ClassDeclarations) {
Ahh this is still disabled? Orion were you going to look into re-enabling these
tests or should I do it?

https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
test/cctest/interpreter/test-bytecode-generator.cc:8427: B(LoadICSloppy),       
                                              //
Is this right? LoadIC needs operands to specify the name / etc - ahh, just
spotted the reason this didn't fail, see above.

[Michael Starzinger, 2016-02-08 15:09:30]

https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
File test/cctest/interpreter/test-bytecode-generator.cc (right):

https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
test/cctest/interpreter/test-bytecode-generator.cc:8427: B(LoadICSloppy),       
                                              //
On 2016/02/08 15:05:21, rmcilroy wrote:
> Is this right? LoadIC needs operands to specify the name / etc - ahh, just
> spotted the reason this didn't fail, see above.

Yeah, this is just me being lazy and hoping we get a "rebase bytecode generator
expectation" script before these tests are being enabled. If you want me to
update tests, then I am happy to do so, but it doesn't make much sense without
also enabling them. Let me know. :)

[rmcilroy, 2016-02-08 15:14:09]

On 2016/02/08 15:09:30, Michael Starzinger wrote:
>
https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
> File test/cctest/interpreter/test-bytecode-generator.cc (right):
> 
>
https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
> test/cctest/interpreter/test-bytecode-generator.cc:8427: B(LoadICSloppy),     
 
>                                               //
> On 2016/02/08 15:05:21, rmcilroy wrote:
> > Is this right? LoadIC needs operands to specify the name / etc - ahh, just
> > spotted the reason this didn't fail, see above.
> 
> Yeah, this is just me being lazy and hoping we get a "rebase bytecode
generator
> expectation" script before these tests are being enabled. If you want me to
> update tests, then I am happy to do so, but it doesn't make much sense without
> also enabling them. Let me know. :)

Feel free to just leave it as-is for now. We can run use this as a test-run for
Stephano's tool once part 1 of it lands :).

[Michael Starzinger, 2016-02-08 15:17:51]

The CQ bit was checked by mstarzinger@chromium.org

[commit-bot: I haz the power, 2016-02-08 15:18:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1678103002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1678103002/1

[commit-bot: I haz the power, 2016-02-08 15:20:05]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-02-08 15:20:43]

Description was changed from

==========
[interpreter] Remove special "prototype" load in class literals.

This allows us to remove the somewhat awkward BuildLoadObjectField
from the BytecodeGraphBuilder and also allows us to simplify the
bytecode stream for class literals.

R=oth@chromium.org
==========

to

==========
[interpreter] Remove special "prototype" load in class literals.

This allows us to remove the somewhat awkward BuildLoadObjectField
from the BytecodeGraphBuilder and also allows us to simplify the
bytecode stream for class literals.

R=oth@chromium.org

Committed: https://crrev.com/5fdf5c1e4d99feb773fa101da5c85d337e406fce
Cr-Commit-Position: refs/heads/master@{#33820}
==========

[commit-bot: I haz the power, 2016-02-08 15:20:44]

Patchset 1 (id:??) landed as
https://crrev.com/5fdf5c1e4d99feb773fa101da5c85d337e406fce
Cr-Commit-Position: refs/heads/master@{#33820}

[oth, 2016-02-09 09:21:58]

On 2016/02/08 15:05:21, rmcilroy wrote:
>
https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
> File test/cctest/interpreter/test-bytecode-generator.cc (right):
> 
>
https://codereview.chromium.org/1678103002/diff/1/test/cctest/interpreter/tes...
> test/cctest/interpreter/test-bytecode-generator.cc:8395:
> DISABLED_TEST(ClassDeclarations) {
> Ahh this is still disabled? Orion were you going to look into re-enabling
these
> tests or should I do it?

Yes, I will fix it today, but have been off on the odyssey of supporting super.

[rmcilroy, 2016-02-09 10:05:11]

> Yes, I will fix it today, but have been off on the odyssey of supporting
super.

Thanks! No rush. Good luck in your super odyssey... :)