[Interpreter] Adds support for DeleteLookupSlot to Interpreter.

[Interpreter] Adds support for DeleteLookupSlot to Interpreter.

Adds support for deleting a variable in a lookup slot. Adds a new bytecode,
its implementation and tests. Also adds support for this bytecode to the
bytecode graph builder.

BUG=v8:4280
LOG=N

Committed: https://crrev.com/6eb00e4ad31030e9eda7fa06b315e8fecd8053e0
Cr-Commit-Position: refs/heads/master@{#33019}

[mythria, 2015-12-22 11:22:54]

mythria@chromium.org changed reviewers:
+ bmeurer@chromium.org, oth@chromium.org

[mythria, 2015-12-22 11:22:55]

Orion, could you please review my changes. 
Benedikt, could you please take a look at compiler changes.

Thanks,
Mythri

https://codereview.chromium.org/1542083002/diff/1/src/interpreter/bytecode-ge...
File src/interpreter/bytecode-generator.cc (right):

https://codereview.chromium.org/1542083002/diff/1/src/interpreter/bytecode-ge...
src/interpreter/bytecode-generator.cc:1784:
builder()->LoadLiteral(variable->name()).DeleteLookupSlot();
I load the name to the accumulator instead of passing it to the bytecode as an
operand to be inline with other variants of Delete operation. We could pass it
as an explicit operand to reduce the size. I think it is not a frequent case, so
may not save much. We would also need an additional bytecode for the wide
variant if we pass name as an operand.

[Benedikt Meurer, 2015-12-22 11:23:33]

LGTM on compiler.

[oth, 2015-12-22 16:00:56]

lgtm, thanks.

https://codereview.chromium.org/1542083002/diff/1/src/interpreter/bytecode-ge...
File src/interpreter/bytecode-generator.cc (right):

https://codereview.chromium.org/1542083002/diff/1/src/interpreter/bytecode-ge...
src/interpreter/bytecode-generator.cc:1784:
builder()->LoadLiteral(variable->name()).DeleteLookupSlot();
On 2015/12/22 11:22:55, mythria wrote:
> I load the name to the accumulator instead of passing it to the bytecode as an
> operand to be inline with other variants of Delete operation. We could pass it
> as an explicit operand to reduce the size. I think it is not a frequent case,
so
> may not save much. We would also need an additional bytecode for the wide
> variant if we pass name as an operand.

Yes, this seems reasonable. We should have more data on bytecode frequencies
next quarter to guide these issues.

[mythria, 2015-12-23 09:10:01]

The CQ bit was checked by mythria@chromium.org

[commit-bot: I haz the power, 2015-12-23 09:10:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1542083002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1542083002/1

[commit-bot: I haz the power, 2015-12-23 09:11:20]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-12-23 09:12:01]

Description was changed from

==========
[Interpreter] Adds support for DeleteLookupSlot to Interpreter.

Adds support for deleting a variable in a lookup slot. Adds a new bytecode,
its implementation and tests. Also adds support for this bytecode to the
bytecode graph builder.

BUG=v8:4280
LOG=N
==========

to

==========
[Interpreter] Adds support for DeleteLookupSlot to Interpreter.

Adds support for deleting a variable in a lookup slot. Adds a new bytecode,
its implementation and tests. Also adds support for this bytecode to the
bytecode graph builder.

BUG=v8:4280
LOG=N

Committed: https://crrev.com/6eb00e4ad31030e9eda7fa06b315e8fecd8053e0
Cr-Commit-Position: refs/heads/master@{#33019}
==========

[commit-bot: I haz the power, 2015-12-23 09:12:02]

Patchset 1 (id:??) landed as
https://crrev.com/6eb00e4ad31030e9eda7fa06b315e8fecd8053e0
Cr-Commit-Position: refs/heads/master@{#33019}