Revert I+TF for lexical variables

Revert I+TF for lexical variables

This reverts the following commits to fix a Node.js regression:
* 5529430dec0d8997319d46e02c473a7a4faf1933 "[compiler] Consistently use Ignition+TurboFan for lexical variables."
* 78691367163ad318a6b441f03de5e7f5bf400df7 "[compiler] Improve let+const decision in AstNumbering."


R=adamk@chromium.org, bmeurer@chromium.org, hablich@chromium.org
BUG=
NOTRY=true

Review-Url: https://codereview.chromium.org/2647523002
Cr-Commit-Position: refs/heads/master@{#42484}
Committed: https://chromium.googlesource.com/v8/v8/+/89b7a4d7d0a2539c8e3d2422f90c6f6fbdecd0a6

[Benedikt Meurer, 2017-01-18 20:29:00]

lgtm

[ofrobots, 2017-01-18 20:46:06]

The CQ bit was checked by ofrobots@google.com to run a CQ dry run

[commit-bot: I haz the power, 2017-01-18 20:46:10]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-01-18 21:12:05]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-01-18 21:12:06]

Dry run: Try jobs failed on following builders:
  v8_linux_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_rel_ng/builds/19452)
  v8_linux_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_rel_ng_triggered/b...)

[ofrobots, 2017-01-18 22:02:05]

On 2017/01/18 20:29:00, Benedikt Meurer wrote:
> lgtm

Hmm.. not quite sure why some of the test262 Intl tests are failing in the CQ on
some build bots. They don't fail locally.

Looking at the test 6.2.2_b.js for example, the copy of the test in the V8 tree
seems to be different from what the bots are actually running. For example, the
CQ complains:

=== test262/intl402/6.2.2_b ===
--- stdout ---
/b/swarm_slave/w/irbuRwjx/test/test262/data/harness/testIntl.js:30:
Test262Error: Invalid language tag de-tester-Tester was not rejected. (Testing
with Collator.)
            throw e;
            ^
Command: /b/swarm_slave/w/irbuRwjx/out/Release/d8 --test
--random-seed=-564235616 --nohard-abort --nodead-code-elimination
--nofold-constants /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/sta.js
/b/swarm_slave/w/irbuRwjx/test/test262/data/harness/assert.js
/b/swarm_slave/w/irbuRwjx/test/test262/harness-adapt.js
/b/swarm_slave/w/irbuRwjx/test/test262/data/harness/testIntl.js
/b/swarm_slave/w/irbuRwjx/test/test262/local-tests/test/intl402/6.2.2_b.js


Looking at the file 6.2.2_b.js, there is no listed tag named 'de-tester-Tester'.

Any ideas what might be going wrong?

[Dan Ehrenberg, 2017-01-18 23:22:13]

On 2017/01/18 22:02:05, ofrobots wrote:
> On 2017/01/18 20:29:00, Benedikt Meurer wrote:
> > lgtm
> 
> Hmm.. not quite sure why some of the test262 Intl tests are failing in the CQ
on
> some build bots. They don't fail locally.
> 
> Looking at the test 6.2.2_b.js for example, the copy of the test in the V8
tree
> seems to be different from what the bots are actually running. For example,
the
> CQ complains:
> 
> === test262/intl402/6.2.2_b ===
> --- stdout ---
> /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/testIntl.js:30:
> Test262Error: Invalid language tag de-tester-Tester was not rejected. (Testing
> with Collator.)
>             throw e;
>             ^
> Command: /b/swarm_slave/w/irbuRwjx/out/Release/d8 --test
> --random-seed=-564235616 --nohard-abort --nodead-code-elimination
> --nofold-constants /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/sta.js
> /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/assert.js
> /b/swarm_slave/w/irbuRwjx/test/test262/harness-adapt.js
> /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/testIntl.js
> /b/swarm_slave/w/irbuRwjx/test/test262/local-tests/test/intl402/6.2.2_b.js
> 
> 
> Looking at the file 6.2.2_b.js, there is no listed tag named
'de-tester-Tester'.
> 
> Any ideas what might be going wrong?

This is very weird. That test is out for review in
https://codereview.chromium.org/2639333003/ . Seems like there must be some sort
of infrastructure issue that I didn't handle in
https://docs.google.com/document/d/16bj7AIDgZLv4WOsUEzQ5NzcEN9_xo095e88Pz8FC5...
properly. Maybe the bots don't clear out newly added files in that directory
between runs. +machenbach do you have any thoughts?

[adamk, 2017-01-18 23:26:27]

On 2017/01/18 23:22:13, Dan Ehrenberg wrote:
> On 2017/01/18 22:02:05, ofrobots wrote:
> > On 2017/01/18 20:29:00, Benedikt Meurer wrote:
> > > lgtm
> > 
> > Hmm.. not quite sure why some of the test262 Intl tests are failing in the
CQ
> on
> > some build bots. They don't fail locally.
> > 
> > Looking at the test 6.2.2_b.js for example, the copy of the test in the V8
> tree
> > seems to be different from what the bots are actually running. For example,
> the
> > CQ complains:
> > 
> > === test262/intl402/6.2.2_b ===
> > --- stdout ---
> > /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/testIntl.js:30:
> > Test262Error: Invalid language tag de-tester-Tester was not rejected.
(Testing
> > with Collator.)
> >             throw e;
> >             ^
> > Command: /b/swarm_slave/w/irbuRwjx/out/Release/d8 --test
> > --random-seed=-564235616 --nohard-abort --nodead-code-elimination
> > --nofold-constants
/b/swarm_slave/w/irbuRwjx/test/test262/data/harness/sta.js
> > /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/assert.js
> > /b/swarm_slave/w/irbuRwjx/test/test262/harness-adapt.js
> > /b/swarm_slave/w/irbuRwjx/test/test262/data/harness/testIntl.js
> > /b/swarm_slave/w/irbuRwjx/test/test262/local-tests/test/intl402/6.2.2_b.js
> > 
> > 
> > Looking at the file 6.2.2_b.js, there is no listed tag named
> 'de-tester-Tester'.
> > 
> > Any ideas what might be going wrong?
> 
> This is very weird. That test is out for review in
> https://codereview.chromium.org/2639333003/ . Seems like there must be some
sort
> of infrastructure issue that I didn't handle in
>
https://docs.google.com/document/d/16bj7AIDgZLv4WOsUEzQ5NzcEN9_xo095e88Pz8FC5...
> properly. Maybe the bots don't clear out newly added files in that directory
> between runs. +machenbach do you have any thoughts?

FWIW this is clearly an unrelated failure and the rest of the bots are green, so
I think it'd be fine to land this particular patch manually (or to add
NOTRY=true to the CL description and re-tick the button).

[ofrobots, 2017-01-19 00:37:12]

Description was changed from

==========
Revert I+TF for lexical variables

This reverts the following commits to fix a Node.js regression:
* 5529430dec0d8997319d46e02c473a7a4faf1933 "[compiler] Consistently use
Ignition+TurboFan for lexical variables."
* 78691367163ad318a6b441f03de5e7f5bf400df7 "[compiler] Improve let+const
decision in AstNumbering."


R=adamk@chromium.org, bmeurer@chromium.org, hablich@chromium.org
BUG=
==========

to

==========
Revert I+TF for lexical variables

This reverts the following commits to fix a Node.js regression:
* 5529430dec0d8997319d46e02c473a7a4faf1933 "[compiler] Consistently use
Ignition+TurboFan for lexical variables."
* 78691367163ad318a6b441f03de5e7f5bf400df7 "[compiler] Improve let+const
decision in AstNumbering."


R=adamk@chromium.org, bmeurer@chromium.org, hablich@chromium.org
BUG=
NOTRY=true
==========

[ofrobots, 2017-01-19 00:37:21]

The CQ bit was checked by ofrobots@google.com

[commit-bot: I haz the power, 2017-01-19 00:37:29]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-01-19 00:39:26]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1484786241843440, "parent_rev":
"f3dcdf886213b9bb36fe40fbde85d4fca6587fb7", "commit_rev":
"89b7a4d7d0a2539c8e3d2422f90c6f6fbdecd0a6"}

[commit-bot: I haz the power, 2017-01-19 00:39:31]

Description was changed from

==========
Revert I+TF for lexical variables

This reverts the following commits to fix a Node.js regression:
* 5529430dec0d8997319d46e02c473a7a4faf1933 "[compiler] Consistently use
Ignition+TurboFan for lexical variables."
* 78691367163ad318a6b441f03de5e7f5bf400df7 "[compiler] Improve let+const
decision in AstNumbering."


R=adamk@chromium.org, bmeurer@chromium.org, hablich@chromium.org
BUG=
NOTRY=true
==========

to

==========
Revert I+TF for lexical variables

This reverts the following commits to fix a Node.js regression:
* 5529430dec0d8997319d46e02c473a7a4faf1933 "[compiler] Consistently use
Ignition+TurboFan for lexical variables."
* 78691367163ad318a6b441f03de5e7f5bf400df7 "[compiler] Improve let+const
decision in AstNumbering."


R=adamk@chromium.org, bmeurer@chromium.org, hablich@chromium.org
BUG=
NOTRY=true

Review-Url: https://codereview.chromium.org/2647523002
Cr-Commit-Position: refs/heads/master@{#42484}
Committed:
https://chromium.googlesource.com/v8/v8/+/89b7a4d7d0a2539c8e3d2422f90c6f6fbde...
==========

[commit-bot: I haz the power, 2017-01-19 00:39:32]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/v8/v8/+/89b7a4d7d0a2539c8e3d2422f90c6f6fbde...