[es6] Initial support for let/const bindings in sloppy mode

[es6] Initial support for let/const bindings in sloppy mode

Allow let in sloppy mode with --harmony-sloppy

Allow ES'15 const in sloppy mode with --harmony-sloppy --no-legacy-const

Functions in block are not done yet. They are only let bound in the block
at this point.

BUG=v8:3305, v8:2198
LOG=N
R=littledan@chromium.org, rossberg@chromium.org, adamk@chromium.org

Committed: https://crrev.com/3b1aabc960ea7a1107c8b6ebb8f2b4ce48e9b610
Cr-Commit-Position: refs/heads/master@{#29536}

[arv (Not doing code reviews), 2015-07-07 21:57:35]

This depends on https://codereview.chromium.org/1218803006/

[arv (Not doing code reviews), 2015-07-07 21:58:06]

On 2015/07/07 21:57:35, arv wrote:
> This depends on https://codereview.chromium.org/1218803006/

I see it landed... rebasing.

[arv (Not doing code reviews), 2015-07-07 22:16:26]

Cleanup

[arv (Not doing code reviews), 2015-07-07 22:19:41]

PTAL

I'm not yet sure what is going on with eval. It creates a DYNAMIC_LOCAL.

I'm sure there are a million edge cases related to eval that are not covered in
these tests since these are just the strict mode tests but I thought it was a
good milestone.

https://codereview.chromium.org/1219853004/diff/20001/test/cctest/test-parsin...
File test/cctest/test-parsing.cc (left):

https://codereview.chromium.org/1219853004/diff/20001/test/cctest/test-parsin...
test/cctest/test-parsing.cc:6778: "const x",
This is not valid in ES'15

https://codereview.chromium.org/1219853004/diff/20001/test/mjsunit/harmony/bl...
File test/mjsunit/harmony/block-let-semantics-sloppy.js (right):

https://codereview.chromium.org/1219853004/diff/20001/test/mjsunit/harmony/bl...
test/mjsunit/harmony/block-let-semantics-sloppy.js:98: // TestAll('function f()
{ eval("var y = 2;"); x = 1; }; f(); let x;');
There is something going on with eval scopes here...

[arv (Not doing code reviews), 2015-07-07 22:20:03]

arv@chromium.org changed reviewers:
+ adamk@chromium.org

[rossberg, 2015-07-08 11:49:56]

LGTM modulo comments.

I'm not too happy with duplicating all the tests via copy & paste, but I see
that some would probably be difficult to parameterise.

https://codereview.chromium.org/1219853004/diff/40001/test/cctest/test-parsin...
File test/cctest/test-parsing.cc (right):

https://codereview.chromium.org/1219853004/diff/40001/test/cctest/test-parsin...
test/cctest/test-parsing.cc:6787: static const ParserFlag always_flags[] =
{kNoLegacyConst};
It seems a bit weird to run the LegacyConst test with NoLegacyConst, in
particular when it requires removing one (currently valid) legacy const test.
Can you create a separate ConstSloppy test instead?

https://codereview.chromium.org/1219853004/diff/40001/test/mjsunit/harmony/bl...
File test/mjsunit/harmony/block-leave-sloppy.js (left):

https://codereview.chromium.org/1219853004/diff/40001/test/mjsunit/harmony/bl...
test/mjsunit/harmony/block-leave-sloppy.js:178: // from with.
Not your change, but I just noticed this comment and don't understand what
'with' it is talking about. Probably means 'eval'?

https://codereview.chromium.org/1219853004/diff/40001/test/mjsunit/harmony/bl...
File test/mjsunit/harmony/block-let-semantics-sloppy.js (right):

https://codereview.chromium.org/1219853004/diff/40001/test/mjsunit/harmony/bl...
test/mjsunit/harmony/block-let-semantics-sloppy.js:98: // TestAll('function f()
{ eval("var y = 2;"); x = 1; }; f(); let x;');
Can you add a TODO explaining why this is commented out right now? And file a
bug, too, so it doesn't get lost.

[arv (Not doing code reviews), 2015-07-08 13:43:31]

Updated tests and comments

[arv (Not doing code reviews), 2015-07-08 13:43:45]

https://codereview.chromium.org/1219853004/diff/40001/test/cctest/test-parsin...
File test/cctest/test-parsing.cc (right):

https://codereview.chromium.org/1219853004/diff/40001/test/cctest/test-parsin...
test/cctest/test-parsing.cc:6787: static const ParserFlag always_flags[] =
{kNoLegacyConst};
On 2015/07/08 11:49:56, rossberg wrote:
> It seems a bit weird to run the LegacyConst test with NoLegacyConst, in
> particular when it requires removing one (currently valid) legacy const test.
> Can you create a separate ConstSloppy test instead?

Agreed and done.

https://codereview.chromium.org/1219853004/diff/40001/test/mjsunit/harmony/bl...
File test/mjsunit/harmony/block-let-semantics-sloppy.js (right):

https://codereview.chromium.org/1219853004/diff/40001/test/mjsunit/harmony/bl...
test/mjsunit/harmony/block-let-semantics-sloppy.js:98: // TestAll('function f()
{ eval("var y = 2;"); x = 1; }; f(); let x;');
On 2015/07/08 11:49:56, rossberg wrote:
> Can you add a TODO explaining why this is commented out right now? And file a
> bug, too, so it doesn't get lost.

Done.

[arv (Not doing code reviews), 2015-07-08 13:45:50]

I also agree that duplicating these tests is sad.

One option might be to introduce a way for the test runner to expand the
variants similarly to how we do it for test262.

[rossberg, 2015-07-08 13:51:59]

https://codereview.chromium.org/1219853004/diff/60001/test/mjsunit/harmony/bl...
File test/mjsunit/harmony/block-let-semantics-sloppy.js (right):

https://codereview.chromium.org/1219853004/diff/60001/test/mjsunit/harmony/bl...
test/mjsunit/harmony/block-let-semantics-sloppy.js:98: // TODO(arv):
https://code.google.com/p/v8/issues/detail?id=4284
Nit: feel free to put in littledan.

[arv (Not doing code reviews), 2015-07-08 14:58:44]

The CQ bit was checked by arv@chromium.org

[arv (Not doing code reviews), 2015-07-08 14:58:45]

The patchset sent to the CQ was uploaded after l-g-t-m from
rossberg@chromium.org
Link to the patchset: https://codereview.chromium.org/1219853004/#ps60001
(title: "Updated tests and comments")

[commit-bot: I haz the power, 2015-07-08 14:58:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1219853004/60001

[commit-bot: I haz the power, 2015-07-08 15:04:09]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-07-08 15:04:25]

Patchset 4 (id:??) landed as
https://crrev.com/3b1aabc960ea7a1107c8b6ebb8f2b4ce48e9b610
Cr-Commit-Position: refs/heads/master@{#29536}