Optimize add/set/delete operations for string keys in Maps and Sets

Optimize add/set/delete operations for string keys in Maps and Sets

This was previously landed in commit 66e2f60, but failed the
collections mjsunit test with --deopt-every-n=1 due to a typo
in the shrinking code. This patch corrects and simplifies the
shrinking logic, and the tests now pass.

R=dslomov@chromium.org

Committed: https://chromium.googlesource.com/v8/v8/+/8599f5f047b4400f7f61de9e449b2246c9aad471

[adamk, 2014-12-04 22:17:54]

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

[Dmitry Lomov (no reviews), 2014-12-05 13:38:45]

lgtm, but what was the typo?

[adamk, 2014-12-05 17:09:26]

On 2014/12/05 13:38:45, Dmitry Lomov (chromium) wrote:
> lgtm, but what was the typo?

There were actually two issues. One was the typo: loading NumberOfElements where
I claimed to be loading NumberOfBuckets.

The second was that I didn't subtract 1 from the num_elements before doing the
shrink check.

[adamk, 2014-12-05 17:09:29]

The CQ bit was checked by adamk@chromium.org

[commit-bot: I haz the power, 2014-12-05 17:10:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/777663003/1

[adamk, 2014-12-05 17:18:44]

The CQ bit was unchecked by adamk@chromium.org

[adamk, 2014-12-05 17:19:38]

On 2014/12/05 17:10:04, I haz the power (commit-bot) wrote:
> CQ is trying da patch. Follow status at
>  https://chromium-cq-status.appspot.com/patch-status/777663003/1

Realized I might have jumped the gun here with the CQ; happy to try to come up
with a small repro of the deopt failure (had to do with a delete causing a
"hole" to appear in a set.

[Dmitry Lomov (no reviews), 2014-12-05 20:07:12]

On 2014/12/05 17:19:38, adamk wrote:
> On 2014/12/05 17:10:04, I haz the power (commit-bot) wrote:
> > CQ is trying da patch. Follow status at
> >  https://chromium-cq-status.appspot.com/patch-status/777663003/1
> 
> Realized I might have jumped the gun here with the CQ; happy to try to come up
> with a small repro of the deopt failure (had to do with a delete causing a
> "hole" to appear in a set.

I didn't mean to not let you commit, we have tests that catch deopt failures,
although of course a targeted test will be nice.

[adamk, 2014-12-05 20:56:28]

On 2014/12/05 20:07:12, Dmitry Lomov (chromium) wrote:
> On 2014/12/05 17:19:38, adamk wrote:
> > On 2014/12/05 17:10:04, I haz the power (commit-bot) wrote:
> > > CQ is trying da patch. Follow status at
> > >  https://chromium-cq-status.appspot.com/patch-status/777663003/1
> > 
> > Realized I might have jumped the gun here with the CQ; happy to try to come
up
> > with a small repro of the deopt failure (had to do with a delete causing a
> > "hole" to appear in a set.
> 
> I didn't mean to not let you commit, we have tests that catch deopt failures,
> although of course a targeted test will be nice.

Okay, thanks, I'll commit and see if I can make a regression test in parallel.

[adamk, 2014-12-05 21:04:08]

Committed patchset #1 (id:1) manually as
8599f5f047b4400f7f61de9e449b2246c9aad471 (presubmit successful).

[adamk, 2014-12-05 22:24:16]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/742353006/ by adamk@chromium.org.

The reason for reverting is: Deopt fuzzer reports that something's still broken,
reverting until
I can create a minimal repro..