Speedup access to global_proxy.* attributes/accessors.

Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower than evaluating just 'f'. This CL aims to perform the necessary checks at compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716, chromium:656959
Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Committed: https://crrev.com/36f3f90907bea457591c6c484fb554b737bbaeac
Cr-Original-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Original-Commit-Position: refs/heads/master@{#40365}
Cr-Commit-Position: refs/heads/master@{#40671}

[vogelheim, 2016-10-10 13:50:41]

vogelheim@chromium.org changed reviewers:
+ jochen@chromium.org, verwaest@chromium.org

[vogelheim, 2016-10-10 13:50:41]

ptal.

This is essentially the same as the - already reviewed & lgtm-ed -
crrev.com/72369933005. It's rebase to current tip-of-tree and addresses the
remaining nitpicks.

All 'real' work here is due to peterssen@.

[jochen (gone - plz use gerrit), 2016-10-10 13:57:18]

note that this merges the approved CL from Alfonso to speed up global property
access with the almost ready CL to add surrogates. Is that on purpose? I think
the former could be just rebased and landed.

https://codereview.chromium.org/2403003002/diff/20001/src/crankshaft/hydrogen.cc
File src/crankshaft/hydrogen.cc (right):

https://codereview.chromium.org/2403003002/diff/20001/src/crankshaft/hydrogen...
src/crankshaft/hydrogen.cc:5495: LookupIterator TryReadingFromCache(Isolate*
isolate, LookupIterator* it) {
can this function be moved to LookupIterator, so it can be used from ic.cc and
objects.cc as well?

[vogelheim, 2016-10-10 14:59:23]

Patchset #1 (id:1) has been deleted

[vogelheim, 2016-10-10 14:59:30]

Patchset #1 (id:20001) has been deleted

[vogelheim, 2016-10-10 15:09:12]

ptal. I've now properly re-created the original patch set, plus one additional
nitpick. Patch Set 1 should now be the one you already saw.

On 2016/10/10 13:57:18, jochen wrote:
> note that this merges the approved CL from Alfonso to speed up global property
> access with the almost ready CL to add surrogates. Is that on purpose? I think
> the former could be just rebased and landed.
> 
>
https://codereview.chromium.org/2403003002/diff/20001/src/crankshaft/hydrogen.cc
> File src/crankshaft/hydrogen.cc (right):
> 
>
https://codereview.chromium.org/2403003002/diff/20001/src/crankshaft/hydrogen...
> src/crankshaft/hydrogen.cc:5495: LookupIterator TryReadingFromCache(Isolate*
> isolate, LookupIterator* it) {
> can this function be moved to LookupIterator, so it can be used from ic.cc and
> objects.cc as well?

Makes sense; but this functionality is no longer in this CL. Therefore not done.

[jochen (gone - plz use gerrit), 2016-10-10 15:17:31]

lgtm

[vogelheim, 2016-10-10 15:28:08]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-10 15:28:10]

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, 2016-10-10 15:54:19]

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

[commit-bot: I haz the power, 2016-10-10 15:54:19]

Dry run: This issue passed the CQ dry run.

[Toon Verwaest, 2016-10-11 06:52:21]

lgtm

[vogelheim, 2016-10-11 08:19:47]

The CQ bit was checked by vogelheim@chromium.org

[commit-bot: I haz the power, 2016-10-11 08:19:54]

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, 2016-10-11 08:22:10]

Committed patchset #3 (id:80001)

[commit-bot: I haz the power, 2016-10-11 08:22:28]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.


BUG=chromium:634276
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

[commit-bot: I haz the power, 2016-10-11 08:22:28]

Patchset 3 (id:??) landed as
https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}

[Michael Achenbach, 2016-10-11 12:48:31]

A revert of this CL (patchset #3 id:80001) has been created in
https://codereview.chromium.org/2408133002/ by machenbach@chromium.org.

The reason for reverting is: Blocks roll:
https://codereview.chromium.org/2406213002/.

[vogelheim, 2016-10-13 14:19:29]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

[vogelheim, 2016-10-13 14:29:39]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-13 14:29:51]

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

[vogelheim, 2016-10-13 14:47:45]

Patchset #4 (id:100001) has been deleted

[vogelheim, 2016-10-13 14:47:52]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-13 14:47:57]

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, 2016-10-13 15:27:22]

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

[commit-bot: I haz the power, 2016-10-13 15:27:23]

Dry run: This issue passed the CQ dry run.

[vogelheim, 2016-10-13 15:41:46]

ptal. Preparing to re-land.

It seems a bit odd to put the ...->ReturnValue into BuildNamedAccess and then
return nullptr, but neither putting it 'higher' or 'lower' in the callstack
seemed to work out.

[vogelheim, 2016-10-14 08:02:10]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

[jochen (gone - plz use gerrit), 2016-10-14 08:11:48]

lgtm

[Toon Verwaest, 2016-10-14 08:15:32]

https://codereview.chromium.org/2403003002/diff/140001/src/crankshaft/hydroge...
File src/crankshaft/hydrogen.cc (right):

https://codereview.chromium.org/2403003002/diff/140001/src/crankshaft/hydroge...
src/crankshaft/hydrogen.cc:7740: }
You do this since the previous inline call already added the nide to the graph I
presume. You could also just return the node without adding?

[vogelheim, 2016-10-14 08:50:50]

https://codereview.chromium.org/2403003002/diff/140001/src/crankshaft/hydroge...
File src/crankshaft/hydrogen.cc (right):

https://codereview.chromium.org/2403003002/diff/140001/src/crankshaft/hydroge...
src/crankshaft/hydrogen.cc:7740: }
On 2016/10/14 08:15:32, Toon Verwaest wrote:
> You do this since the previous inline call already added the nide to the graph
I
> presume. You could also just return the node without adding?

Well, sort of.

The InlineGlobalPropertyStore adds the instruction, and adds an HSimulate thing
after it. I assume the order must be preserved, so if I don't add the
instruction there, I also cannot add the HSimulate, which means I need to copy
that code to each call site.

It also just seems odd, since OnlineGlobalPropertyStore potentially adds other,
previous instructions, too, so it would then add most of its instructions,
except the last one.


The reason why I didn't put the ReturnValue thing into InlineGlobalPropertyStore
is that one call site needs this, and the other doesn't.


I'm trying a version that just returns the node, but so far it looks to be
uglyfying the code.

[vogelheim, 2016-10-14 11:42:47]

This would be fairly elegant if I could just use
ast_context()->ReturnInstruction(..) in the caller - which almost does
everything we need. But that would return the value of of HStoreNamedField, not
the value that goes into the store, which causes later tests to fail.

Am I overlooking something?

[Toon Verwaest, 2016-10-14 12:03:38]

The simulates are also added by the caller I thought? perhaps I'm overlooking
something?

[vogelheim, 2016-10-17 09:45:12]

On 2016/10/14 12:03:38, Toon Verwaest wrote:
> The simulates are also added by the caller I thought? perhaps I'm overlooking
> something?

I don't get this.

The part I'm worried is at the end of InlineGlobalPropertyStore (lines 6746 -
6751). That adds a HSimulate (line 6750) in the callee, not the caller. If I
want to return that HStoreNamedField without adding it, then the caller (or the
caller's caller) will have to add the HSimulate that currently gets added at the
end of InlineGlobalPropertyStore.

One usage of InlineGlobalPropertyStore (HandleGlobalVariableAssignment) doesn't
add an HSimulate, the other usage of BuildNamedAccess... depends.
BuildNamedAccess has 3 usages, one - BuildStore - does exactly what we need. The
other two... HandleKeyedElementAccess + BuildLoad - seems to do something subtly
different.

If I'm seeing this correctly, I'll need to consider 4 callers:
- HandleGlobalVariableAssignment -> InlineGlobalPropertyStore
- BuildStore -> BuildNamedAccess -> InlineGlobalPropertyStore
- HandleKeyedElementAccess -> BuildNamedAccess -> InlineGlobalPropertyStore
- BuildLoad -> BuildNamedAccess -> InlineGlobalPropertyStore

For the 2nd, it should all work out elegantly, because all the necessary logic
is already in place. For the other 3, I'll have to somehow compensate. The best
idea I've had was to copy the HSimulate+ReturnValue logic into those three
callers, but that seems... unfortunate.


I do wonder whether we're talking about the same thing, though.. :-/  Is the
above what you meant? If so, am I overlooking something for the case 1 + 3 + 4?

[vogelheim, 2016-10-17 10:16:13]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-17 10:16:24]

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, 2016-10-17 10:40:19]

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

[commit-bot: I haz the power, 2016-10-17 10:40:20]

Dry run: This issue passed the CQ dry run.

[vogelheim, 2016-10-17 12:17:23]

ptal. I think I got it now.

[Toon Verwaest, 2016-10-17 13:29:30]

lgtm

[vogelheim, 2016-10-17 13:33:28]

The CQ bit was checked by vogelheim@chromium.org

[vogelheim, 2016-10-17 13:33:29]

The patchset sent to the CQ was uploaded after l-g-t-m from jochen@chromium.org
Link to the patchset: https://codereview.chromium.org/2403003002/#ps160001
(title: "Return instruction from InlineGlobalPropertyStore")

[commit-bot: I haz the power, 2016-10-17 13:33:38]

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, 2016-10-17 13:36:17]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

[commit-bot: I haz the power, 2016-10-17 13:36:18]

Committed patchset #6 (id:160001)

[commit-bot: I haz the power, 2016-10-17 13:37:13]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Cr-Commit-Position: refs/heads/master@{#40153}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

[commit-bot: I haz the power, 2016-10-17 13:37:16]

Patchset 6 (id:??) landed as
https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Commit-Position: refs/heads/master@{#40365}

[vogelheim, 2016-10-20 16:09:38]

A revert of this CL (patchset #6 id:160001) has been created in
https://codereview.chromium.org/2434233002/ by vogelheim@chromium.org.

The reason for reverting is: Revert, because of crbug.com/656959..

[vogelheim, 2016-10-25 17:09:28]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

[vogelheim, 2016-10-25 17:27:14]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716, chromium:656959

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

[vogelheim, 2016-10-25 17:33:04]

ptal.

Due to the rebase the exact difference to the last version is a bit difficult to
see; search for "is_same_context_global_proxy_access".

I've updated crbug.com/656959 with a bit of the reasoning behind this. The gist
is this: The old version used isolate->MayAccess to check whether to inline.
This was very secure, and also semantically incorrect when the access was
cross-context and we'd the inline-ing would inline a presumed result from the
'wrong' context.

This version checks whether the contexts are identical, and just defaults to the
slow case if they are different at all.

... hope this makes sense... :)

[vogelheim, 2016-10-25 17:33:14]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-25 17:33:24]

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, 2016-10-25 18:10:08]

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

[commit-bot: I haz the power, 2016-10-25 18:10:09]

Dry run: This issue passed the CQ dry run.

[jochen (gone - plz use gerrit), 2016-10-26 08:17:16]

at first, I thought, maybe we can just always load the constant out of the CL,
as long as the contexts' maps are equal. However, one context might lose access
to another context (via setting document.domain), so just doing this for
same-context makes sense.

lgtm

[vogelheim, 2016-10-26 09:37:34]

On 2016/10/26 08:17:16, jochen wrote:
> at first, I thought, maybe we can just always load the constant out of the CL,
> as long as the contexts' maps are equal. However, one context might lose
access
> to another context (via setting document.domain), so just doing this for
> same-context makes sense.


Hmmm... like: Keep the MayAccess check, but instead of
current_context->global_proxy() use map_context->global_proxy()?

I didn't think of that, but if we keep MayAccess, that would also work, wouldn't
it?

[vogelheim, 2016-10-26 11:19:47]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-26 11:19:55]

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

[vogelheim, 2016-10-26 11:20:34]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-26 11:20:45]

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

[vogelheim, 2016-10-26 11:24:02]

Patch Set 8 is a slightly cleaned version of the previous code.

Patch Set 9 is a version that inlines the global object from the *map* context,
rather the current one. (I also best-guessed a sligthly different
MayAccess-check; please do have a look.)

I need help deciding between the two. What do you think? Provided further
testing confirms both work, which one should we use?

[commit-bot: I haz the power, 2016-10-26 11:49:04]

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

[commit-bot: I haz the power, 2016-10-26 11:49:05]

Dry run: This issue passed the CQ dry run.

[jochen (gone - plz use gerrit), 2016-10-26 13:10:14]

I think patchset 8 is correct(er)

[vogelheim, 2016-10-31 11:52:04]

Patchset #9 (id:220001) has been deleted

[vogelheim, 2016-10-31 11:52:15]

The CQ bit was checked by vogelheim@chromium.org

[vogelheim, 2016-10-31 11:52:15]

The patchset sent to the CQ was uploaded after l-g-t-m from
verwaest@chromium.org, jochen@chromium.org
Link to the patchset: https://codereview.chromium.org/2403003002/#ps200001
(title: "Slightly simplify & prettify the code. No functionality changes.")

[commit-bot: I haz the power, 2016-10-31 11:52:27]

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, 2016-10-31 11:53:37]

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

[commit-bot: I haz the power, 2016-10-31 11:53:39]

Try jobs failed on following builders:
  v8_linux64_asan_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_asan_rel_ng/buil...)
  v8_linux64_avx2_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng/buil...)
  v8_linux64_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_rel_ng/builds/15429)
  v8_linux_arm64_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm64_rel_ng/build...)
  v8_linux_arm_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng/builds/...)
  v8_linux_dbg_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_dbg_ng/builds/15459)
  v8_linux_gcc_compile_rel on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_gcc_compile_rel/bu...)
  v8_linux_mipsel_compile_rel on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_mipsel_compile_rel...)
  v8_linux_nodcheck_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_nodcheck_rel_ng/bu...)
  v8_mac_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_mac_rel_ng/builds/11713)
  v8_presubmit on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/27570)

[vogelheim, 2016-10-31 12:24:45]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-31 12:24:49]

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, 2016-10-31 12:48:20]

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

[commit-bot: I haz the power, 2016-10-31 12:48:21]

Dry run: This issue passed the CQ dry run.

[vogelheim, 2016-10-31 13:24:41]

The CQ bit was checked by vogelheim@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-31 13:24:51]

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, 2016-10-31 13:51:31]

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

[commit-bot: I haz the power, 2016-10-31 13:51:32]

Dry run: This issue passed the CQ dry run.

[vogelheim, 2016-10-31 14:25:52]

The CQ bit was checked by vogelheim@chromium.org

[vogelheim, 2016-10-31 14:25:52]

The patchset sent to the CQ was uploaded after l-g-t-m from jochen@chromium.org,
verwaest@chromium.org
Link to the patchset: https://codereview.chromium.org/2403003002/#ps260001
(title: "Check for nullptr return in InlineGlobalPropertyStore.")

[commit-bot: I haz the power, 2016-10-31 14:26:00]

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, 2016-10-31 14:28:15]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716, chromium:656959

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716, chromium:656959

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

[commit-bot: I haz the power, 2016-10-31 14:28:16]

Committed patchset #10 (id:260001)

[commit-bot: I haz the power, 2016-11-17 22:18:03]

Description was changed from

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716, chromium:656959

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Cr-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Commit-Position: refs/heads/master@{#40365}
==========

to

==========
Speedup access to global_proxy.* attributes/accessors.

Using a global proxy (e.g. 'window.f', 'w.f' or 'this.f') is considerably slower
than evaluating just 'f'. This CL aims to perform the necessary checks at
compile time and inline the accesses.

This is a follow-on CL to crrev.com/2369933005:
- The initial upload is crrev.com/2369933005 + a rebase.
- The remaining issues are the fixes requested by the reviewers on that CL.

BUG=chromium:634276, chromium:654716, chromium:656959

Committed: https://crrev.com/8f43d748272536117008aa6a1b53ea52126261c1
Committed: https://crrev.com/041314524952a3c1bc71bd3beafbbb37319f1d22
Committed: https://crrev.com/36f3f90907bea457591c6c484fb554b737bbaeac
Cr-Original-Original-Commit-Position: refs/heads/master@{#40153}
Cr-Original-Commit-Position: refs/heads/master@{#40365}
Cr-Commit-Position: refs/heads/master@{#40671}
==========

[commit-bot: I haz the power, 2016-11-17 22:18:05]

Patchset 10 (id:??) landed as
https://crrev.com/36f3f90907bea457591c6c484fb554b737bbaeac
Cr-Commit-Position: refs/heads/master@{#40671}