WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
Bug 224171
[WTF] Introduce FixedVector and use it for FixedOperands
https://bugs.webkit.org/show_bug.cgi?id=224171
Summary
[WTF] Introduce FixedVector and use it for FixedOperands
Yusuke Suzuki
Reported
2021-04-03 22:15:56 PDT
[WTF] Introduce FixedVector and use it for FixedOperands
Attachments
Patch
(52.37 KB, patch)
2021-04-03 22:19 PDT
,
Yusuke Suzuki
no flags
Details
Formatted Diff
Diff
Patch
(52.51 KB, patch)
2021-04-03 23:07 PDT
,
Yusuke Suzuki
mark.lam
: review+
ews-feeder
: commit-queue-
Details
Formatted Diff
Diff
Patch
(53.51 KB, patch)
2021-04-05 12:25 PDT
,
Yusuke Suzuki
no flags
Details
Formatted Diff
Diff
Patch
(67.68 KB, patch)
2021-04-05 21:40 PDT
,
Yusuke Suzuki
ews-feeder
: commit-queue-
Details
Formatted Diff
Diff
Patch
(67.82 KB, patch)
2021-04-06 07:56 PDT
,
Yusuke Suzuki
ews-feeder
: commit-queue-
Details
Formatted Diff
Diff
Show Obsolete
(3)
View All
Add attachment
proposed patch, testcase, etc.
Yusuke Suzuki
Comment 1
2021-04-03 22:19:13 PDT
Created
attachment 425126
[details]
Patch
Yusuke Suzuki
Comment 2
2021-04-03 23:07:28 PDT
Created
attachment 425127
[details]
Patch
Mark Lam
Comment 3
2021-04-05 09:44:10 PDT
Comment on
attachment 425127
[details]
Patch View in context:
https://bugs.webkit.org/attachment.cgi?id=425127&action=review
r=me
> Source/JavaScriptCore/llint/LLIntOffsetsExtractor.cpp:90 > using RefCountedArray = WTF::RefCountedArray<int>;
I think we can remove this now.
> Source/JavaScriptCore/runtime/CachedTypes.cpp:625 > + void encode(Encoder& encoder, const FixedVector<SourceType<T>>& vector)
It's a pity that we're copy-pasting this code 3 times. Can we use just one template, and if needed, std::enable_if the container type to be one of the 3 acceptable types?
> Source/JavaScriptCore/runtime/CachedTypes.cpp:658 > + void decode(Decoder& decoder, FixedVector<SourceType<T>>& vector, Args... args) const
Ditto.
Yusuke Suzuki
Comment 4
2021-04-05 11:33:29 PDT
Comment on
attachment 425127
[details]
Patch View in context:
https://bugs.webkit.org/attachment.cgi?id=425127&action=review
>> Source/JavaScriptCore/llint/LLIntOffsetsExtractor.cpp:90 >> using RefCountedArray = WTF::RefCountedArray<int>; > > I think we can remove this now.
Unfortunately, LLInt still uses RefCountedArray::m_data in loadp CodeBlock::m_argumentValueProfiles + FixedVector::m_storage + RefCountedArray::m_data[t1], t3
>> Source/JavaScriptCore/runtime/CachedTypes.cpp:625 >> + void encode(Encoder& encoder, const FixedVector<SourceType<T>>& vector) > > It's a pity that we're copy-pasting this code 3 times. Can we use just one template, and if needed, std::enable_if the container type to be one of the 3 acceptable types?
Changed them to use templatized ones :)
>> Source/JavaScriptCore/runtime/CachedTypes.cpp:658 >> + void decode(Decoder& decoder, FixedVector<SourceType<T>>& vector, Args... args) const > > Ditto.
Fixed.
Yusuke Suzuki
Comment 5
2021-04-05 12:25:03 PDT
Created
attachment 425186
[details]
Patch
Yusuke Suzuki
Comment 6
2021-04-05 21:40:22 PDT
Created
attachment 425239
[details]
Patch
Yusuke Suzuki
Comment 7
2021-04-06 07:56:30 PDT
Created
attachment 425278
[details]
Patch
Yusuke Suzuki
Comment 8
2021-04-06 12:44:27 PDT
Existing flaky crashes.
Yusuke Suzuki
Comment 9
2021-04-06 12:45:55 PDT
mac-debug-wk1 is getting stuck.
Yusuke Suzuki
Comment 10
2021-04-06 12:47:57 PDT
Committed
r275542
(
236198@main
): <
https://commits.webkit.org/236198@main
>
Radar WebKit Bug Importer
Comment 11
2021-04-06 12:48:17 PDT
<
rdar://problem/76279966
>
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug