[webkit-reviews] review granted: [Bug 213106] Bytecode liveness should be aware of checkpoints : [Attachment 401769] Patch

[bugzilla-daemon at webkit.org]

Keith Miller <keith_miller at apple.com> has granted  review:
Bug 213106: Bytecode liveness should be aware of checkpoints
https://bugs.webkit.org/show_bug.cgi?id=213106

Attachment 401769: Patch

https://bugs.webkit.org/attachment.cgi?id=401769&action=review




--- Comment #10 from Keith Miller <keith_miller at apple.com> ---
Comment on attachment 401769
  --> https://bugs.webkit.org/attachment.cgi?id=401769
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=401769&action=review

>> Source/JavaScriptCore/bytecode/BytecodeLivenessAnalysis.cpp:178
>> +Vector<Operand, maxNumCheckpointTmps> livenessForCheckpoint(const
CodeBlock& codeBlock, BytecodeIndex bytecodeIndex)
> 
> It feels like this should take a scoped lambda rather than return a vector.
Almost every case is just going to forward it to someone else.

After talking offline, I'm cool with this design if we have an assert we didn't
exceed inline capacity before returning the Vector. There may also be a better
WTF container type for this, might be worth a look.