[webkit-reviews] review denied: [Bug 195279] DFG should not force KnownInt32 edges during fixup : [Attachment 363642] Patch

[bugzilla-daemon at webkit.org]

Saam Barati <sbarati at apple.com> has denied Tadeu Zagallo <tzagallo at apple.com>'s
request for review:
Bug 195279: DFG should not force KnownInt32 edges during fixup
https://bugs.webkit.org/show_bug.cgi?id=195279

Attachment 363642: Patch

https://bugs.webkit.org/attachment.cgi?id=363642&action=review




--- Comment #7 from Saam Barati <sbarati at apple.com> ---
Comment on attachment 363642
  --> https://bugs.webkit.org/attachment.cgi?id=363642
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=363642&action=review

You need to make sure we do the right thing for these nodes in the other parts
of the compiler. As an example, you need to vet and make sure we don't switch
on KnownInt32use (and now it's Int32Use). Another example, you need to make
SpeculateStrictInt32Operand SpeculateInt32Operand for
GetEnumeratorStructurePname in DFGSpeculativeJIT

> Source/JavaScriptCore/ChangeLog:11
> +	   For example, if the incoming value is a local and its underlying
slot is reused,

I don't think the reused bit here is what's interesting. This can happen for
general reasons.

> Source/JavaScriptCore/ChangeLog:12
> +	   its type might be widened and it might no longer satify this
requirement.

also widened isn't exactly correct. We may pick a different type for it.