[webkit-reviews] review granted: [Bug 190923] Re-introduce op_bitnot : [Attachment 355636] Patch

[bugzilla-daemon at webkit.org]

Yusuke Suzuki <yusukesuzuki at slowstart.org> has granted Caio Lima
<ticaiolima at gmail.com>'s request for review:
Bug 190923: Re-introduce op_bitnot
https://bugs.webkit.org/show_bug.cgi?id=190923

Attachment 355636: Patch

https://bugs.webkit.org/attachment.cgi?id=355636&action=review




--- Comment #35 from Yusuke Suzuki <yusukesuzuki at slowstart.org> ---
Comment on attachment 355636
  --> https://bugs.webkit.org/attachment.cgi?id=355636
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=355636&action=review

r=me with nits. Please ensure it does not cause any performance regression
before landing.

> Source/JavaScriptCore/dfg/DFGAbstractInterpreterInlines.h:388
> +	       setConstant(node, JSValue(~a));

Insert `break;` here.

> Source/JavaScriptCore/dfg/DFGNodeType.h:114
> +    macro(ArithBitNot, NodeResultInt32) \

Add `NodeMustGenerate` since the Untyped case can cause anything.
And clear `NodeMustGenerate` when the edge is fixed by `Int32` in the fixup
phase.
Later, if you add ValueBitNot, then you can make this non-MustGenerate, and
handle Untyped cases in ValueBitNot.