Bug 206329 – Try to simplify the template deduction used by callOperation in DFGSpeculativeJIT

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Robin Morisset]

Might slightly help compile times and be more readable to boot.

[Robin Morisset]

Created attachment 387879 [details]
Patch

[Keith Miller]

Comment on attachment 387879 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=387879&action=review

r=me with question.

> Source/JavaScriptCore/dfg/DFGSpeculativeJIT.h:-954
> -    callOperation(OperationType operation, NoResultTag, Args... args)

Does anyone use NoResultTag now? If not, we should just get rid of it.

[Robin Morisset]

(In reply to Keith Miller from comment #2)
> Comment on attachment 387879 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=387879&action=review
> 
> r=me with question.
> 
> > Source/JavaScriptCore/dfg/DFGSpeculativeJIT.h:-954
> > -    callOperation(OperationType operation, NoResultTag, Args... args)
> 
> Does anyone use NoResultTag now? If not, we should just get rid of it.

It is still used by various other functions. It is not clear whether it can also be removed from those.

[WebKit Commit Bot]

Comment on attachment 387879 [details]
Patch

Clearing flags on attachment: 387879

Committed r254717: <https://trac.webkit.org/changeset/254717>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/58662089>

[Saam Barati]

Did it make compile times faster?

[Robin Morisset]

(In reply to Saam Barati from comment #7)
> Did it make compile times faster?

I was hoping that it would, but it apparently did not (or at least not significantly):
https://perf.safari.apple.com/v3/#/charts?since=1578678298555&paneList=((1117-4968-null-null-(5-2.5-500))-(1117-4990-null-null-(5-2.5-500)))