[Webkit-unassigned] [Bug 156147] [JSC] implement async functions proposal

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=156147

--- Comment #25 from Saam Barati <sbarati at apple.com> ---
(In reply to comment #24)
> (In reply to comment #23)
> > (In reply to comment #19)
> > > Created attachment 275529 [details]
> > > Patch
> > > 
> > > don't tokenize async/await as keywords, hopefully not costing too much perf
> > 
> > Can you test parser performance on this? Specifically, octane's
> > jquery/closure benchmarks.
> > 
> > I'm working on making our parser faster as we've
> > regressed 10% since we've implemented arrow functions
> > and I don't want to introduce anything that slows the
> > parser down.
> 
> It was regressing CodeLoad by quite a lot. I've managed to get it down to
> about a ~5% regression on average with some improvements for arrow function
> and method parsing, and can probably squeeze a few more out of it in other
> ways.

Based on this information I think it's worth having a compile
time flag instead of a runtime flag. That way async functions
are disabled we don't slow down the parser. Unless the runtime
flag doesn't slow down the parser with async functions disabled.

That said, I think it's worth getting the performance regression as down
as possible because we will presumably remove the runtime/compile time flag
at some point.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160405/2aae6c70/attachment.html>