...
Created attachment 292268 [details] patch
Comment on attachment 292268 [details] patch Looks good!
Comment on attachment 292268 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=292268&action=review r=me > Source/JavaScriptCore/ChangeLog:11 > + that it allocated for it after the TryGetById node executed. Yeah, if base's use count remains, we still should keep this in that register!
Comment on attachment 292268 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=292268&action=review > JSTests/stress/try-get-by-id-should-spill-registers-dfg.js:10 > +let f = createBuiltin(`(function (arg) { > + let r = @tryGetById(arg, "prototype"); > + if (arg !== true) throw new @Error("Bad clobber of arg"); > + return r; > + })`); > +noInline(f); > + > +for (let i = 0; i < 10000; i++) { > + f(true); > +} Looks like you used tabs instead of spaces?
Comment on attachment 292268 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=292268&action=review >> JSTests/stress/try-get-by-id-should-spill-registers-dfg.js:10 >> +} > > Looks like you used tabs instead of spaces? I don't think I did but my indentation is a bit off. I'll fix before landing.
Created attachment 292403 [details] patch for landing
Comment on attachment 292403 [details] patch for landing Clearing flags on attachment: 292403 Committed r207697: <http://trac.webkit.org/changeset/207697>
All reviewed patches have been landed. Closing bug.