It's like an 8% speedup in metal compile time
Created attachment 376919 [details] patch
Comment on attachment 376919 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=376919&action=review > Source/WebCore/ChangeLog:12 > + 4. Lazily produce lvalues for GlobalVariableReference. I forgot to add a comment her on speedup. Will do.
Comment on attachment 376919 [details] patch Clearing flags on attachment: 376919 Committed r248994: <https://trac.webkit.org/changeset/248994>
All reviewed patches have been landed. Closing bug.
<rdar://problem/54583805>
Comment on attachment 376919 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=376919&action=review >> Source/WebCore/ChangeLog:12 >> + 4. Lazily produce lvalues for GlobalVariableReference. > > I forgot to add a comment her on speedup. Will do. This got committed by the commit queue. For posterity, I was seeing a 7-8% speedup in metal compile times with p=0.0001