display: inline-block
Created attachment 414009 [details] for bots
Created attachment 414031 [details] for bots
Created attachment 414032 [details] for bots
Created attachment 414053 [details] Patch
Created attachment 414191 [details] Patch
Created attachment 414351 [details] Patch
Created attachment 414404 [details] Patch
Created attachment 414505 [details] Patch
Created attachment 414616 [details] Patch
Created attachment 414620 [details] Patch
Created attachment 414638 [details] Patch
Created attachment 414647 [details] Patch
Committed r270070: <https://trac.webkit.org/changeset/270070> All reviewed patches have been landed. Closing bug and clearing flags on attachment 414647 [details].
<rdar://problem/71615879>
Reverted r270070 for reason: broke an API test Committed r270108: <https://trac.webkit.org/changeset/270108>
Committed r270123: <https://trac.webkit.org/changeset/270123>
Looks like this was a 9% - 11% regression on StyleBench. You can see the initial regression when the patch landed, then a return to normal, then another regression when the patch landed a second time. The bot is a little confusing because the runloop regression landed immediately after. But that regression has been rolled out, and this one remains visible on the bot. https://perf.webkit.org/v3/#/charts?since=1600721032751&zoom=(1605422135834.8188-1606923328416.2063)&paneList=((21-1649-(1605785343399.1538-1605903635771.0242)-null-(5-2.5-500))-(21-1006-null-null-(5-2.5-500))-(21-1019-null-null-(5-2.5-500))-(21-1032)-(21-1045-null-null-(5-2.5-500))-(21-1407-null-null-(5-2.5-500))-(21-1648-null-null-(5-2.5-500))-(21-1974-null-null-(5-2.5-500)))