Just for our own sanity when reasoning about what the slow paths do, we don't want a random value in property when the base isn't a cell.
Created attachment 436926 [details] patch
Comment on attachment 436926 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=436926&action=review r=me too > Source/JavaScriptCore/jit/JITPropertyAccess.cpp:1519 > emitArrayProfilingSiteWithCell(regT0, profile, regT2); Can you also ensure that AccessCase IC code for InByVal / InById reserves the above registers if we go to the slow path?
(In reply to Yusuke Suzuki from comment #2) > Comment on attachment 436926 [details] > patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=436926&action=review > > r=me too > > > Source/JavaScriptCore/jit/JITPropertyAccess.cpp:1519 > > emitArrayProfilingSiteWithCell(regT0, profile, regT2); > > Can you also ensure that AccessCase IC code for InByVal / InById reserves > the above registers if we go to the slow path? Confirmed that they do not clobber these registers.
Committed r281826 (241160@main): <https://commits.webkit.org/241160@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 436926 [details].
<rdar://problem/82600217>