DFG TypeOf implementation should have its backend code aligned to what the CFA does
https://bugs.webkit.org/show_bug.cgi?id=109385
Reviewed by Sam Weinig.
The problem was that if we ended up trying to constant fold, but didn't succeed
because of prediction mismatches, then we would also fail to do filtration.
Rearranged the control flow in the CFA to fix that.
As far as I know, this is asymptomatic - it's sort of OK for the CFA to prove less
things, which is what the bug was.
* dfg/DFGAbstractState.cpp:
(JSC::DFG::AbstractState::execute):
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@142508 268f45cc-cd09-0410-ab3c-d52691b4dbfc
2 files changed
