Bunch of octane tests hit this.

Some refactoring of the testObjectTruthy code found in ion's CodeGenerator.cpp is involved. If you think it's appropriate, I can check that in separately to inbound, and push the baseline-specific changes after that gets merged back.

Comment on attachment 719693 [details] [diff] [review] Patch Review of attachment 719693 [details] [diff] [review]: ----------------------------------------------------------------- (In reply to Kannan Vijayan [:djvj] from comment #1) > Some refactoring of the testObjectTruthy code found in ion's > CodeGenerator.cpp is involved. If you think it's appropriate, I can check > that in separately to inbound, and push the baseline-specific changes after > that gets merged back. Should be fine to land on IonMonkey. I don't expect changes to this code in the coming weeks, and even if there are it should be easy to merge them. ::: js/src/ion/BaselineIC.cpp @@ +1909,5 @@ > + Register objReg = masm.extractObject(R0, ExtractTemp0); > + Register scratch = R1.scratchReg(); > + masm.branchTestObjectTruthy(Assembler::NotEqual, objReg, scratch, &ifFalse, &slowPath); > + > + // If object doesn't emulate undefined, it evaulates to true. Nit: typo, evaluates

Pushed with a minor fix to ToBool(Object) - wasn't notting the result of the EmulatesUndefined ABI call before boxing: https://hg.mozilla.org/projects/ionmonkey/rev/80642ec465bf Waiting for tbpl before closing