Patch forthcoming.
Created attachment 290283 [details] it works on simple things
Created attachment 290355 [details] the patch
Comment on attachment 290355 [details] the patch View in context: https://bugs.webkit.org/attachment.cgi?id=290355&action=review > Source/JavaScriptCore/b3/air/AirFixObviousSpills.cpp:194 > - > + I'll revert.
Created attachment 290358 [details] the patch
Comment on attachment 290358 [details] the patch Attachment 290358 [details] did not pass mac-debug-ews (mac): Output: http://webkit-queues.webkit.org/results/2176850 New failing tests: js/slow-stress/call-spread.html imported/w3c/canvas/2d.text.draw.fill.maxWidth.negative.html imported/w3c/canvas/2d.gradient.interpolate.zerosize.fillText.html fast/forms/input-maxlength.html fast/workers/worker-cloneport.html js/slow-stress/nested-function-parsing-random.html js/slow-stress/new-spread.html js/dom/stack-trace.html imported/w3c/canvas/2d.gradient.interpolate.zerosize.strokeText.html inspector/debugger/call-frame-this-strict.html js/dfg-inline-function-dot-caller.html inspector/debugger/call-frame-this-nonstrict.html js/slow-stress/array-prototype-filter.html inspector/debugger/breakpoint-columns.html fast/forms/input-implicit-length-limit.html webgl/1.0.2/conformance/uniforms/out-of-bounds-uniform-array-access.html
Created attachment 290366 [details] Archive of layout-test-results from ews117 for mac-yosemite The attached test failures were seen while running run-webkit-tests on the mac-debug-ews. Bot: ews117 Port: mac-yosemite Platform: Mac OS X 10.10.5
Comment on attachment 290358 [details] the patch View in context: https://bugs.webkit.org/attachment.cgi?id=290358&action=review Looks good to me but the EWS result indicates a bug. > Source/JavaScriptCore/b3/B3MoveConstants.cpp:146 > + if (criteria(value)) { FWIW, I like to call arguments like this "predicate". One downside to criteria is that it presupposes a plural. One upside to predicate is that it matches the STL.
(In reply to comment #7) > Comment on attachment 290358 [details] > the patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=290358&action=review > > Looks good to me but the EWS result indicates a bug. > > > Source/JavaScriptCore/b3/B3MoveConstants.cpp:146 > > + if (criteria(value)) { > > FWIW, I like to call arguments like this "predicate". > > One downside to criteria is that it presupposes a plural. > > One upside to predicate is that it matches the STL. I changed it to predicate.
Created attachment 290373 [details] the patch
Comment on attachment 290373 [details] the patch View in context: https://bugs.webkit.org/attachment.cgi?id=290373&action=review > Source/JavaScriptCore/ChangeLog:20 > + This change makes moveConstants() pick the most dominant constant that works for an value. In s/an/a > Source/JavaScriptCore/b3/B3MoveConstants.cpp:193 > + // @q + c + @p - @q nifty
I'll wait for the bots, since they found things last time.
landed in https://trac.webkit.org/changeset/206682