<rdar://problem/51099753>
Created attachment 371438 [details] Patch
Comment on attachment 371438 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=371438&action=review r=me > Source/JavaScriptCore/dfg/DFGAbstractInterpreterInlines.h:3331 > + setTypeForNode(node, SpecObjectOther); can you add a test for this when we don't constant fold? I think if you run with validateAIState option it should fail in the FTL if you have a test where you create a bunch of global objects.
Comment on attachment 371438 [details] Patch Attachment 371438 [details] did not pass win-ews (win): Output: https://webkit-queues.webkit.org/results/12389465 New failing tests: imported/blink/fast/canvas/canvas-clip-stack-persistence.html fast/shadow-dom/svg-text-path-href-change-in-shadow-tree.html fast/shadow-dom/svg-use-href-change-in-shadow-tree.html
Created attachment 371463 [details] Archive of layout-test-results from ews211 for win-future The attached test failures were seen while running run-webkit-tests on the win-ews. Bot: ews211 Port: win-future Platform: CYGWIN_NT-10.0-17763-3.0.5-338.x86_64-x86_64-64bit
Comment on attachment 371438 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=371438&action=review >> Source/JavaScriptCore/dfg/DFGAbstractInterpreterInlines.h:3331 >> + setTypeForNode(node, SpecObjectOther); > > can you add a test for this when we don't constant fold? I think if you run with validateAIState option it should fail in the FTL if you have a test where you create a bunch of global objects. I don't think that's possible for globals. Since they always have different structures and we don't emit GetGetterByOffset unless that get_by_id is monomorphic. It's possible to do that with regular objects, but that's already covered by tl-try-catch-patchpoint-with-volatile-registers.js (my original patch broke it...)
Comment on attachment 371438 [details] Patch Clearing flags on attachment: 371438 Committed r246210: <https://trac.webkit.org/changeset/246210>
All reviewed patches have been landed. Closing bug.