...
The consequence here is that we end up monotonically growing that cursor, which will eventually lead to us never stealing empty blocks.
Created attachment 336495 [details] patch
Is that a known flaky test? ** The following JSC stress test failures have been introduced: stress/ftl-put-by-id-setter-exception-interesting-live-state.js.no-cjit-collect-continuously
Comment on attachment 336495 [details] patch Clearing flags on attachment: 336495 Committed r229962: <https://trac.webkit.org/changeset/229962>
All reviewed patches have been landed. Closing bug.
<rdar://problem/38843812>
(In reply to Alexey Proskuryakov from comment #3) > Is that a known flaky test? > > > ** The following JSC stress test failures have been introduced: > stress/ftl-put-by-id-setter-exception-interesting-live-state.js.no-cjit- > collect-continuously Yes. I've seen it flake before. Do you know how to grab the crashlog from the bots? I'll file a bug or you can and assign it to me.
There is no crash log on the bot, looks like the test aborts the process without crashing. Running stress/ftl-put-by-id-slow-exception-no-catch.js.no-cjit-collect-continuously stress/ftl-put-by-id-setter-exception-interesting-live-state.js.no-cjit-collect-continuously: Exception: Error42 Running stress/ftl-put-by-id-slow-exception-no-catch.js.dfg-eager Running stress/ftl-put-by-id-slow-exception-no-catch.js.dfg-eager-no-cjit-validate stress/ftl-put-by-id-setter-exception-interesting-live-state.js.no-cjit-collect-continuously: ERROR: Unexpected exit code: 3 FAIL: stress/ftl-put-by-id-setter-exception-interesting-live-state.js.no-cjit-collect-continuously
Found an existing bug for that, bug 183266.