Patch Set 2:
I'm still uncertain as to how https://github.com/dart-lang/sdk/issues/36232 is still causing breakages after the commit in question being reverted. Do we know why we're seeing these failures?
Unsure. Is it possible that they already were approved errors, but when your CL came through and fixed the tests, that the approvals for failures went away, and then when the revert happened the tests now started visibly failing due to the lack of approvals?
You can see the diff, as it's very small, and it shows why the failures were happening. The lifecycle test looks for a suffix (using `endsWith`) of "main" to filter out non-spawned isolates, but the actual string ends in "main()" with parens. When set_rpc_name_test is compiled, it's compiled to <...>/out.dill and so the isolate is named "out.dill:main()", but that wasn't a possibility considered by the test.
The latter could have happened if the test infrastructure changed to compile to out.dill instead of, say, out.jitsnapshot, but I don't think that was a recent change. In the former case, it could be that the string representations of methods changed recently to add the parens, but since both original examples in set_name_rpc_test end in "main()", I also don't expect that to have been a recent change, hence the speculation about approvals because I don't see how these tests passed previously.--
You received this message because you are subscribed to the Google Groups "Dart Code Reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email to reviews+u...@dartlang.org.