chelcassanova wrote: @labath
> Well.. IIUC, this test tries to attach to a process called "a.out" -- and > hopes that the process doesn't appear. That's pretty much it, I just need to wait-attach to _some_ process to trigger the progress report I'm expecting and then drop the process once we've triggered that report and check that it was emitted correctly. > If you run the test suite in parallel, you'll have any number of processes > like that going around all the time. Not only will this test fail, but it > will mess up the state of a random other test. Great point, I hadn't considered this. I'm also a little surprised this seems to specifically affect ARM Ubuntu here 🤔 . > I guess you could just wait for a sufficiently unique process name... This is probably the right way to go for this, I'll update the test to do this. @DavidSpickett > Though I cannot reproduce the failure, so it could be to do with the high > load when run on the bots. Some process being deprioritised. With this and Pavel's comment about the generic process name it's probably how this test works on the bots. I'm not sure if there's a way to preview how this test would run on the bots before actually landing (outside of running it locally or having the GH pre-commit checks), but I'll try relanding and changing the process name to wait on for now since I don't think I have another way to ensure the test would be fine on the bots. https://github.com/llvm/llvm-project/pull/144768 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
