LeeTibbert

j-mie6 Thank you for this discovering this defect and creating this Issue. I believe that it captures the Discord discussion and will be useful to future maintainers. A clarification, if...

> exitValue: calls with WNOHANG, but this means that there could be an exception thrown here as well. It's unclear how this should proceed The gold standard is: What does...

> askZombiesForTheirExitStatus: this does not WNOHANG, so should be set to false (ECHILD here is a legitimate bug) I concur that ECHILD here is totally unexpected and should throw. The...

How expensive is the execution run which you mentioned in the base note? Can it easily be done using 0.5.7-SNAPSHOT code (aside: is there a Highway 61 in Britain ?)...

> definition of waitpidImpl to swallow ECHILD, since this might affect people doing lower-level C-style operations. Discourse is a terse medium, helpful for triage & iteration but can leave lingering...

Good discussion, makes for a stronger, more maintainable project. In situations like this I take two approaches: 1) return to primary sources. In this case, linux & macOS documentation. 2)...

@j-mie6 Are you seeing the problem in the base Issue on linux, macOS, or something else? I started my journey with the Wikipedia topic: [Zombie process](https://en.wikipedia.org/wiki/Zombie_process). Not authoritative but authoritative-adjacent...

Perhaps but a comment in the code referring to this Issue and the reasoning this Issue gives for the third argument value used at each individual waitpidImpl() call site?

re: bug shows on Linux Thank you for the info. Based on that I studied the SN `Process` `posix_spawn()` call site. Nothing suspicious in either the Open Grouj (2024) description...

* I concur that `isAlive()` and `exitValue()` should suppress `ECHILD`. * I believe that `askZombiesForTheirExitStatus()` should throw `ECHILD` but would not object to it suppressing it. Both branches of the...