RFR: 8374341: Use of "\s" instead of "\\s" in CoreUtils.java results in splitting on backspace instead of whitespace [v4]

[Chris Plummer]

On Mon, 12 Jan 2026 12:06:04 GMT, Kevin Walls <kevinw at openjdk.org> wrote:

> It looks wrong. This match can never have worked?

Yes, I agree with that, but as I pointed out, this check was added late in the review, so the code after the check was likely getting exercised and worked while the review was in process, but probably not tested well enough at the end of the review.

>  And an option is to remove the code if not tested or known good...

I think the goal of the code is to give the core processing time to complete. I think for this use case, the core still ends up in the same location, but you just have to wait for it. Is that correct? If so, it would be useful for the work I'd like to see done to get the core file handling to work when the core file is zipped. Right now that results in SkippedException because the core file is not found. We could look for a .gz file, but I'm guessing we have to wait for it to appear, which is where this existing code comes in handy. If I'm not understanding this correctly, please let me know.

> So I see that recognising the pipe symbol is really benficial, as it prompts me to fix the system. But being so specific about matching systemd-coredump I'm not sure about the benefit.

I'm still fuzzy on what systemd-coredump is expected to do in the context of this CoreUtils code.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/28984#issuecomment-3747812759