Hi,

I've attached a small patch for this.

Thanks, Laurence

On Wed, Nov 20, 2024 at 2:46 PM Robert Pluim <rpluim@gmail.com> wrote:
>>>>> On Sun, 27 Oct 2024 11:50:32 -0400 (EDT), Michael Albinus via Mailing list for Emacs changes <emacs-diffs@gnu.org> said:

    Michael> branch: emacs-30
    Michael> commit 7a8ca202c5eeb810e5f86510c3ea46d3ec519222
    Michael> Author: Laurence Warne <laurencewarne@gmail.com>
    Michael> Commit: Michael Albinus <michael.albinus@gmx.de>

    Michael>     Fix flakey proced refine tests (Bug#73441)

    Michael>     * test/lisp/proced-tests.el (proced-refine-test)
    Michael>     (proced-refine-with-update-test): Use the much simpler CPU refinement
    Michael>     for testing 'proced-refine'.  The previous tests made the incorrect
    Michael>     assumption that refining on the PID of process A only filtered the
    Michael>     buffer to contain process A and its children, whereas in actuality
    Michael>     the children of process A's children, their children, and so on will
    Michael>     also be shown.
    Michael>     (proced-update-preserves-pid-at-point-test): Mark as unstable.

This causes proced-refine-test and proced-refine-with-update test
failures on macOS, where %CPU is not implemented [1]. Probably we should
stick a check for darwin around them in emacs-30.

Robert

Footnotes:
[1]  Probably because it appears to be non-trivial.

--