On 2025-04-24  15:09, Eli Zaretskii wrote:

> LGTM, and the tests succeeded on my system, so I've now installed this
> on master, and I'm closing the bug.

Thanks.  Two follow-up nits, not sure whether and how you'd like to
handle them:

- Your windows-nt-specific follow-up commit e38401e7 isn't at all
  windows-nt-specific, as I learned after I noticed it.  On Debian
  GNU/Linux I have both executables "7z" and "7za" available with
  7z(1) saying:

    7z uses plugins to handle archives, so it might be slightly slower than
    7za(1) but can handle more archive formats.

  So using executable "7z" for all ports seems to be TRT.

  (And yes, I have tested using

    (define-arc-mode-test-on-type "7z" '("7z" "a") "7z" '7z)

  on GNU/Linux successfully.)

- And that part of my doc string of macro `define-arc-mode-test-on-type'
  is incorrect:

    "Define and execute a test that tests function `archive-find-type'.
            ^^ wrong ^^

  Since the macro, strictly speaking, only *defines* a test, IIUC.
  Execution happens when command `ert' (or one of its derivatives)
  runs.

Attached is a patch that fixes both of these.  Please let me know what
you think.  Thanks again.