bug#81394: 31.0.90; unexpected buffer completion behavior

Ping!  Spencer, could you please fix the test, so this change could
be installed?

> Cc: [email protected], "Charles A. Roelli" <[email protected]>,
>  [email protected]
> Date: Thu, 23 Jul 2026 10:03:33 +0300
> From: Eli Zaretskii <[email protected]>
> 
> > From: "Charles A. Roelli" <[email protected]>
> > Cc: [email protected], [email protected], [email protected]
> > Date: Wed, 22 Jul 2026 14:25:11 +0000
> > 
> > > Thanks for the detailed report and bisecting.
> > >
> > > Can you try this patch?
> > 
> > Thanks for the patch, it's working on emacs-31.
> 
> I was going to install this on the emacs-31 branch, but the updated
> test fails here:
> 
>   Test completion-pcm-test-7 backtrace:
>     signal(ert-test-failed (((should (equal (completion-pcm--merge-try '
>     ert-fail(((should (equal (completion-pcm--merge-try '(prefix any) '(
>     (if (unwind-protect (setq value-224 (apply fn-222 args-223)) (setq f
>     (let (form-description-226) (if (unwind-protect (setq value-224 (app
>     (let ((value-224 'ert-form-evaluation-aborted-225)) (let (form-descr
>     (let* ((fn-222 #'equal) (args-223 (condition-case err (list (complet
>     #f(lambda () [t] (let* ((fn-197 #'equal) (args-198 (condition-case e
>     #f(compiled-function () #<bytecode -0xa0f4154856001a2>)()
>     handler-bind-1(#f(compiled-function () #<bytecode -0xa0f4154856001a2
>     ert--run-test-internal(#s(ert--test-execution-info :test #s(ert-test
>     ert-run-test(#s(ert-test :name completion-pcm-test-7 :documentation
>     ert-run-or-rerun-test(#s(ert--stats :selector ... :tests ... :test-m
>     ert-run-tests((not (or (tag :unstable) (tag :nativecomp))) #f(compil
>     ert-run-tests-batch((not (or (tag :unstable) (tag :nativecomp))))
>     ert-run-tests-batch-and-exit((not (or (tag :unstable) (tag :nativeco
>     eval((ert-run-tests-batch-and-exit '(not (or (tag :unstable) (tag :n
>     command-line-1(("-L" ";." "-l" "ert" "--eval" "(setq treesit-extra-l
>     command-line()
>     normal-top-level()
>   Test completion-pcm-test-7 condition:
>       (ert-test-failed
>        ((should
> 	 (equal (completion-pcm--merge-try ... ... "" "") '("foo" . 0)))
> 	:form (equal ("foo" . 3) ("foo" . 0)) :value nil :explanation
> 	(cdr (different-atoms (3 "#x3" "?\3") (0 "#x0" "?\0")))))
>      FAILED  25/51  completion-pcm-test-7 (0.000546 sec) at lisp/minibuffer-tests.el:291
> 
> The original test passed with the code before the patch.
> 
> Can the test and/or the code be fixed not to fail, please?
> 
> 
> 
>