Re: Question about test failure on Hydra

[Stephen Berman]

On Tue, 01 Aug 2017 13:39:50 +0900 Tino Calancha <address@hidden> wrote:

> Stephen Berman <address@hidden> writes:
>
>> On Mon, 31 Jul 2017 06:26:55 +0300 Eli Zaretskii <address@hidden> wrote:
>>
>> (dired test-dir) reverts an existing Dired buffer, because of setting
>> dired-auto-revert-buffer to t.  (When I step through the code, it is
>> after reverting that point moves to the subdirectory line, which the
>> should make "sanity check" true (and does, both when I step through the
>> code and just run the test in any way).  When I comment out the
>> dired-auto-revert-buffer line, then point stays at point-min, which
>> makes the sanity check fail (dired-file-name-at-point returns nil),
> When i comment out i see the point in point-max, and the test fails
> same as you but: (dired-file-name-at-point returns nil)

I'm seeing point at point-max now, too, but I did see it at point-min
yesterday and also when I first tried again after reading your mail,
though, strangely, I can't reproduce that now.  I was testing with two
frames and switched back and forth between them, maybe that affects
point.

>> though not in the way Hydra reports (it shows point being on the ".."
>> entry).)
> If somehow, we have in such Dired buffer the point at ".." _before_
> call `dired-revert', then _after_ revert the point is preserved:

Yes, but the question is, how does point get there in Hydra?

> We can insert additional `should' calls in the failing tests and
> wait until next hydra failoure.
>
> (Following just add more should forms into `dired-test-bug27243-01'; we
> might do the same in `dired-test-bug27243-02' and `dired-test-bug27243-03').

That's probably a good idea; can you add them?  (AFAIK
dired-test-bug27243-03 hasn't failed in Hydra, though I can't access the
logs right now to check more recent builds, but I guess it's fine to add
more sanity checks.  As for replacing switch-to-buffer by
pop-to-buffer-same-window, I used the former in an attempt to mimic
typing `C-x b', which the recipes in the bug reports use, but since it's
not being called interactively, it's probably not a valid attempt
anyway.)

Steve Berman