Re: [PATCH] lisp/ob-octave.el, was [PATCH] rfc: using ert-deftest with side-effects

[Leo Butler]

On Mon, Jan 23 2023, Ihor Radchenko <yantar92@posteo.net> wrote:

> Ihor Radchenko <yantar92@posteo.net> writes:
>
>> Ihor Radchenko <yantar92@posteo.net> writes:
>>
>>> So, the test failure is real.
>>
>> The error buffer contents when the test fails is the following:
>>
>> warning: using the gnuplot graphics toolkit is discouraged
>> ...
>> [ Babel evaluation exited with code 0 ]
>>
>> Exit code is 0, so octave does finish.
>>
>> Hence, test assertion that
>> (should-not (buffer-live-p (get-buffer "*Org-Babel Error Output*")))
>> does not appear to be accurate.
>>
>> Leo, should we simply remove the assertion?
>
> I decided to remove the assertion from the tests.
> https://git.savannah.gnu.org/cgit/emacs/org-mode.git/commit/?id=59228e51345ab522d9db611c8e73caa078d86d2f
>
> While there could be a value in making sure that ob-octave calls octave
> without generating errors, it is not strictly what the test is checking
> for.
>
> Moreover, we currently have no reliable way to disambiguate mere
> warnings from non-zero exit code.

Yes, as I said in a previous email, I can live with that.

The origin of the errors that you documented is not, I believe, due to
either of the causes suggested by either Max or you. But, investigating
obscure test failures like this is probably not the best expenditure of
time and effort.

Best regards,
Leo