bug#52380: ERT, buffers and windows

[Kévin Le Gouguec]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Kévin Le Gouguec <kevin.legouguec@gmail.com>
>> Date: Fri, 10 Dec 2021 21:09:30 +0100
>> Cc: 52380@debbugs.gnu.org
>> 
>> If I amend the test as follows:
>> 
>> diff --git a/test/lisp/progmodes/python-tests.el 
>> b/test/lisp/progmodes/python-tests.el
>> index 15bda5c197..783f115fc7 100644
>> --- a/test/lisp/progmodes/python-tests.el
>> +++ b/test/lisp/progmodes/python-tests.el
>> @@ -5454,6 +5454,7 @@ python-tests--bug31398
>>    "Test for https://debbugs.gnu.org/31398 ."
>>    (skip-unless (executable-find python-tests-shell-interpreter))
>>    (let ((buffer (process-buffer (run-python nil nil 'show))))
>> +    (should (eq (get-buffer-window buffer) (selected-window)))
>>      (should (eq buffer (current-buffer)))
>>      (pop-to-buffer (other-buffer))
>>      (run-python nil nil 'show)
>> 
>> … then (should (eq (get-buffer-window buffer) (selected-window)))
>> passes, but (should (eq buffer (current-buffer))) does not.  Does that
>> make sense to anyone?
>
> You assume that the selected window always shows the current buffer?
> That's not in general true.

Mmm, thanks.  I see that "(elisp) Current Buffer" has things to say on
the subject:

>    When an editing command returns to the editor command loop, Emacs
> automatically calls ‘set-buffer’ on the buffer shown in the selected
> window (*note Selecting Windows::).

In the context of ERT, what would be the correct way to test that after
calling M-x run-python (≡ (run-python nil nil 'show)), the *Python*
buffer is visible, and its window is selected?

Should we just disregard (current-buffer) when writing ERT tests, and
assert things in terms of (selected-window)?  Or is there a way to make
ERT "return to the editor command loop", thereby selecting the buffer
shown in the selected window?