bug#12906: 24.2.50; Parts of the screen are completely white

["Fabrice Niessen" <fni-djc/iPCCuDYQheJpep6IedvLeJWuRmrY@public.gmane.org>]

Eli,

> Eli Zaretskii wrote:
>>> From: "Fabrice Niessen" <fni-djc/iPCCuDYQheJpep6IedvLeJWuRmrY@public.gmane.org>
>>>
>>> Because of heavy workload (with the end of year holidays and so on), I cannot
>>> commit on being able to free up the time needed to do so, if this is
>>> configuration related.
>>
>> Not really configuration related.  Rather, something in your
>> configuration triggers the bug.  And it's undoubtedly a bug, since
>> redraw-display fixes the display: some redisplay optimization
>> somewhere decides it's okay to take a shortcut, which it shouldn't.
>>
>>> However, the first settings I'd suspect would be in the "Controlling the
>>> display" block. I'm copying it in case it'd trigger some alarm for someone:
>>
>> Thanks, I will try with some of these settings.
>
> Another thought: I've observed that in Org buffers -- but I almost exclusively
> work in such buffers, nowadays -- with "literate programming" aspects (I mean:
> with code blocks inserted in the flow of the text).
>
> And, something that's not on by default about those blocks is:
>
>   ;; fontify code in code blocks
>   (setq org-src-fontify-natively t)
>
> Please add that as well in your test config, it could count!
>
> Thinking more about it, I'm not sure (anymore) I've ever seen the display
> problem in pure Emacs Lisp buffers, for example -- though, this may prove
> wrong (assumption to be taken with lots of caution).

After one day without my settings about "Controlling the display" (given
previously), I did not see the problem (yet?). Hence, a furious envy to say
that the "trigger" settings provoking the problem would be in that given
configuration. Just for you to know.

Take care that this conclusion is only based on time and observation, though.
But I'm quite confident about that first observation.

FYI, I did not undo the "fontify code" setting in my config for the moment.
Hence, that Org setting should not be the problem. Only one (or several) of
the other ones.

Best regards,
Fabrice