Re: emacs-29 b44a7ff85dc: Allow 'icon-title-format' to have the value t

[Óscar Fuentes]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Óscar Fuentes <ofv@wanadoo.es>
>> Date: Tue, 21 Feb 2023 16:02:22 +0100
>> 
>> What about the other part on my message?
>
> You mean, this part:
>
>> A reason for keeping icon-title-format in sync with frame-title-format
>> is for not having to also set icon-title-format whenever
>> frame-title-format is set (because most users will want to retain the
>> title of a frame when it is iconified.)
>
> ?

Yes.

> If that's what users want, how did they do it before this change?

They didn't need to do anything, because icon-frame-title was
ineffective and Emacs used frame-title-format when minimized too.

>> The current explanation for having `t' in icon-frame-title should be
>> removed.
>
> It was the motivation for the change, AFAIU, and I'm not at all sure
> KDE is the only environment where the problem happens.

The problem with KDE fudging with the focus when switching virtual
desktops is a bug that manifests itself under very specific conditions
that involve mechanisms exclusive of KDE. That bug was reported to KDE
and Emacs does not need to care further about it.

As for the motivation for the change, I don't know. I proposed in #61496
that the *default* value for icon-frame-title should be `nil' (or `t',
that detail was irrelevant) for meaning "same as frame-title-format" so
the observable behavior of Emacs wrt previous releases doesn't change,
so the users have no need to tweak their configs. You rejected the
proposal and implemented support for the "same as frame-title-format"
(which is good to have) but didn't changed the default.

> If the desktop doesn't behave like KDE, users could simply set both
> formats to two identical strings.  The only reason to force Emacs to
> use the _same_ string (which can only be done internally) is that
> those desktops misbehave.

The reason to force Emacs to use the same string is to behave as most
users expect, and to avoid hard-to-detect breakage on desktop automation
setups that depend on the title of the window.

That was explained in #61496. No need to rehash the discussion here.

> So I don't see why we should remove the explanation.
>
>> OTOH, maybe we could explain that icon-frame-title it is effective
>> in cases where the user does not explicitly iconify the frame, as
>> when he moves to another virtual desktop.
>
> Sorry, I don't think I understand what you mean here.

Again, it was explained in #61496.