bug#33870: 27.0.50; xref-goto-xref not configurable

[martin rudalics]

>> We have 'display-buffer-alist' for quite some time now.
>> So please consider making this an action alist entry.
>
> Yes, it would be much better, but it never seems to reach the top of my
> todo list.
>
>> That way a user can decide whether all buffers displayed by
>> 'display-buffer' should be dedicated or only certain ones and which
>> 'dedicated' value they should get.
>
> Historically, special-display-buffer-alist always caused
> the created frames/windows to be dedicated, so
> display-buffer-mark-dedicated extends this to those windows created for
> other reasons.
>
> I haven't looked in detail, but this seems to make it less trivial to
> just add a new action alist parameter: it should default to `t` if we
> matched in display-buffer-alist but to nil if we only rely on
> display-buffer-base-action?

I'm missing you here.  An ALIST argument is equally passed to all
buffer display actions regardless of whether they are specifed by
'display-buffer-base-action' or by someone else.  It's their choice
whether they want to obey or disregard it.  The same currently holds
for 'display-buffer-mark-dedicated'.

> Also, some (all?) let-bindings of display-buffer-mark-dedicated should

I don't see any such bindings in our current code base.

> now be unnecessary (because of the features you added so bury-buffer (or
> was it quit-window?) automatically deletes the window).

This use case of dedicated windows should be no more necessary indeed.

I attach a patch of my proposed changes.  After applying that I have
no more objections against renaming 'window--display-buffer' any way
people want.

martin

window--display-buffer.diff
Description: Text document