emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Emacs's set-frame-size can not work well with gnome-shell?


From: martin rudalics
Subject: Re: Emacs's set-frame-size can not work well with gnome-shell?
Date: Sun, 19 Jan 2020 09:52:55 +0100

> Maybe. But it might also mean that it's possible to provide some hint
> to the window manager that resizing is indeed allowed.

Maybe.  But how would we find out?  Did others try your "fix" already?

> Do you have a compositing window manager (on your GNU/Linux system)?

Xfce comes with compositing capabilities but my graphic card might not
comply.  Setting an 'alpha' parameter for a child frame effectively does
not work (it seemed to work on my Windows machine but that is broken now
so I can't check).  Mostly I don't care because normal Emacs frames
always flicker during resizing anyway.

> It doesn't flicker always, but like 3 out of 4 times is does. And running 
resize-test in a loop always shows 1-2 flickers, at least.

So when you now resize a child frame by dragging its corners, it usually
flickers.  It does not flicker at all here.  Probably because its buffer
is empty, possibly because certain compositing effects are ignored or
not implemented.

>> You could try to play around with 'x-wait-for-event-timeout'.
>
> You mean to bind it to a larger value around the body of resize-test?
>
> I tried 1, as well as 10. No change.

A floating point value less than 0.1.  Non-floating point values stand
for 0 ("no wait").  It stands for waiting that many seconds for a
configure notify event that would confirm that your frame has really
become visible or invisible.  So IIUC (Noam might correct me) you may
have to wait twice that time with WMs that don't send such events.


BTW: I just found out that with emacs -Q my latest code samples produce
a red border when I do 'eval-buffer' via M-x but don't produce that red
border when I put an

; (eval-buffer)

line into *scratch* and hit C-x C-e with point after the ")".  So much
about the reproducibility of experiments here.

martin



reply via email to

[Prev in Thread] Current Thread [Next in Thread]