GNU bug report logs - #44002
27.1; Small window height with VcXsrv X server when built with gtk3 toolkit

Previous Next

Full log

View this message in rfc822 format

From: martin rudalics <rudalics <at> gmx.at>
To: Eli Zaretskii <eliz <at> gnu.org>, "J. Scott Berg" <jsberg-bnl <at> outlook.com>
Cc: 44002 <at> debbugs.gnu.org
Subject: bug#44002: 27.1; Small window height with VcXsrv X server when built with gtk3 toolkit
Date: Sat, 17 Oct 2020 11:32:39 +0200

>> I did a bit more debugging. What seems to be happening is that we
>> receive a pair of ConfigureNotify events in rapid succession. The
>> first is for the top level window, and has a reasonable width and
>> height. The second is for the window associated with the frame, but
>> it has a width=1 height=1. This then leads to the resizing of the
>> frame. Presumably the frame's window is not yet mapped, and thus the
>> bogus width and height. Checking for visibility on the frame fixes
>> the problem. I've attached a patch that works for me, but I haven't
>> exhaustively tested all the various possible cases.
>
>
> Martin, any comments on the proposed patch?  Or on the problem in
> general?

I applied it here and didn't see any detrimental effects.  I suppose we
should just install it and look whether it causes problems with any of
our window managers.

For me the most interesting aspect is that it apparently works with
GTK2.  Maybe comparing the traces of GTK2 and GTK3 runs could reveal
more.  Or comparing a GTK3 trace with that of a Lucid build.

> Btw, do we understand why this happens only with VcXsrv?

I never used VcXsrv so I don't know how it works.  As an example, I have
no idea what the "top level window" is and why it is not the same as the
"window associated with the frame".  Also, the fact that a "bogus
height" is propagated while a "bogus width" is not, hints at yet another
unresolved mystery.

martin

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.