RE: Multiwindows problem

[Jean-Claude Gervais <jc dot gervais at videotron dot ca>]

Hi Colin,

Just wondering; have you read the threads about this problem in the
mailing-list archives?
If not, you should definitely take a look: Harold, Myself and others have
done a fair amount of investigation as to what could be causing this, and it
may serve you well, if only to help you exclude hypothesis.


-----Original Message-----
From: cygwin-xfree-owner@cygwin.com [mailto:cygwin-xfree-owner@cygwin.com]On
Behalf Of Colin Harrison
Sent: Wednesday, May 21, 2003 4:58 AM
To: cygwin-xfree@cygwin.com
Subject: RE: Multiwindows problem

Hi Earle,

I must admit I'd prefer to find the cause and not fix the symptoms!

The winwndproc.c code contains WM message handler code that may have been
missed from the
multiwindow case. For example, mode-key state retention is handled between
WM_KILLFOCUS and WM_SETFOCUS. The WM_KEYDOWN stuff looks fairly well
'evolved', with repeats on key events etc, which are maybe required in
multiwindow?

The communication with X clients is confusing me, as to where it should
occur, in X or windoze 'servers'. Ensuring 'bomb proof' synchronization of
the different manager's messages must be a problem? Maybe I need to get
back-to-basics here and get away for the empirical approach, or else I'll
only add an unnecessary layer of bodge!

I'm still tracing what we have, to try and come up with a 'safe/minimum but
quick and dirty' fix first. One mystery (to me anyway!) is why some of the
WM_KILLFOCUS and WM_ACTIVATEAPP messages have 'NULL' pWin and hwndScreen
handles...

The 'dragging X scrollbar' problem is somewhere in the buttonup messages,
which I'm about to trace (and try to fathom out!)

On the side, how can I avoid my XWin.exe being dependent on libfreetype.dll?
Harold's test builds aren't and I like developing like-with-like (some mod.
to the config/cf files no doubt, but I'm lazy!)


Colin