nanogui: Re:Re: a question about flnx on framebuffer
Subject:
Re:Re: a question about flnx on framebuffer
From:
Jordan Crouse ####@####.####
Date:
30 Apr 2001 18:08:38 -0000
Message-Id: <01043012091408.00902@cosmic>
I have yet to run into a situtation where GPM couldn't bail me out. I do
agree that we should probably have some more native mouse drivers, but not at
the expense of needlessly expanding the Microwindows code. Personally, I
prefer the "when its needed" approach. I like that we can still fit the
Microwindows code on a floppy disk, I would like to keep it that way for a
while.
Jordan
On Monday 30 April 2001 11:30, Alex Holden mentioned:
> On Mon, 30 Apr 2001, Jordan Crouse wrote:
> > When Microwindows was
> > developed, the decision was made to be as generic as possible (so it
> > could be run on many different devices with a minimum of effort). So we
> > used the GPM feature of "repeating" the mouse info as PS2 data in the
> > device file /dev/gpmdata. That way, we can handle any mouse as a PS2
> > mouse in a predetermined location.
>
> Actually it was more of a matter of hacking something together in an
> evening to get mouse input working. Using GPM to do that seemed a good
> idea at the time, but in retrospect I'd like to implement a few simple
> native drivers for the most common mouse types as well. That way we can
> cleanly support wheel mice and other more-than-three-button types of mice,
> allow Nano-X to play nicely on a system that is also running XFree86
> (actually, you can already do that if you tell X to use the repeater, but
> again that doesn't work for wheel mice), and remove the uglyness of
> having two mouse drivers one on top of the other, with one of them
> running as a seperate process to the windowing system. If you've ever
> tried hacking GPM, you'll know it's not the cleanest of programs either.
