[gpm]Re: two bugs in gpm

Alessandro Rubini rubini@gnu.org
Mon, 1 Oct 2001 16:06:30 +0200


>> One is the way the baud rate setting is done. It worked fine in old
>> versions of gpm, but the special case for mman in the baud rate
>> setting code now actually *breaks* support for my Logitech MoiseMan
>> (the version with 4 buttons). Could you provide an way to initialize
>> it to 9600 baud just like all the other mice, without using the
>> special mman code that doesn't work for this mouse?
> 
> For now I have to pass this question to Alessandro.

Well, the mman stuff was an hack. How does it break your mouse? (no,
I'm in a hurry now (and mostly always) and I can't check the code).
Nico, please take a look at it and report any question to me.
 
>> The second bug, which probably affects more people, is that gpm fails
>> to detect increases in the console size

I've had the same problem lately, after switching to a mac as my
personal workstation. I was pretty sure I fixed that, and wondered why
it didn't work in my box. I though it was the older Debian gpm I was
using.

>> The way I normally fix this problem is just to make gpm
>> re-check the console size on every mouse motion event, but there
>> should be a better way...

Since it checks what the current console is at any event, I think the
best thing to do is making it check the size only when the current
console changes. Thus, after a setfont or other size-changing command
you just need to switch to another console and then back. It's a
minimal issue for a user, and anyone will see how it works even
without reading the manual (I must do the double change when I press
ctrl-alt-f1 from X for example, otherwise the text console thinks the
control key is still pressed).

When you change console size you might sent winch to gpm, but since
each console can be of a different size, you still need to resync
everytime the console is switched.

Hope this helps
/alessnadro