Numlock

[Mike A. Harris]

On Tue, 16 Sep 2003, Thomas Dodd wrote:

>> So, without modifying the kernel and/or X server, you'll have to 
>> run setleds, or xset for X.  I dunno if the kernel guys are open 
>> to the idea of patching the kernel to be sane for numlock by 
>> default or not, but I'd consider doing it for X, especially if 
>
>I've seen several application misbehave when num-lock is on.
>A big one form me is HP's SmarTest. Its's motif based. It's the UI for a 
>some test equipment, that I use for testing SRAM (single, dual, and quad 
>ports, FIFOs , and even some CPLDs). I often display it remotely to my 
>Linux and Solaris machines. It won't work correctly if numlock is on, on 
>either machine.

That's not an argument against having numlock enabled by default, 
but rather an argument to file bug reports to HP's engineering 
team for the faulty product.


>Sure, it's a bug in the application, but not one I can fix. HP is very 
>slow to fix bugs, and the fixes usually require the purchase of a new 
>version.

Are you suggesting that Linux development and user friendliness 
be held up waiting for commercial vendors to update broken 
software?

>I've seen similar behavior in other apps on Solaris. So there are valid 
>reason not to have numlock on. I've seen a few apps on Solaris that use 
>the numpad keys as numbers when numlock is off too. As I recal, they 
>didn't behave well when numlock was on.

Wrong.  You set defaults to be for the general case.  People 
using the faulty software can apply workarounds/reconfigure 
around the bugs in their software.


-- 
Mike A. Harris     ftp://people.redhat.com/mharris
OS Systems Engineer - XFree86 maintainer - Red Hat