Bug 72151

Nothing has changed in X that would cause such behaviour so this
is highly unlikely to be an X bug.

Key repeat works fine for me on all of my machines.

You'll have to provide more information on what hardware you are using,
and what specific steps you have tried doing.

>Key repeat works fine for me on all of my machines.

It currently works! There's probably some instability, but I really don't know
why it didn't work (probably it doesn't work in some circumstances, which
existed yesterday when I tried it, but which don't exist currently; sorry!).

>You'll have to provide more information on what hardware you are using,
>and what specific steps you have tried doing.

My PS/2 keyboard is configured as US International, and I tried to change the
key repitition using Preferences|Keyboard (nothing else). I tried everything in
that program I could think of, but it wouldn't change a thing. (Now it works.)

I'm afraid I can't tell why it didn't work (there are no logical circumstances,
or changes since yesterday). Maybe the best solution would be to resolve this
bug as NOTABUG..

(Side issue: the scroll bar in that same preferences program for the delay works
exactly the other way around as is indicated.)

joe has seen this with gnome2 without our beta so it's probably a gnome bug.

[13:00:35] <joe> I've seen it a couple of times in the g2 snaps and in xd2
snaps, so it's not a red hat-only problem
[13:08:20] <joe> Well, I don't know much about the problem; it's very sporadic.
 Sometimes it happens on login, sometimes shortly after an unsuspend
[13:08:42] <joe> Doing an "xset r on" or "xset r rate ..." doesn't work
[13:08:47] <joe> Jody told me to try the following, though:
[13:08:54] <joe> 12:13 < jody> 1) kill settings deamon
[13:08:54] <joe> 12:14 < jody> 2) xset r rate 500 50
[13:08:54] <joe> 12:14 < jody> 3) xset -q
[13:08:54] <joe> 12:14 < jody> a) Does it keep the 500 50 ?
[13:08:54] <joe> 12:14 < jody> b) Do you get any repeat ?
[13:09:05] <joe> but I haven't been able to try it out because it's working for
me now, and I need to get work done. :)

I've seen rumors of X screwing up the repeat rate after an unsuspension.  Are
either of these machines laptops?

[13:52:43] <joe> blizzard: just on my laptop

My machine isn't a laptop. If that's interesting, I have a SiS 730S mainboard
(the AMD variant of 730). Please ask for specific information if you can use it.

So, I've seen the behavior a bit more.  I'm still not totally sure if suspending
is what's causing the problem.

If I am having the problem and I log out, I don't see the problem in gdm or my
ssh-askpass-gnome, but it does show up after I log in.

If I kill the X server, though, and log back in, I don't see the problem (or at
least, have not yet).  It usually happens then after I suspend, and
occassionally when I switch between the internal display and the external VGA.

"xset r" doesn't seem to ever work, nor does killing gnome-settings-daemon.

Oh, I should note that this first happened when I installed null and uninstalled
it because of this.  But it happens for me now in 7.3, since I started running
GNOME 2 on it.  Previously I ran 7.2 on this machine with the daily gnome 2
snaps (as recently as about 3 weeks ago) and never saw any problems.  I suspect
it may be a change between XFree 4.1 and 4.2.

Yes, I also found the power-saving configuration in my BIOS relates to this
problem. It was set to suspend after 15 minutes without using. Then after I 
come back. The key repeat is disabled. Later I disabled suspending my machine
in BIOS setting. Then this problem disappeared.

BTW, previously I ran Redhat 7.2 and this problem did not exist. Only after I
upgraded/reinstalled to 7.3/8.0, this problem crept into my machine.

You've indicated changing a BIOS setting makes the problem go away, so I
consider this a weirdo hardware/BIOS issue.  It is not going to be
reproduceable on any hardware that I have available to me in any case,
and so wether it is a hardware bug, BIOS bug, or an XFree86 bug in the end,
there is virtually nothing I can do about it.  In any case it most certainly
isn't a Red Hat specific problem, and since you've worked around the issue,
I'm closing this WONTFIX.

If the problem recurs or you would like someone to investigate it further
I suggest reporting it directly to xfree86 so that the upstream
developers and maintainers are aware of the problem and if someone has the
hardware on hand they can investigate.