Bug 10308

I hope I am using the correct terminology.

The problem happens occasionally (this latest time I was using SeaMonkey Mozilla Email), 5 times I can remember since I installed FC4 14 months ago.  The keyboard stops working a normal typing speed, I have to hold a key down for 1 second (estimated) before any application registers the keypress.  The auto-repeat timer appears unaffected by this bug.  I would guess my default X11 auto-repeat is in the order of 250ms @ 24 per second.

This means that while the bug occurs:

If I hold a key down for 150ms it does not register in the app.
If I hold a key down for 1001ms it registers as a single key press.
If I hold a key down for 1221ms it registers as a key press and 1 addtional press due to auto-repeat.
If I hold the key down for 1501ms it registers as a key press and 6 additional keypressed due to auto-repeat (on the basis I have it set to 24 per second, guess)

The above information is more estimate than fact but its the best way I can describe to someone else that happens.


Other interesting facts I can say (whilst the bug is in effect) :

 * In order to use multiple keys, like Shift or Ctrl+Alt+<key> I have to hold down each modifier key for 1 second first, before attempting to press the next key.  So even modifiers like shift will not register in the emitted char if I didn't hold the shift down on its own down for 1 seconds first.  When I am using 3 keys, I hold down Ctrl, then wait 1 second, then hold down Ctrl and Alt (without letting go of the Ctrl), then wait 1 second, then hold down Ctrl and Alt and F1 and wait for 1 second, then it will switch out to a text mode VC.

 * If I use Ctrl + Alt + F1 (to switch back to a text mode VC on linux), then the text mode VC has perfect keyboard operation, it is unaffected by this problem.

 * Swithing back to X11 (from the text mode VC) and the problem persists.

 * Mouse operations are unaffected when the bug occurs.

 * The only way I know to fix the problem is to logout of the X11 session, I believe Fedora Core 4 shuts down the X11 server and respawns a new one.


Maybe this is an SMP problem, in the area of modifying the internal keyboard timing parameters withing the server during keyboard grabs ?

From memory at the time I had operated a popup menu within Mozilla between the time the keyboard last worked normally and when the bug occured.

Is there a command line utility I can use to dump out the complete keyboard configuration, a bit like "xset" butI am not aware "xset" can be used to inspect the current configuration, it can only be used to set a new configuration.  I also don't believe "xset" allow configuration of the debounce timeout parameter.


My system is based on FC4, its a Dual CPU x86_64 clocked at 3.2 GHz.  The X server came from the package xorg-x11-6.8.2-37.FC4.49.2.

If there is any other information I can provide please ask.  I am not able to provide a procedure to follow to cause the problem, as I say its only happened a few times over the year or so I have been using this version and am hoping that a developer might know where to look from the information given above.