FW: problem: occasional keystroke doubling (a known issue?)

["Fries, David D" <david dot fries at boeing dot com>]

The correct solution if you asked me is to ignore the windows repeat
events and use this X "feature" that is causing the problem.  I
would say this is the correct thing to do because of the very reason
that the feature was included in X in the first place (at least I
assume it is the reason).  It gives you rates and delays the hardware
doesn't provide, especially lower delays than 250ms and 30cps on
pc hardware.  If you use, `xset r rate 150 80` you can see it work
under cygwin.

I'm new to windows programming so I have no idea if you can tell
windows you only want key down and up events, or if you can even
detect and throw away the repeats and let X take care of them.  I
would suggest that is the way to go if possible.

One note though if you fiddle with the rates even on a Linux computer
that is heavily loaded it can duplicate keys when you didn't actually
hold it long enough for the delay to kick in, so it isn't always
foolproof even on Linux.

I've noticed that the modifiers under cygwin (control, alt, shift etc)
show that they repeat under xev just like other keys where they don't
on Linux.

While input is the topic is it possible to claim the alt and window keys
in cygwin?  I have fvwm setup to use them a ton and it works great under
Linux, but not at all in cygwin.  I was debating if it would be useful to
let windows still control alt-tab or not.  I've seen other programs
take control of the window and alt keys, so it is possible, but I have no
idea what it would take.

Harold L Hunt II wrote:
> Excellent.
> 
> Now, if I could only find a way to turn this off programmatically, once 
> and for all.  This X key repeat feature is completely unnecessary since 
> Windows handles key repeats for us.  I tried to interface with the same 
> code that xset interfaces with, but it did not seem that this interface 
> was meant to be built into an X Server.  Perhaps someone else can help 
> me figure out how to disable this.  It could be a simple as a build
> flag, but I have never come across it.
> 
> Harold

--
David Fries <dfries@mail.win.org> (for personal)