(In reply to comment #18)
> Okay, after a quite a few tests I'd say a value of 2000 makes a pretty safe
> worksforme value.
>
> Do you have any thoughts on what kind of things in X configuration, for
> example, might affect this? I'm curious to understand why I'm (so heavily (as
> in 0.03 vs. 0.2 seconds)) affected by the problem and you're not.
Nope, no idea. Do you have a blazing-fast computer with lots of RAM? It occurs to me that 0.03 seconds isn't much time to release a key. My assumption is that the time to load xfdesktop and get to the point where it's doing the grab test dominates, and the 0.03 seconds is just enough for most people. But if you have a really fast computer, the xfdesktop load is really quick? I dunno - just pulling stuff out of my ass here, really.
(Actually, the original value was 100; I upped it to 300 before I asked you to test; so 0.01 seconds was the original timeout.)
Olivier, any ideas? I can change the value to 2000; I don't think 0.2 seconds in a busy-loop is too terrible, but it would be nice to know *why* this is necessary for some people but not others.
(In reply to comment #18)
> Okay, after a quite a few tests I'd say a value of 2000 makes a pretty safe
> worksforme value.
>
> Do you have any thoughts on what kind of things in X configuration, for
> example, might affect this? I'm curious to understand why I'm (so heavily (as
> in 0.03 vs. 0.2 seconds)) affected by the problem and you're not.
Nope, no idea. Do you have a blazing-fast computer with lots of RAM? It occurs to me that 0.03 seconds isn't much time to release a key. My assumption is that the time to load xfdesktop and get to the point where it's doing the grab test dominates, and the 0.03 seconds is just enough for most people. But if you have a really fast computer, the xfdesktop load is really quick? I dunno - just pulling stuff out of my ass here, really.
(Actually, the original value was 100; I upped it to 300 before I asked you to test; so 0.01 seconds was the original timeout.)
Olivier, any ideas? I can change the value to 2000; I don't think 0.2 seconds in a busy-loop is too terrible, but it would be nice to know *why* this is necessary for some people but not others.