[MPlayer-dev-eng] Release... How about it?

[Reimar Döffinger]

Hi,
> RD>Wrong. It sure was read, and I even discussed it with beastd (I don't
> RD>know if I referred specifically to this mail, but still...).
> 
> OK, sorry.. That's good, but there was no reaction in mailing lists and
> you sent this patch only after I wrote here - so my comment still
> applies.

I sure applies partially. There was no reply because I had not yet made
up my mind on how to fix it and was a bit lazy ;-)

> RD>> code, maybe someone kept the original patch and can send it again so it
> RD>> can be reversed?
> RD>
> RD>This patch did not cause the bug, actually just several bugs were hiding
> RD>behind this one...
> RD>
> RD>Please try the attached patch, I can't see any point in using a pipe for
> RD>buffering keyboard input anyway...
> 
> It doesn't help. Well, actually it does, it reduces the effect (so the
> maximum reaction latency is about 1 second now, much better than
> before), but that's far from what it was before that stuck buttons
> patch.

Well, I changed that one define. It specifies how many unprocessed keypresses are kept in the queue. You would get the behaviour from before the patch by setting it to one, but that means that if you press more than one key during the display time of one frame, it will simply be discarded. Maybe setteing it to 5 would be okay to, but less seems not acceptable to me.

Greetings,
Reimar Döffinger