Nathaniel Smith wrote:
>
> I found the bug in GGI::Visual that was preventing events from being
> forwarded; had to fill in the event.any.target field, or it was discarded
> by ggiEventSend.
strange. Why is this needed ? Events are sent to visuals. So what is the target
good for ? The SendEvent already takes the visual as argument. It seems redundant
somehow...
> Unfortunately, this still isn't enough to make everything
> work perfectly; applications that poll for any key (say, the dots demo)
> work correctly. Applications that pause and wait (eg, the standard
> demo) do not -- there seems to be a failure in giiEventPoll for the SHM
> target.
keep in mind that our synthesized ggi events aren't correctly initialized.
Half of the information is missing. We have to look into the gii docs and
reconstruct the missing part, or else rethink our key event structure, if
it turns out to be essential.
> X just crashes when I press a key -- I'll have to look more
> closely; probably we have some fields set up that it doesn't like, or
> don't contain enough information. In any case, though, we're getting
> much closer to getting X etc. (I was going to try quake, but realized I
> don't have the data files around anymore) running in a Berlin window.
well, and we don't yet generate positional events. I'm working at the
PositionalFocus / PickTraversal again, so I'll look into that part shortly.
Regards, Stefan
_______________________________________________________
Stefan Seefeld
Departement de Physique
Universite de Montreal
email: [EMAIL PROTECTED]
_______________________________________________________
...ich hab' noch einen Koffer in Berlin...
