[vdr] Re: [ANNOUNCE] vdr-actuator-0.0.5 plugin

[Luca Olivetti <luca@ventoso.org>]

Christian Gmeiner wrote:
> > Yes, I understand that, but it's just a workaround until the dxr3 
> > plugin is fixed , at least I think it's caused by the dxr3 plugin, 
> > because if I don't load that plugin I don't see the problem, but I 
> > don't see tv either ;-) .
>
> Could you explain me the problem with the dxr3 plugin?
Well, I don't know exactly. I just know that a short while after the 
call to cDevice::Shutdown() vdr crashes, so not all the plugins have 
enough time to execute their finalization code. I scratched my head for 
a while and I thought that my plugin was at fault, because the crash 
happened when my plugin was exiting.
After many hours of debugging and trying various things, I could not 
find anything wrong with my plugin, I only saw that, by changing the 
order of some instructions[*], the crash happened at a different points, 
so I though that it was something else causing the crash after a more or 
less fixed amount of time. I tried removing the plugins one by one and I 
saw that without the dxr3 plugin, the problem went away. So, since 
removing the dxr3 plugin is not an option, I just changed the order of 
the calls in vdr.

[*] I actually put a printf before and after a sleep, the printf before 
printer but it never reached the printf after. I thought that "sleep" is 
a function that cannot be called in a thread, so I tried removing it, 
and saw that the crash was happening later on. That's when I started 
thinking that something else was crashing and it was just a matter of 
timing that my plugin was the affected one.

Bye
--
- Yo también quiero una Europa libre de Patentes de Software  -
- I want a Software Patents Free Europe too! And you?         -
---------------------------------------------------------------
  EuropeSwPatentFree - http://EuropeSwPatentFree.hispalinux.es

Attachment: signature.asc
Description: OpenPGP digital signature