Re: [gst-devel] GstMad problems

[Thomas V. S. <th...@ap...>]

Hi,


> 1- Is there a reason why the chain function checks only one time if the first 
> frame of an mp3 has a Xing header? What I mean is that if I change the state 
> of my pipeline to GST_STATE_READY, and that I change the location of my 
> filesrc for an mp3 that has a Xing header and I reset the state of the 
> pipeline to GST_STATE_PLAYING, what would happen? Should we not set 
> mad->check_for_xing to TRUE in the gst_mad_change_state again?

Sounds like that should be done, yes.  Feel free to supply a patch in
bugzilla, preferably with a small test app (can be in python for
example) and a link to such an mp3 so we can verify your patch fixes
something.

> 2- Let's say that I have an MP3 that have a bad frame (the first one 
> actually), and let's say that mad_frame_decode returns MAD_ERROR_BADDATAPTR 
> which is not handled in the plugin. Does someone know what this error means? 
> Should we not skip that kind of frame in the plugin?

I have no idea what it means, you should check mad.  but it sounds like
it would be good to check for - I can't say if we should skip it without
knowing what it means.  Again, patch accepted :)

Thomas


Dave/Dina : future TV today ! - http://www.davedina.org/
<-*- thomas (dot) apestaart (dot) org -*->
Ten generations of selective cross-breeding
and you still have a fat ass ?
<-*- thomas (at) apestaart (dot) org -*->
URGent, best radio on the net - 24/7 ! - http://urgent.fm/