[MPlayer-dev-eng] Patch: Jack audio output

[Ed Wildgoose]

Hi Reimar,

>>We can do, but the actual API changed in bio2jack to make it an unsigned 
>>char...  For whatever reason we aren't loading the bio2jack.h file and 
>>instead portions of it are reproduced here directly...
>>    
>>
>
>Don't know why, I wouldn't mind changing it, although people will have
>to have it lie around somewhere like this...
>  
>

OK, well since it is working today, and if you aren't going to add in 
the bio2jack.c file as well, then I propose that we consider to 
basically agree to do it, but lets put it on hold until the API settles 
down a little and there seems to be a reason to do it.  Sound OK?


>Now that there are bio2jack releases and not only CVS everyone can
>downgrade to the last release if it breaks...
>Having bio2jack directly included would require someone regularly
>updating it - a waste of developer time at least I won't be happy with,
>which is why it was already refused at the time of first commit.
>And when I look at what a mess every update of libfaad or libmpeg2 seems
>to cause I am even more against it.
>  
>

I'm only going to lightly protest and say that the author feels that for 
such a small library it seems silly to build it as a shared library.  We 
are only talking about a medium sized c file here.  It may well be a 
waste of developer time, but at least I guess that mplayer doesn't 
suddenly break if you upgrade the shared library...  I would be happy to 
maintain this part of mplayer for example.

However, I think there will be a bit more rapid evolution of bio2jack 
and then it will settle down a little.  It's only really changing right 
now, I think because I have been submitting requests/changes to the 
author in order to make it work so nicely with video.  Once I get 
mplayer and mythtv working well then I think it will settle down a bit.

Anyway, mplayer is the only app which needs it as a shared library.  I 
will protest lightly that I think it's a strange thing to do, but I 
don't have strong feelings on the issue, so will go with whatever you 
decide.  Just bear in mind that it is not really much of a library, more 
a C file and a header file.  In mythtv I just included the relevant 
files directly.


>It wouldmake much more sense if those "bytes in 
>the jack subsystem" were included in the value from the
>JACK_GetJackLatency...
>  
>

Possibly, but it might well be useful to know the individual values 
under some circumstances.  One of them will change dynamically for 
example, the other is relatively static (although some functions 
appeared in recent jack builds to actually allow you to change the size 
of the buffer size at runtime, but I can't imagine why anyone is using 
that in anger...?)


>Here you use spaces instead of tabs for intendation... that makes the
>code look bad with a different tabsize...
>  
>

Aha, ooops.  Sorry, I did try quite hard to keep tabs wherever possible 
since that is what was used originally.  Do you need me to resubmit this 
patch with the tabs changed?


So barring the tabs to spaces change, is this good to commit? 

Thanks for your helpful comments and review on this patch.  Much appreciated

Ed W