Re: mpg123 porting: libmp3lame non-parallel init code?

Am Sun, 27 Aug 2017 13:41:37 +0200
schrieb Thomas Orgis <[email protected]>: 

> Am I right in assuming that lame_init() is _not_ the right place, as it
> can be called many times to create multiple instances in a parallel
> program?

Well, so far I settled for calling mpg123_init() on each call to
hip_decode_exit() (it has internal flag to only run once, and even if
tho concurring instances end up running, the filling of static buffers
should be safe even then). I dropped mpg123_exit() as that does nothing
and I don't see what it should do in future. I will mark it as a dud in
mpg123 API.

I am ignoring the deprecated lame_decode API for now. Should it stay?
The name lame_decode() is arguably better than hip_decode().

I am confident that a patch will arrive … for sure within September. I
do not have so much to change in libmpg123. I got most of the pinfo
extraction already, even. One might add the option to do a minimal
build with existing libmpg123 that does not extract the raw ID3 nor
pinfo for the analyser.

Oh, and one thing so far is just a stub: the use of error reporting
functions. But I guess I can configure libmpg123 to be silent also on
errors and instead reconstruct a message in the lame binding using
mpg123_strerror(), which then is handed to the user-supplied callback.
This is more tricky for debugging and general messages. There so far is
no way in mpg123 to route those through a callback. It might be an idea
to add it, though. And in the end, if I tell libmpg123 to be quiet, it
won't produce those …

I see that the libmp3lame API nicely supports integer or float
decoding. I still would like the get_audio() stuff also get floats from
libmpg123. I guess this is a separate patch, though, that should be
obvious once I demonstrated float output in the libmp3lame patch.


Alrighty then,

Thomas
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Lame-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/lame-dev