Re: fix for the winemp3 module

[Mike McCormack]

Deun wrote:
> Changelog by Denis Huguet ([e-mail])
> - 2006/10/17 :
>
> Major diff between winemp3's layer3.c file and the one from
> MPlayer-1.0pre7try2/mp3lib/layer3.c for fixing decoding
> glitches found out when opening mp3 file with GoldWave 4.21
> with wine-0.9.22 (same with 0.9.23)

Hi Denis,

That's a big change.  It would be nicer to submit it piece by piece, fix 
by fix, clarifying which bug you are fixing and to get rid of the 
extraneous #if 0's and C++ comments.

That would make regression analysis easier in case the patch introduces 
a bug.

thanks,

Mike

fix for the winemp3 module

[Eric Pouech]

Mike McCormack wrote:
> Deun wrote:
> > Changelog by Denis Huguet ([e-mail])
> > - 2006/10/17 :
> >
> > Major diff between winemp3's layer3.c file and the one from
> > MPlayer-1.0pre7try2/mp3lib/layer3.c for fixing decoding
> > glitches found out when opening mp3 file with GoldWave 4.21
> > with wine-0.9.22 (same with 0.9.23)
>
> Hi Denis,
>
> That's a big change.  It would be nicer to submit it piece by piece, 
> fix by fix, clarifying which bug you are fixing and to get rid of the 
> extraneous #if 0's and C++ comments.
>
> That would make regression analysis easier in case the patch 
> introduces a bug.
>
> thanks,
>
> Mike

also, mplayer is GPL:ed, so you cannot use their work in wine (which is 
LGPL)
so the patch has to be made from the code of the original library 
(mpglib), which is LGPL
A+

fix for the winemp3 module

[Deun]

Hi,

Well, I'm not so familiar with detailed licencing stuff, sorry
for not noticing.
However, I've worked alot on mp3 codec a few years ago to make
my own assembly codec (sorry, not x86) and after comparing many
different sources, all I can say is that there aren't so many
ways to decode the stream datas, so in the end all sources
looks very similar.
Also note that I haven't simply copied code from Mplayer's
source, I've adapted it to current code. But I guess I'm being
too naive saying this regarding to licence stuff. And the code
which has most changed in my patch (about linbits decoding) is
not something made by MPlayer's team as there was exactly the
same method used a few years before in the old AMP 0.7.1 source
(I'm not sure about exact version but it was around that one).
Anyway if you think that my patch would lead to licence issues,
then it's ok to drop it, don't worry.

Regards,

Denis.

--- Eric Pouech  wrote:
> Mike McCormack wrote:
> > Deun wrote:
> > > Changelog by Denis Huguet ([e-mail])
> > > - 2006/10/17 :
> > >
> > > Major diff between winemp3's layer3.c file and the one
>
> from
> > > MPlayer-1.0pre7try2/mp3lib/layer3.c for fixing decoding
> > > glitches found out when opening mp3 file with GoldWave
>
> 4.21
> > > with wine-0.9.22 (same with 0.9.23)
> >
> > Hi Denis,
> >
> > That's a big change.  It would be nicer to submit it piece
>
> by piece,
> > fix by fix, clarifying which bug you are fixing and to get
>
> rid of the
> > extraneous #if 0's and C++ comments.
> >
> > That would make regression analysis easier in case the
>
> patch
> > introduces a bug.
> >
> > thanks,
> >
> > Mike
>
> also, mplayer is GPL:ed, so you cannot use their work in wine
> (which is 
> LGPL)
> so the patch has to be made from the code of the original
> library 
> (mpglib), which is LGPL
> A+

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com

fix for the winemp3 module

[Eric Pouech]

Deun wrote:
> Hi,
>
> Well, I'm not so familiar with detailed licencing stuff, sorry
> for not noticing.
> However, I've worked alot on mp3 codec a few years ago to make
> my own assembly codec (sorry, not x86) and after comparing many
> different sources, all I can say is that there aren't so many
> ways to decode the stream datas, so in the end all sources
> looks very similar.
> Also note that I haven't simply copied code from Mplayer's
> source, I've adapted it to current code. But I guess I'm being
> too naive saying this regarding to licence stuff. And the code
> which has most changed in my patch (about linbits decoding) is
> not something made by MPlayer's team as there was exactly the
> same method used a few years before in the old AMP 0.7.1 source
> (I'm not sure about exact version but it was around that one).
> Anyway if you think that my patch would lead to licence issues,
> then it's ok to drop it, don't worry.

I thought at first you just diff:ed the file against mplayer's code
however, as mplayer's uses the old mpglib code, one could also ask 
mplayer's team to release the diffs they made to mpglib as LGPL (which 
they have to...)
the bad side of this is that mpglib is no longer maintained, so we can't 
expect fixes from that lib (the only LGPL mpg3 decoding lib I know of)
updates to mpglib now appear in mpg123 a player (LGPL), but with lots of 
changes since the original code (and no longer a library on its own)

anyway, the fixes seem to be needed... that's Alexandre's final call
A+