<p>On Jul 28, 2011 3:15 AM, "Rogério Brito" <<a href="mailto:rbrito@ime.usp.br">rbrito@ime.usp.br</a>> wrote:<br>
><br>
> Hi, Andreas and others.<br>
><br>
> On Jul 27 2011, Andres Mejia wrote:<br>
> > 2011/7/27 Rogério Brito <<a href="mailto:rbrito@ime.usp.br">rbrito@ime.usp.br</a>>:<br>
> > > Oh, shit. I'll take a look at that... I thought that the code were<br>
> > > ifdef'ed... Isn't it?<br>
> ><br>
> > I have a patch here with some changes we do to the build system.<br>
><br>
> OK.<br>
><br>
> > We outright disable the use of the GTK frontend, since it requires a very<br>
> > old version of GTK. I think there are better tools out now (audacity<br>
> > maybe). Perhaps the GTK frontend should be removed from lame entirely.<br>
><br>
> Sure, it has not compiled for ages and I think that killing it leaves us<br>
> with less cruft. If someone happens to want to port it forward to GTK2 or<br>
> GTK3, though, we may reinstate it...<br>
><br>
> > The other change simply changes how the debian directory is included<br>
> > in the generated tarball. The debian directory still gets included,<br>
> > but there's no need to have any "Makefile*" files in the debian<br>
> > directory. This would remove our need to patch the build system for<br>
> > each release of lame.<br>
> (...)<br>
> ><br>
> > CLEANFILES = lclint.txt $(EXTRA_PROGRAMS)<br>
><br>
> Can you please send "git format-patch/git send-email" patches against my<br>
> github tree (branch origin), with one topic per patch?<br>
><br>
> Your patches came foo-bared and they are changing at least two things (the<br>
> GTK stuff and the debian subdir).<br>
><br>
> > I've started replacing frontend/portableio.* code. I should have<br>
> > something available later on tonight.<br>
><br>
> Please see the attached patch to kill portableio.c if you are using<br>
> libsndfile (which we should be in debian, anyway).</p>
<p>Please note my earlier email about sndfile. Enabling sndfile for lame breaks k3b and any app using kio slaves.</p>
<p>If you look around, the only fix is to use lame's internal io functions when this breakage happens.</p>
<p>> It *is* dirty and there's one function (the one to read stuff) that is not<br>
> used, but works for the purpose of decoding some files (I tested).<br>
><br>
> I should probably adapt something from a patch that I once wrote to mplayer<br>
> ages ago when I was trying to make the generated PCM files works for little<br>
> and big-endian arches (hey, I *care* about PowerPC).<br>
><br>
> Or if libsdnfile does that already, that's even better. But for the sake of<br>
> just shipping a first package in unstable, I am very anxious... :-) I want<br>
> to see this darn thing in Debian proper! :-)<br>
><br>
> > One final thing, I delete the "LICENSE" file from the upstream tarball<br>
> > since the "COPYING" file has the actual license text and the contents<br>
> > of "LICENSE" seem to what is already said in the "README" file.<br>
> > Perhaps you might want to look into that too.<br>
><br>
> Will look into that for the next release. Please, *do* ping me if I forget<br>
> about this.<br>
><br>
><br>
> Thanks,<br>
><br>
> --<br>
> Rogério Brito : rbrito@{<a href="http://ime.usp.br">ime.usp.br</a>,<a href="http://gmail.com">gmail.com</a>} : GPG key 4096R/BCFCAAAA<br>
> <a href="http://rb.doesntexist.org">http://rb.doesntexist.org</a> : Packages for LaTeX : <a href="http://algorithms.berlios.de">algorithms.berlios.de</a><br>
> DebianQA: <a href="http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br">http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br</a><br>
</p>