Packaging current lame CVS snapshot for Debian -- good or bad idea?

classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
Dear lame-devs,

I am (one of the) maintainer(s) of the lame package in Debian:

https://tracker.debian.org/pkg/lame

We are currently packaging the latest stable release 3.99.5 and over
time, we have added more and more patches to our package, most of them
have already been applied in the lame CVS repository as well (thanks
rbrito):

http://anonscm.debian.org/cgit/pkg-multimedia/lame.git/tree/debian/patches

Now that Debian "jessie" has been released, I'd like to prepare a new
Debian package for lame based on a current CVS snapshot, so I can get
rid of most of the patches we currently apply. Do you know of any issues
with the current state of the code in the repository or another reason
why this might turn out as a bad idea? Is there anything that should
keep me from packaging it? Anything I should know?

Thank you very much already!

Cheers,

- Fabian


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Rogério Brito
Hi, Fabian and others.

On Fri, May 8, 2015 at 4:32 AM, Fabian Greffrath <[hidden email]> wrote:
> We are currently packaging the latest stable release 3.99.5 and over
> time, we have added more and more patches to our package, most of them
> have already been applied in the lame CVS repository as well (thanks
> rbrito):

Thanks. I have been applying some patches that people have sent me,
but I still have quite a bit of work to do, as we have to triage
contributions from a lot of people and sourceforge's interface is not
entirely manipulable by e-mail or other convenient ways.

Anyway, as I have already talked privately with Fabian, I think that
the best person to answer about packaging a snapshot of LAME would be
Robert (in CC), since he has some patches fine-tuning the audio parts
of the project. Robert, is generating a package from the current
HEAD/master branch of CVS likely to have negative effects in
comparison with lame 3.99.5?

If so, what if we released a small bug fix of 3.99, say, 3.99.6? This
way, I think that we could address things that other people want and
we would save people from patching their local copies of lame.

> http://anonscm.debian.org/cgit/pkg-multimedia/lame.git/tree/debian/patches
>
> Now that Debian "jessie" has been released, I'd like to prepare a new
> Debian package for lame based on a current CVS snapshot, so I can get
> rid of most of the patches we currently apply.

Now, this is a question to Fabian: can you remind me why you are
repackaging lame instead of using the tarball that we ship?

Also, I saw that debian/rules uses two switches:

* --enable-expopt=full: I am not really sure if this has been touched
in the last decade or not, but I recall a discussion from a few years
ago that we had developers using (for the sake of simplicity) one of
the ready made Makefile that are in the repository.
* --with-fileio=lame: any reason to not use sndfile? It allows feeding
a considerable greater amount of file types into lame (including
FLAC). I personally use a local copy of lame recompiled with sndfile
just for transcoding mp3 and flac files into files with lower bitrate
(yes, file size is still something that matters).

OK, I will stop here. I hope that Robert can give everybody a position
on what he considers the status of HEAD.


Regards,

--
Rogério Brito : rbrito@{ime.usp.br,gmail.com} : GPG key 4096R/BCFCAAAA
http://cynic.cc/blog/ : github.com/rbrito : profiles.google.com/rbrito
DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
Hi Rogério et al.,

thank you for taking up on the issue!

Am Montag, den 11.05.2015, 03:49 -0300 schrieb Rogério Brito:
> Now, this is a question to Fabian: can you remind me why you are
> repackaging lame instead of using the tarball that we ship?

There are three reasons given by Andres in:
http://anonscm.debian.org/cgit/pkg-multimedia/lame.git/tree/debian/README.source

* Removal of upstream's "debian" directory.
* Inclusion of GTK-1 autoconf directives.
* Regeneration of build system using autoreconf.

The first should not be an issue anymore with Debian's "3.0 (quilt)"
source format, the second has already been fixed by yourself im lame's
CVS repository and the third is not an issue anymore with using
dh-autoreconf. So, no, I am not going to repack the lame source tarball
for Debian anymore and instead use the pristine sources as released.

> * --enable-expopt=full: I am not really sure if this has been touched
> in the last decade or not, but I recall a discussion from a few years
> ago that we had developers using (for the sake of simplicity) one of
> the ready made Makefile that are in the repository.

Frankly, I have no idea about this switch. :/

> * --with-fileio=lame: any reason to not use sndfile? It allows feeding
> a considerable greater amount of file types into lame (including
> FLAC). I personally use a local copy of lame recompiled with sndfile
> just for transcoding mp3 and flac files into files with lower bitrate
> (yes, file size is still something that matters).

The sndfile backend did not allow reading from stdin and that broke some
"KDE apps relying on kio_audiocd slave for CD ripping". Please see the
changelog for lame_3.98.4+repack1-1 in Debian and how Andres disabled
the feature even after I enabled it and applied the patch to allow for
reading from stdin from Gentoo.

However, I have lost track of that issue and if the patch in question
has been applied upstream in lame (I am sure it is) we should consider
building the Debian package with the sndfile backend again.

> OK, I will stop here. I hope that Robert can give everybody a position
> on what he considers the status of HEAD.

I hope so, too!

Again, thank you very much for the discussion.

Best regards,

- Fabian


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
Am Montag, den 11.05.2015, 09:10 +0200 schrieb Fabian Greffrath:
> Frankly, I have no idea about this switch. :/

So, I have checked what the --extopt parameter does and it turns out it
merely enables some additional compiler switches. However, it enables
these switches based on arch and gcc version, but fails to detect the
version of gcc that I am currently running, i.e. gcc 5.1.

So, I believe that these compiler switches should be either thoroughly
reviewed or maybe even removed altogether -- gcc has really improved on
optimizations in the last decade. ;)

Cheers,

- Fabian


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Robert Hegemann
In reply to this post by Fabian Greffrath
Hi,

I was already thinking about to release 3.99.6, containing almost all fixes from the main branch. Current psy tunings are not ready for a beta release, there is more work needed to be done.

Ciao Robert






-------- Ursprüngliche Nachricht --------
Von: [hidden email]
Datum: 11.05.2015  08:49  (GMT+01:00)
An: Fabian Greffrath <[hidden email]>
Cc: Robert Hegemann <[hidden email]>, [hidden email]
Betreff: Re: [Lame-dev] Packaging current lame CVS snapshot for Debian --  good or bad idea?

...
Anyway, as I have already talked privately with Fabian, I think that
the best person to answer about packaging a snapshot of LAME would be
Robert (in CC), since he has some patches fine-tuning the audio parts
of the project. Robert, is generating a package from the current
HEAD/master branch of CVS likely to have negative effects in
comparison with lame 3.99.5?

If so, what if we released a small bug fix of 3.99, say, 3.99.6? This
way, I think that we could address things that other people want and
we would save people from patching their local copies of lame.

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
Hi Robert,

Am Mittwoch, den 13.05.2015, 10:14 +0200 schrieb robert:
> I was already thinking about to release 3.99.6, containing almost all
> fixes from the main branch. Current psy tunings are not ready for a
> beta release, there is more work needed to be done.

thank you for your reply and sorry for me asking again, but I have no
idea what the "psy tunings" are and if "main branch" is the same as
"HEAD" in CVS language. So, this means that the current CVS HEAD is not
in a state that should be released and/or packaged, right?

Cheers,

Fabian

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Rogério Brito
Hi to all of you.

On May 13 2015, Fabian Greffrath wrote:
> Am Mittwoch, den 13.05.2015, 10:14 +0200 schrieb robert:
> > I was already thinking about to release 3.99.6, containing almost all
> > fixes from the main branch. Current psy tunings are not ready for a
> > beta release, there is more work needed to be done.
>
> thank you for your reply and sorry for me asking again, but I have no
> idea what the "psy tunings"

"psy tunings" = "Psycho-acoustic model adjustments". Is that clearer? :)

> are and if "main branch" is the same as "HEAD" in CVS language.

I think that that's what Robert meant, yes.

> So, this means that the current CVS HEAD is not in a state that should be
> released and/or packaged, right?

Yes, that's what I thought, but I also thought that it would be best to ask
here.  If Robert is thinking the same things that I am thinking, the release
3.99.6 would be essentially what you have in Debian, plus (perhaps?) a few
minor patches from other projects.

They should all be contained in my git repository and releasing that would
be, essentially, cherry-picking the releases and putting them onto the 3.99
branch (I don't remember if my git repo has one such branch).

It would be a good thing to check with the people from handbrake if they
have some patches (they usually do).

Also regarding handbrake, the last time I checked, they were disabling the
joint-stereo mode of LAME when the bitrate was "high", but I suspect that
this is not exactly a recommended setting.

Robert, just confirming: joint-stereo is the preferred mode of compression
at all bitrates, right? If so, that may have to be patched (in handbrake
upstream or in Debian and other distributions).


Thanks,

--
Rogério Brito : rbrito@{ime.usp.br,gmail.com} : GPG key 4096R/BCFCAAAA
http://cynic.cc/blog/ : github.com/rbrito : profiles.google.com/rbrito
DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
Hi again,

Am Mittwoch, den 13.05.2015, 19:15 -0300 schrieb Rogério Brito:
> "psy tunings" = "Psycho-acoustic model adjustments". Is that clearer? :)

At least that's more words to describe it. I still have no idea what it
actually does, though, so I'll leave it up to you guys to evaluate if it
is in a releasable state or not. ;)

> Yes, that's what I thought, but I also thought that it would be best to ask
> here.  If Robert is thinking the same things that I am thinking, the release
> 3.99.6 would be essentially what you have in Debian, plus (perhaps?) a few
> minor patches from other projects.

This sounds like a very good idea!

> It would be a good thing to check with the people from handbrake if they
> have some patches (they usually do).

Gentoo tends to have good patches, too:
https://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/media-sound/lame/files/

Cheers,

Fabian


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
In reply to this post by Rogério Brito
Am Mittwoch, den 13.05.2015, 19:15 -0300 schrieb Rogério Brito:
> Yes, that's what I thought, but I also thought that it would be best to ask
> here.  If Robert is thinking the same things that I am thinking, the release
> 3.99.6 would be essentially what you have in Debian, plus (perhaps?) a few
> minor patches from other projects.

PS: That would not only be the actual patches from Debian, i.e. the ones
in debian/patches, but also (1) the demotion of the debian/ directory in
the lame sources from a build target to an extra directory and (2) the
inclusion of the GTK+ autoconf directives. Both are currently patched
directly into the Debian source tarball.

Thanks!

Fabian

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Rogério Brito
In reply to this post by Fabian Greffrath
Hi again.

On May 14 2015, Fabian Greffrath wrote:
> Hi again,
>
> Am Mittwoch, den 13.05.2015, 19:15 -0300 schrieb Rogério Brito:
> > "psy tunings" = "Psycho-acoustic model adjustments". Is that clearer? :)
>
> At least that's more words to describe it. I still have no idea what it
> actually does, though, so I'll leave it up to you guys to evaluate if it
> is in a releasable state or not. ;)

The psychoacoustic model is, broadly speaking, the simulation that the
encoder does of our auditory system (and our perception of what we hear).
The better we can model the auditory system, the better we know what matters
and what doesn't in a given piece of music/audio and what we can discard so
that we use the limited space for audio for that.

Robert is fine-tuning that critical part of LAME, which means that HEAD is
not releasable in his words.

> > Yes, that's what I thought, but I also thought that it would be best to ask
> > here.  If Robert is thinking the same things that I am thinking, the release
> > 3.99.6 would be essentially what you have in Debian, plus (perhaps?) a few
> > minor patches from other projects.
>
> This sounds like a very good idea!

Great that you guys like it. :)

> > It would be a good thing to check with the people from handbrake if they
> > have some patches (they usually do).
>
> Gentoo tends to have good patches, too:
> https://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/media-sound/lame/files/

I took a look at some of their patches (which I could have sworn to have
applied before), but I bumped into this one:

   https://bugs.gentoo.org/show_bug.cgi?id=454322

Which is reported here:

   http://sourceforge.net/p/lame/bugs/436/

Don't we have this problem on Debian also? I would have thought that we
would have had something like that, given that we compile tinfo separetely
from ncurses (or something like that).


Thanks,

--
Rogério Brito : rbrito@{ime.usp.br,gmail.com} : GPG key 4096R/BCFCAAAA
http://cynic.cc/blog/ : github.com/rbrito : profiles.google.com/rbrito
DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
Am Donnerstag, den 14.05.2015, 21:33 -0300 schrieb Rogério Brito:
> Don't we have this problem on Debian also? I would have thought that we
> would have had something like that, given that we compile tinfo separetely
> from ncurses (or something like that).

In Debian, lame is linked against libncurses which contains the initscr
symbol that configure checks for. The libncurses library, in turn, is
linked against libtinfo which contains the tgetent symbol that lame
actually uses. Checking directly for the presence of the tgetent symbol
in libtinfo first and then in libncurses thus makes sense, IMHO.

- Fabian


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
Reply | Threaded
Open this post in threaded view
|

Re: Packaging current lame CVS snapshot for Debian -- good or bad idea?

Fabian Greffrath
In reply to this post by Rogério Brito
Hi Rogério and Robert,

Am Mittwoch, den 13.05.2015, 19:15 -0300 schrieb Rogério Brito:
> Yes, that's what I thought, but I also thought that it would be best to ask
> here.  If Robert is thinking the same things that I am thinking, the release
> 3.99.6 would be essentially what you have in Debian, plus (perhaps?) a few
> minor patches from other projects.

I hope you are all well!

So, has anything happened yet regarding a 3.99.6 point release?

Cheers,

Fabian


------------------------------------------------------------------------------

_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev