Bug#410426: xine-lib: misleading package descriptions

2007-02-11 Thread Fabian Greffrath
Am Samstag, den 10.02.2007, 21:39 +0100 schrieb Reinhard Tartler:
 Yes, the selection is indeed quite random. The main motivation for the
 split was to promote the dependencies from 'Recommends' to 'Depends'. In
 order to not get too many dependencies from libxine1 (which has a lot of
 reverse depends), I moved some plugins to seperate packages.
 
 I tried to keep the libxine1 package functional, without imposing too
 many and 'heavyweight' dependencies.

This sounds rational, maybe the plugin selection for the packages could 
be motivated like this:

- The `libxine1' package contains all plugins which do not pull in
further dependencies, except those who are expected on a desktop system
anyway (e.g. libogg*, libvorbis*, libtheora*). [Did you notice that
libxine1 has dependencies on libxine1?]

- The `libxine1-ffmpeg' contains all plugins which pull in further
dependencies, which may not be part of a minimal installation. However,
the suffix `-ffmpeg' would be inappropriate and not describe the
package's contents. Something like `libxine1-extras' might fit better.

I think this is similar to the way the gstreamer maintainers split up
their plugin packages (see README.Debian and HACKING.Debian in the
`gst-plugins-base0.10-0.10.10' tarball). 

By the way, the separation of KDE- and GNOME-related plugins was the
best idea anyway! ;)

 Yes, this is indeed necessary. Perhaps you can make some suggestions for
 that?

Sure! As soon as the criteria for the plugin selection are defined. ;)

 Thanks again for your report!

Thanks for your great work on this package!

Nice greetings,
Fabian



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#410426: xine-lib: misleading package descriptions

2007-02-10 Thread Fabian Greffrath
package: xine-lib
version: 1.1.4-2
severity: minor
tags: experimental

Hello,

since the split-up of the `libxine1' package the short descriptions of
the fragmented packages are misleading. Now all of them are `the xine
video/media player library, binary files' without further explanation.

Secondly, in the new package's description there is a dot `.' visible
where it should most propably show an empty line.

Thirdly, the selection of plugins which make it into `libxine1' and
`libxine1-ffmpeg' seems quite random:
In `libxine1-ffmpeg' there is the `xineplug_decode_mad.so' plugin which
obviously has nothing to do with ffmpeg itself. The same applies to the
a52 and dts plugins. Now, one might think that this package is for input
plugins in general, maybe even for decoding of patented formats (i.e.
mp3). But then there are even more input plugins in `libxine1', like
those for vorbis. And, more puzzling, there are even the plugins for
decoding of real-audio etc.

Maybe I get something wrong, but could you please clean up the packages'
descriptions to explain more precisely what kind of codecs they contain
and why they do?

Thank you very much!

Nice greetings,
Fabian



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#410426: xine-lib: misleading package descriptions

2007-02-10 Thread Reinhard Tartler
Fabian Greffrath [EMAIL PROTECTED] writes:

 package: xine-lib
 version: 1.1.4-2
 severity: minor
 tags: experimental

Thanks for your bug report!

 since the split-up of the `libxine1' package the short descriptions of
 the fragmented packages are misleading. Now all of them are `the xine
 video/media player library, binary files' without further explanation.

Oh, you're right. I'll fix the description in the next upload

 Secondly, in the new package's description there is a dot `.' visible
 where it should most propably show an empty line.

I'll have a log

 Thirdly, the selection of plugins which make it into `libxine1' and
 `libxine1-ffmpeg' seems quite random:
 In `libxine1-ffmpeg' there is the `xineplug_decode_mad.so' plugin which
 obviously has nothing to do with ffmpeg itself. The same applies to the
 a52 and dts plugins. Now, one might think that this package is for input
 plugins in general, maybe even for decoding of patented formats (i.e.
 mp3). 

Yes, the selection is indeed quite random. The main motivation for the
split was to promote the dependencies from 'Recommends' to 'Depends'. In
order to not get too many dependencies from libxine1 (which has a lot of
reverse depends), I moved some plugins to seperate packages.

 But then there are even more input plugins in `libxine1', like
 those for vorbis. And, more puzzling, there are even the plugins for
 decoding of real-audio etc.

I tried to keep the libxine1 package functional, without imposing too
many and 'heavyweight' dependencies.

 Maybe I get something wrong, but could you please clean up the packages'
 descriptions to explain more precisely what kind of codecs they contain
 and why they do?

Yes, this is indeed necessary. Perhaps you can make some suggestions for
that?

Thanks again for your report!

-- 
Gruesse/greetings,
Reinhard Tartler, KeyID 945348A4


pgpvYFSa9bML9.pgp
Description: PGP signature