On Mi, Mai 26, 2010 at 10:10:18 (CEST), Jonas Smedegaard wrote:

> On Wed, May 26, 2010 at 09:33:57AM +0200, Reinhard Tartler wrote:
>>On Mi, Mai 26, 2010 at 09:17:14 (CEST), Jonas Smedegaard wrote:
>>> On Wed, May 26, 2010 at 06:59:00AM +0000, siret...@users.alioth.debian.org 
>>> wrote:
>>>>The following commit has been merged in the master branch:
>>>>commit e0636d22570edd78dcc81797f84336ffbd810b95
>>>>Author: Reinhard Tartler <siret...@tauware.de>
>>>>Date:   Wed May 26 08:30:37 2010 +0200
>>>>    copy in mencoder.c from upstream
>>>>    this is a cowboy approach that places mencoder.c in
>>>>    debian/mencoder.c. This is of course a gross hack and should be reverted
>>>>    on the next upstream upgrade.
>>> [ huge patch snippet ]
>>>>diff --git a/debian/rules b/debian/rules
>>>>index 0ba540f..c9c289d 100755
>>>>--- a/debian/rules
>>>>+++ b/debian/rules
>>>>@@ -93,8 +93,12 @@ endif
>>>> # https://wiki.ubuntu.com/DistCompilerFlags
>>>>+# cowboy in mencoder.c manually fetched from upstream to avoid having to 
>>>>+# a new upstream tarball. Will be dropped with a new upstream upgrade
>>>>+mencoder.c: debian/mencoder.c
>>> I fail to see the point in hiding upstream code in the Debian packaging
>>> - even without mentioning it in debian/copyright!
>>mencoder has exactly the same copyright as mplayer itself. My reading of
>>debian/copyright does not leave any concerns about the licensing of
>>mencoder.  What parts are unclear according to your reading?
> debian/copyright states that the packaging (which I read as the contents
> of the debian/ subdir) is owned by Dariush Pietrzak and A Mennucci.

yes, from lines 1 to 20. The rest of the file talks about the upstream

> Thank you for telling me here(!) the source and copyright of that
> particular file below debian/ - I would prefer if that information was
> contained in debian/copyright too, or at least in the header of the code
> (stored as a patch, conveniently leaving room for such meta info).

Since mencoder is part of mplayer, I thought the licensing was clear,
but if you find it confusing, we could clarify that in a sentence or
two in debian/copyright.

>>> I strongly suggest to either place it as a proper patch with DEP3
>>> header, or roll a new tarball.
>> I disagree here. IMO, DEP3 is still way too much in flux to be
>> seriously considered, please don't force me to use it.  Moreover, DEP3
>> (currently) mandates a lot of very annoying and hairsplitting work by
>> considering each and every source file which is not exactly required
>> by debian policy. My opinion might change if DEP3 matures and #472199
>> makes progress.
> It seems to me that you are talking about DEP5 - the proposed (status:
> draft) machine-readable debian/copyright file format.  Indeed that one
> is in flux (but not a lot) and even when/if decided it is only optional.
> I am talking about DEP3 - the proposed (status: candidate)
> machine-readable debian/patches/ header format.  More info here:
> http://dep.debian.net/deps/dep3/

You are completely right that I horribly confused DEP3 and DEP5. We
talked about debian/copyright and using patches for that matter feels
really strange to me so that I must have skipped that word.

I've already been working far too long on the mplayer and ffmpeg package
today, finally need to get back to my day-job.

>>> And to document its licensing if placed below debian/ .
>>I could also have added it as patch in debian/patches, but I think that
>>would have been even sillier.
> Why do you find that sillier?
it requires additional work overhead to work (diff, update, etc.)
with. Commit logs don't contain diff-on-diffs. In case mencoder.c is
changed, debdiffs become more readable. Morover, changes are less likely
to confuse 'git annotate'.  In short: I find this approach much more
practicable and easier to work with.

Reinhard Tartler, KeyID 945348A4

pkg-multimedia-maintainers mailing list

Reply via email to