requesting sponsor for pd-cyclone

Hans-Christoph Steiner hans at at.or.at
Sat Nov 20 20:13:03 UTC 2010


On Nov 19, 2010, at 5:30 PM, Felipe Sateler wrote:

> On Thu, Nov 18, 2010 at 14:03, Hans-Christoph Steiner  
> <hans at at.or.at> wrote:
>> On Wed, 2010-11-17 at 16:50 -0300, Felipe Sateler wrote:
>>> On Sun, Nov 14, 2010 at 14:33, Hans-Christoph Steiner  
>>> <hans at at.or.at> wrote:
>>>>
>>>> On Nov 13, 2010, at 8:37 PM, Felipe Sateler wrote:
>>>>
>>>>> On Thu, Nov 11, 2010 at 23:17, Hans-Christoph Steiner <hans at at.or.at 
>>>>> >
>>>>> wrote:
>>>>>>
>>>>>> On Thu, 2010-11-11 at 09:19 +0100, Jonas Smedegaard wrote:
>>>>>>>
>>>>>>> On Thu, Nov 11, 2010 at 12:15:05AM -0500, Hans-Christoph  
>>>>>>> Steiner wrote:
>>>>>>>>
>>>>>>>> pd-cyclone is a huge library of all sorts of objects that are  
>>>>>>>> clones of
>>>>>>>> objects in Max/MSP4.5, Pd's proprietary cousin.  It also  
>>>>>>>> includes
>>>>>>>> utilities for importing Max files.
>>>>>>>>
>>>>>>>> pd-cyclone is a long-form dh package that is a wrapper to a  
>>>>>>>> very
>>>>>>>> complicated Make build system. pd-cxc is in
>>>>>>>> git.debian.org/pkg-multimedia.  It is a library without  
>>>>>>>> depends that
>>>>>>>> are new packages.  It builds on my i386, but it did not build  
>>>>>>>> on the
>>>>>>>> launchpad amd64 build machine, it died on this error, which  
>>>>>>>> has stumped
>>>>>>>> me:
>>>>>>>>
>>>>>>>> dpkg-genchanges -B -mUbuntu/amd64 Build Daemon <buildd at titanium.ppa 
>>>>>>>> >
>>>>>>>>>
>>>>>>>>> ../pd-cyclone_0.1~alpha55-1~maverick_amd64.changes
>>>>>>>>
>>>>>>>> dpkg-genchanges: arch-specific upload - not including arch- 
>>>>>>>> independent
>>>>>>>> packages
>>>>>>>> dpkg-genchanges: error: cannot read files list file: No such  
>>>>>>>> file or
>>>>>>>> directory
>>>>>>>
>>>>>>> It looks odd to me that you use no debhelper tools in the  
>>>>>>> binary-arch
>>>>>>> target - only in the binary-indep one.
>>>>>>
>>>>>> Thanks, that was the key.  This now builds on i386 and amd64  
>>>>>> and is
>>>>>> ready to go.
>>>>>
>>>>> The git repository is not fixed.
>>>>
>>>> I am not sure I understand. What I see on git.debian.org is the  
>>>> same as what
>>>> I have on my machine, and I used git-buildpackage to make the  
>>>> source package
>>>> to submit to Launchpad:
>>>>
>>>> http://launchpad.net/~eighthave/+archive/libdirs/+sourcepub/1362005/+listing-archive-extra
>>>
>>> Oh it was fixed, my bad. But there is a large commit that changes  
>>> many
>>> things at once (d634143), and the change was hidden there.
>>>
>>>>
>>>>> Also, why do you use long-form debhelper?
>>>>
>>>>
>>>> Because I didn't see a way to do it with short-form. This library  
>>>> has a very
>>>> elaborate build system that I've tried for years to figure out.   
>>>> Instead
>>>> I've found the best way is to wrap it into something more  
>>>> understandable,
>>>> and long-form seemed the easiest way to do that.
>>>
>>> Indeed, the makefile is quite convoluted. It also can't handle it  
>>> when
>>> the user passes CFLAGS and other variables. Also, it sets -j4
>>> unconditionally.
>>
>>> On further inspection, the package seems to bundle the pddp source.
>>> Are riddle, sickle and ViCious more code copies? We should not use
>>> code copies. We need to avoid using them.
>>
>> This is how cyclone is released.  The code in 'pddp', 'riddle', and
>> 'ViCious' are unused.  The 'pddp' code shared some code with pd-pddp,
>> but not all.  The code in 'sickle' is fully part of pd-cyclone and  
>> used.
>>
>> .hc
>>
>>>
>>>>
>>>> I am open to changing it to short-form, but I don't see a clear  
>>>> way to do it
>>>
>>> I don't really prefer short form, but it is just a matter of using
>>> override_dh_foo targets.
>
> It looks like it's going to be a busy week. I'll continue checking pd
> packages (hopefully) late next week.

No problem, thanks for uploading all these packages!

.hc


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

“We must become the change we want to see. - Mahatma Gandhi




More information about the pkg-multimedia-maintainers mailing list