Bug#781938: RFS: vid.stab/0.98b-1

[Vincent Pinon]

Hello,

One month  later, I still haven't forgotten that task, sorry for the delay ;)
new upload here:
http://mentors.debian.net/package/vid.stab

Le mardi 27 octobre 2015, 00:16:57 Andreas Cadhalpun a écrit :
[snip]
>  * d/control:
> - 'Priority: optional' not extra
> - packages should be 'Multi-Arch: same'
>  * d/copyright: misses copyright of 'Alexey Osipov'
Done, thanks for guidance!

> The tests could be run e.g. with:
> cd tests; cmake .; make all; ./tests
> However, the return code is wrong [2].
I included your patch, but I'm not sure how to add & run test target in 
debian/rules.
This might go into future improvements, with the runtime loading of SSE on 
i386? (Felipe's suggestion)

I don't know if this packaging activity is worth versioning on alioth, as 
upstream doesn't seem to evolve any/much more?
If yes anyway, am I allowed to initialize a git repo?

Many thanks again,
Cheers,

Vincent.

signature.asc
Description: This is a digitally signed message part.

Bug#781938: RFS: vid.stab/0.98b-1

[Felipe Sateler]

On 29 November 2015 at 18:17, Vincent Pinon  wrote:
> Hello,
>
> One month  later, I still haven't forgotten that task, sorry for the delay ;)
> new upload here:
> http://mentors.debian.net/package/vid.stab
>
> Le mardi 27 octobre 2015, 00:16:57 Andreas Cadhalpun a écrit :
> [snip]
>>  * d/control:
>> - 'Priority: optional' not extra
>> - packages should be 'Multi-Arch: same'
>>  * d/copyright: misses copyright of 'Alexey Osipov'
> Done, thanks for guidance!
>
>> The tests could be run e.g. with:
>> cd tests; cmake .; make all; ./tests
>> However, the return code is wrong [2].
> I included your patch, but I'm not sure how to add & run test target in 
> debian/rules.

You can do:

override_dh_auto_test:
  #test commands here

> This might go into future improvements, with the runtime loading of SSE on 
> i386? (Felipe's suggestion)

I'd add -DSSE2_FOUND=false to the cmake invocation when run on i386.
This appears to be enough to disable the sse flags

>
> I don't know if this packaging activity is worth versioning on alioth, as 
> upstream doesn't seem to evolve any/much more?
> If yes anyway, am I allowed to initialize a git repo?

I think git history is always useful, although pre-first release
history not so much (so we can create the repo with the uploaded
package). Please apply to the pkg-multimedia project on alioth, and
read the wiki

https://alioth.debian.org/projects/pkg-multimedia
https://wiki.debian.org/DebianMultimedia/DevelopPackaging

-- 

Saludos,
Felipe Sateler

Bug#781938: RFS: vid.stab/0.98b-1

[Andreas Cadhalpun]

Hi,

On 26.10.2015 13:15, Felipe Sateler wrote:
> I see that Andreas Cadhalpun (in CC) expressed interest in this
> package and apparently already forwarded that patch upstream. Andreas,
> did you ever get a response on that?

No, unfortunately not. That's why I didn't pursue this any further,
as I assumed that upstream is not active anymore. Am I wrong about that?

Best regards,
Andreas

Bug#781938: RFS: vid.stab/0.98b-1

[Andreas Cadhalpun]

On 26.10.2015 23:00, Felipe Sateler wrote:
> On 26 October 2015 at 18:34, Andreas Cadhalpun
>  wrote:
>> On 26.10.2015 13:15, Felipe Sateler wrote:
>>> I see that Andreas Cadhalpun (in CC) expressed interest in this
>>> package and apparently already forwarded that patch upstream. Andreas,
>>> did you ever get a response on that?
>>
>> No, unfortunately not. That's why I didn't pursue this any further,
>> as I assumed that upstream is not active anymore. Am I wrong about that?
> 
> The last commit in github[1] is from May. So not really active, no.
> But that doesn't mean we shouldn't provide the package (as long as
> there is a willing maintainer).

Yes, it just means that one can't expect any help from upstream.

Regarding Vincent's packaging:

 * d/control:
- 'Priority: optional' not extra
- packages should be 'Multi-Arch: same'
 * d/copyright: misses copyright of 'Alexey Osipov'

The tests could be run e.g. with:
cd tests; cmake .; make all; ./tests

However, the return code is wrong [2].

Best regards,
Andreas


2: 
https://bugs.debian.org/cgi-bin/bugreport.cgi?att=2;filename=0002-tests-return-0-upon-success.patch;bug=781938;msg=20

Bug#781938: RFS: vid.stab/0.98b-1

[Felipe Sateler]

On 26 October 2015 at 18:34, Andreas Cadhalpun
 wrote:
> Hi,
>
> On 26.10.2015 13:15, Felipe Sateler wrote:
>> I see that Andreas Cadhalpun (in CC) expressed interest in this
>> package and apparently already forwarded that patch upstream. Andreas,
>> did you ever get a response on that?
>
> No, unfortunately not. That's why I didn't pursue this any further,
> as I assumed that upstream is not active anymore. Am I wrong about that?

The last commit in github[1] is from May. So not really active, no.
But that doesn't mean we shouldn't provide the package (as long as
there is a willing maintainer).

[1] https://github.com/georgmartius/vid.stab




-- 

Saludos,
Felipe Sateler

Bug#781938: RFS: vid.stab/0.98b-1

[Rogério Brito]

Hi.

On Oct 26 2015, Felipe Sateler wrote:
> I'm adding the ITP (and all participants) back to CC to let them know
> about your work.

Please, go ahead. I'm way too busy right now to maintain this package at
this moment.


Thanks,

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

Bug#781938: RFS: vid.stab/0.98b-1

[Felipe Sateler]

On 26 October 2015 at 02:00, Vincent Pinon  wrote:
>> - You might want to keep the epoch so that people installing those
>> packages can upgrade to this version when it lands in debian.
> I also think that we should start with a clear base in Debian,
> so I reset the changelog (wiping the dmo & ubuntu-ppa mix, but keeping 
> copyright)...
>
>> - Please use the alioth list as the maintainer (check any other team
>> package [eg, csound] for the name and address). In general please read
>> the guidelines in the wiki[1]
> Sorry, fixed, and read again the ref page :)
>
>> - Please drop the builddeb override, xz is the default.
> done
>
>> - On i386, we need to disable sse, as it is not guaranteed to be
>> available on all hosts. I think it can be loaded at runtime by using
>> the hardware capabilities feature of glibc linker:
>>- Compile the package twice, once with and once without sse.
>>- Ship the non-sse version in /usr/lib/i386-linux-gnu/
>>- Ship the sse version in /usr/lib/i386-linux-gnu/sse2
>>- Then ld.so will choose the appropriate version at runtime (see
>> the ld.so manpage for details).
> Hum, a bit beyond my packaging knowledge...
> Do you have an example?

It is a bit convoluted. Maybe we could start by force-disabling SSE on
i386 and then later add a second sse version?

>
>> - debian/copyright needs dates updated (2010-2014 for the upstream part).
> Done
>
> All these changes in forced-upload again:
> https://mentors.debian.net/package/vid.stab
>
>> - You might want to forward the patch upstream.
> Right, I will do it

I see that Andreas Cadhalpun (in CC) expressed interest in this
package and apparently already forwarded that patch upstream. Andreas,
did you ever get a response on that?

I'm adding the ITP (and all participants) back to CC to let them know
about your work.

-- 

Saludos,
Felipe Sateler