[maemo-community] Extras and Fremantle

[From: Jeremiah Foster]

On Mar 4, 2009, at 1:34 PM, David Greaves wrote:

> Jeremiah Foster wrote:
>> On Mar 3, 2009, at 10:47 PM, Eduardo Lima (Etrunko) wrote:
>
>>> What is left for us is a clear definition of
>>> the QA policy,
>>
>> This is important, and involves a lot of community input. It should
>> most likely have a dedicated mailing list.
>
> This bit worries me.
> I've seen the nit-picking over getting the website released.
> I feel that agility is under-weighted in the equation at the moment.

Agility is important, no doubt. I just thought we might keep the  
signal to noise ratio down for those who are uninterested in policy.
>
> (Bear in mind I wouldn't have suggested increasing the 'complexity'  
> of the
> release system if I didn't want to improve quality)
>
>> Unfortunately this is not trivial. The simple reason is that debian
>> uses tools that are built for debian and seem to work poorly other
>> places. Debian uses dak (debian archive kit) to manage repos. Ubuntu
>> uses something else. dak is big, poorly documented, and is not
>> currently being used in maemo, this means the current repo system
>> would have to be adapted to dak. Think new repos, new documentation,
>> new policy, etc.
>>
>> Adapting debian tools to manage the maemo repositories might be a
>> really good idea, but there are some significant technical hurdles.
>
> OTOH if we can support Debian (for example) in documenting dak and  
> fixing bugs
> in  it then we indirectly pull in more resource and provide support  
> to the
> community at large.

Absolutely. This is why I have created  http://wiki.debian.org/ 
DakHowTo and I have subscribed to the dak mailing list and contacted  
dak developers on IRC. :)  Any and all documentation of dak can easily  
be shared by debian and maemo if dak gets chosen to be used inside  
maemo.

>>
>> So far, I have forked debian's policy checker, lintian, into  
>> something
>> called maemian.
>
> I'm not sure this is the best approach :)
> (eg thinking about the maemo fork of gtk)
>
> In the short term I recognise that it is easier but I think we  
> should be *VERY*
> strong on working with upstream wherever possible - and if it's not  
> possible
> then I think we should consider an alternative solution that we can  
> work with.

The problem is that lintian is debian specific - it checks packages  
based on debian policy while maemo policy is different. For example,  
if lintian complains to a developer that they do not have a man page  
for their application, which is a requirement in debian, the developer  
will be pretty confused when they read the maemo policy saying that  
there should be no man pages. A fork must be done because a certain  
amount of the lintian code is useless.

> OTOH it's a positive step - thanks!
>
>> This will inherit all of the debian policy checks, but
>> exlcude those that do not apply to maemo, plus we can add some that  
>> do
>> apply to maemo. It is written in perl so anyone interested in  
>> writing/
>> learning perl is welcome to participate.
> Can we contact lintian's maintainer and see if they would support  
> pluggable
> policies?
> (Having never even run lintian I accept I don't know what I'm  
> talking about)

I am on the lintian mailing list and have been a long time  
collaborator with Russ Alberry, lintnian's current maintainer and one  
of the main debian policy gurus. Russ is part of the debian-perl team  
as am I, the team takes care of about 1,000 debian perl packages.  
lintian is in fact one such package, being largely written in perl.  
Since it is perl it can run in scratchbox and we can send patches up  
to debian to improve lintian, but as I mentioned earlier, I doubt they  
will be interested in our maemo policy specific patches. lintian is  
fully modular and is developed in a pluggable manner so maemo can  
easily extend it.

>
>> I think the best thing to do is to hash out policy, codify it in a
>> document. Adapt the policy checking tool to the agreed upon policies,
>> deploy them over the most experimental repos first.
> Hmmm - OK, but only because...
>
>> In parallel we
>> might deploy dak and a more debian-like repo structure along the  
>> lines
>> that Quim described.
> I think this should be pursued (but then I would wouldn't I)
>>
>> Now if there were only 30 hours in the day . . .
> You don't have coffee?

  I drink tea. :) Maybe that is my problem. :-/

Jeremiah