2015-06-24 17:11 GMT+02:00 Bernhard Rosenkraenzer <[email protected]>:

> On 2015-06-24 16:07, Tomasz Gajc wrote:
>
>  Per what are the real benefits of moving back to rpm.org?
>> Does this needs some extra work on perl-URPM, urpmi, mock-urpm and
>> various ABF subsystems ?
>>
> Not really, for perl-URPM, it would mainly be a matter of just replace use
of rpm5 API with rpm4 API, which perl-URPM used by mageia uses already..
For this one I'm not likely to  just pick up mageia's rpm.org compatible
fork and port our changes to, but rather just port to the different API.
Considering that being the "biges"t job, while already done by Mageia, just
peeking Mageia's should be sufficient and enough toi to pick up...

For compatibility otherwise, I intend to address it in advance to prevent
issues from arrising, with work on rpm.org kept out of usage for quite a
while all untill ready . :)

>
> I think there's both benefits (esp. that it's the same thing opensuse and
> the likes use, so zypper will automatically get all needed changes) and
> drawbacks...
>
> We rely on stuff that was added in rpm5 in quite a few spec files -- I
> haven't looked at rpm.org in a long time, so maybe they have been added
> there by now. (Either way they should be fairly easy to add if we decide to
> go down that route). What's the status of:
>
> Filetype triggers
> %bcond_with and friends
> %apply_patches
> Suggests: handling
> %track
> pkgconfig(*) and cmake(*) dependencies etc.
> noarch subpackages in arch specific packages
>
> in current rpm.org? (And does anyone else remember other rpm5-isms that
> may be important for us?)

>From the guy who's working pretty much exclusively alone on it, done a poor
job of announcing & documenting new features, behaviour & enhancements,
there is actually considerably more.

Just look at the over 300 locally managed patches in our package repo since
November 2011.
Then there's work of mine from involvement in rpm5 dating back to ~2008
IIRC, with activity increased quite a bit in 2009, then considerably in
increased to full time work on in 2010 & 2011 that's part of upstream
before commit access were cut off..

I took a closer (but quite brief look at rpm.org repo, now hosted at github
btw.)  yesterday (made my first & only commit so far in my repo that felt
easier and safiest just to start by applying to begin with;
scripts/pythoneggs.py :), much of the stuff that either myself or others
implemented and  relied on by us seems to have been caught up with in some
areas and years later, so a lot won't be needed to be ported, either  due
to already having the exact same in place now, or some equivalent, which
could be worse, for which I'd  probably would prefer to implement our
replacement/alternative implementation, or in the case of better
implemented, making the obvious choice of adapt to the best candidate. :)

I have a lot of work related to this projet to catch up with on so many
levels, so work on this item on my TODO item isn't ranked the highest, and
there is no plans on preparing it for next OMV LX release... ;)

--
Regards,
Per Øyvind
_______________________________________________
OM-Cooker mailing list
[email protected]
http://ml.openmandriva.org/listinfo.cgi/om-cooker-openmandriva.org

Reply via email to