Re: pkg-perl TODO: SVN, archive, upstream, etc versions
Damyan Ivanov dijo [Thu, Jul 19, 2007 at 01:36:12PM +0300]:
> Hi all,
>
> The scripts/qa/versioncheck-html script compares all five possible
> versions of a package - upstream, SVN, NEW, incoming and the version in
> the archive and outputs a list of packages whose versions differ.
>
> The result is at http://pkg-perl.alioth.debian.org/qa/versions.html
_Very_ nice list. I'll be playing a bit (sooner or later... You know
how I am, I tend to leave things for later :-/ ) - I think adding a
link here and there can be a good extra :)
> This gives the following TO-DO for people with time :)
/me looks around and whistles before anybody notices him ;-)
> (1)
> For the entries with "CPAN contains no /regex/" in the "Upstream"
> column, the watch-file needs to be updated to contain the real location
> of the upstream source.
>
> Similarly with packages claiming to be higher than the upstream -
> perhaps debian/watch points to some old location.
AFAICT, we will have to learn to live with some spurious files here
and there... Maybe we should add something +- like overrides (probably
bound to a specific version) just so we don't clutter the display with
what we are already aware of.
> (...)
> The script is run daily and spends less than a minute on alioth.
Wow, that's amazing! And... Compared to the DEHS, it does not complain
as much about not being able to contact the servers.
> I guess after fixing (2), this more-or-less provides the functionality
> of the "pending/" branch we discussed. That is, packages needing an
> upload are those whose SVN and archive versions differ. Suggestions how
> to make this more visible on the page are welcome.
Yes, my only problem with this is the daily frequence - Of course, I
don't want to tie up Alioth checking upstream versions... But it's not
good for coordinating work anymore. It's not good to know if you and
me are doing the same work twice.
> CAVEATS
> The script still considers "0.2200" and "0.22" to be different.
Well, those version numbers are different as per Debian's standards,
but the same as per CPAN's. Maybe we should interpret them differently
according to each column? ;-)
> This all HTML thing is just some bike-shedding on my side over the work
> Gregor has done. Thanks, Gregor!
Yup, Gregor's scripts rock :) Thank you for the bike-shedding!
Greetings,
--
Gunnar Wolf - gwolf@gwolf.org - (+52-55)5623-0154 / 1451-2244
PGP key 1024D/8BB527AF 2001-10-23
Fingerprint: 0C79 D2D1 2C4E 9CE4 5973 F800 D80E F35A 8BB5 27AF
Reply to: