On Tue, 18 Jun 2013 23:10:32 +0200, Markus Koschany wrote: > >> I am looking for a sponsor for my package mockito which i intend to > >> adopt. It builds also fine with its reverse dependencies. > > Uploaded, from/to the pkg-java git repo. > Thank you very much for the quick upload and your review! You're welcome! > > - In my experience the git tag is set by the person / at the time of > > the upload to make sure it matches what's in the archive > > (and to avoid git troubles). > Sounds good to me. I have been in the lucky position that my previous > sponsors let me tag new releases as soon as we both had agreed about the > final state of the package. I think both ways are fine but i haven't > experienced any git troubles so far. If you or I would have changed something after the tag we would have needed to delete it and re-create it and then overone who has already pulled is unhappy -- at least that's my maybe wrong understanding :) > > - debhelper (>= 7.0.50~): > > I'd go for >= 8, since even oldstable has 8 (and compat level 8 in > > d/compat) > > [7.0.50 was the first one to introduce support for "dh $@ --with > > FOO] > Ok, normally i would even go for debhelper 9 as i did for all other > packages except mediathekview. It appears to me that java packages don't > benefit as much from the latest debhelper version as c or c++ packages > and some people have claimed on mentors that a lower compat level is > then more beneficial for backports. I have no problems to bump the > compat level to 9 with the next upload. Ack, I usually also stick to 8 for (perl or java) arch:all packages to keep them backportable and because there's no real improvement. Just 7.0.50 seemed a bit outdated :) > > - Removing the Forwarded: header from the patches is not a good idea > > IMO; because the next one looking at the patches doesn't know if > > they are forwarded upstream or not. And bonus points for actually > > forwarding them :) (or marking them as "Forwarded: not-needed" if > > they are Debian-specific). > Agreed. I blame it on git-buildpackage (because it can't fight back) > because it automatically removes descriptions and optional DEP-3 headers > and you need to be very careful to preserve old patch descriptions if > you simply run "gbp-pq import". Ah, gbp-pq. I see. > Otherwise i believe the Debian patches > are not upstreamable but i will add "Forwarded: not-needed" again. Ok, thanks. > > Also the Subjects were nicer in dapal's original version of the > > patches. (Slightly ...) > Git-buildpackage again. Subject and patch name have to be identical if > you use gbp-pq import / gbp-pq export but David's patch subject and > patch name differed in the past. I will add a short description to > compensate for that. Thanks! > > - You might also want to add > > echo "compression = xz" > debian/source/options > Ok, but we need xz compression as default...now. :) Indeed :) Cheers, gregor -- .''`. Homepage: http://info.comodo.priv.at/ - OpenPGP key 0xBB3A68018649AA06 : :' : Debian GNU/Linux user, admin, and developer - http://www.debian.org/ `. `' Member of VIBE!AT & SPI, fellow of the Free Software Foundation Europe `- NP: Willi Resetarits & Stubnblues: De Dornen bleibm
Attachment:
signature.asc
Description: Digital signature