Hi Yaroslav, sorry for the delay in the response Il 18/01/2012 03:25, Yaroslav Halchenko ha scritto: > ok -- few more points > > * is it really necessary to depend on +dfsg'ed version of sphinx? i.e. > > python-sphinx (>= 1.0.7+dfsg) > > unless strictly needed, I would appreciate if it becomes > > python-sphinx (>= 1.0.7~) > > asking because in NeuroDebian we still carry backport of 1.0.7-2 ;)
lintian suggestion > > * upstream provides a mighty unittest battery. minimalistic/default > version ran in 280 sec on my box, so imho acceptable and desired to be > enabled to be ran during package build time. I have ran it simply > through > > PYTHONPATH=$PWD/.. python -c 'import tables; tables.test() > > from build directory after building extensions in place. Ideally it > should imho be ran against version "installed" under debian/. > I enable testing for all of my packages where there are tests > batteries and it has been proven indispensable especially for reliable > backporting > good idea I'm going to do it ASAP Not clear how to perform tests on the installed package during the build process, Can you please point me to a package to use as guidance? > * debhelper 8 -- may be I have missed some news (I am somewhat > non-compliant myself and using 7 for the sake of easy backports for > elderly releases), but why (>= 8.1.0) and not just (>= 8) (squeeze > carries 8.0.0) It is still related to sphinx stuffs. 8.1 should have a better handling of .inv files and compression IMHO we could also use 8.0 wit a couple of manual fix in the debian/rules > > Cheers, > > On Tue, 17 Jan 2012, Yaroslav Halchenko wrote: > >> first of all -- thanks for all the work!!! > >> My quick take would be: > >> * I would have added debian/README.source to describe that .orig >> fetched from GIT tarball directly so would not match upstream's >> source distribution tarball > OK, Also in this case it is related to shpinx issues. dh_sphinxxdoc replaces .js files with links to the system ones. I have a build error if the version of .js files bundled in the tarball is not the same of the system ones. >> * > >> -Maintainer: Wen Heping <[email protected]> >> +Maintainer: Debian Science Maintainers >> <[email protected]> >> +Uploaders: Wen Heping <[email protected]>, Antonio Valentino >> <[email protected]> >> +DM-Upload-Allowed: yes > >> I think > >> - premature for DM-Upload-Allowed (nothing personal but at least I >> would not sponsor upload then, since I am not very familiar with >> Antonio's work (yet)) >> - I would drop Wen altogether > perfectly fine with me. I set the DM-Upload-Allowed in order to comply with the debian-science policy: http://debian-science.alioth.debian.org/debian-science-policy.html#idp17887120 anyway, at the moment, I can't upload packages since I'm only a sponsored maintainer. >> * building it now so might whine a bit more later on ;-) > >> Thanks once again! > >> On Tue, 17 Jan 2012, Yaroslav Halchenko wrote: > > >>> Hi guys -- sorry, this email slipped through my eyes ;) > >>>> PS: Yaroslave can you read all the thread 641485 and gives you opinion. > >>> if you meant > >>> - should we get a new version -- SURE >>> - should we wait longer for original maintainer to reply -- NO >>> his uploads were sponsored and he was inactive for way to long to >>> indicate that he cares at all >>> - maintaining under debian-science team umbrella -- SURE > great, thank you >>> on my side -- I will have a look at package either it is easily >>> backportable A final question. Do you prefer me to upload a new package to mentors.debian.net or it is OK for you to grab updates from https://github.com/avalentino/PyTables/tree/debian best regards -- Antonio Valentino -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
