I'm looking at this bug as part of the BSP in Dublin, and I have a few points:
It seems that python-support does support the correct behaviour, and it's documented how to invoke it. It is the responsibility of the user of this package (i.e. the maintainer of a package which build-depends on python-support) to make sure the correct procedure is followed in postinst. It would be awesome if it would just DTRT every time out of the box, but I guess there are valid reasons for it not to do it: - I suppose auto-detecting that "update-python-modules -p" should be run is non-trivial - just running it every time would have major effect on performance, and is only really needed in narrow use cases (judging by the lack of complaints from other package maintainers) Thus it seems that the correct severity for this bug would be "wishlist" rather than "serious". Having said that, I must say I'm surprised that "python-support" is going away, as it seems to be used by ~800 packages. Is there a replacement? What is it? Now for a minor nitpick: the issue is actually documented in README.gz, not README.Debian -- Marcin Owsiany <porri...@debian.org> http://marcin.owsiany.pl/ GnuPG: 2048R/02F946FC 35E9 1344 9F77 5F43 13DD 6423 DBF4 80C6 02F9 46FC -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
