On Thu, Dec 13, 2012 at 06:50:00PM +0100, Santiago Vila wrote: > As we don't support upgrades which skip releases (i.e. Debian 4.0 to > Debian 6.0, or Debian 5.0 to Debian 7.0), why don't we remove all > those obsolete dependencies on essential packages instead?
That makes sense! I investigated the situation on sid amd64 hoping that other architectures don't differ that much. bash depends on base-files (>= 2.1.12) bsd-mailx depends on base-files (>= 2.2.0) debian-edu-config depends on base-files (>= 5.3) dict-foldoc depends on base-files (>> 4) dict-vera depends on base-files (>= 4.0.0) dpkg-dev depends on base-files (>= 5.0.0) heirloom-mailx depends on base-files (>= 2.2.0) kup-server depends on base-files (>= 6.4) libgtk2-imageview-perl depends on base-files (>= 4.0.1) libpam-mount depends on base-files (>= 6.4) libpod-constants-perl depends on base-files (>= 4.0.1) liblog4cxx10-doc depends on base-files (>= 4.0.4) lib32nss-mdns depends on base-files (>= 3.1.10) libnss-mdns depends on base-files (>= 3.1.10) python-parsedatetime depends on base-files (>= 4.0.4) rsync depends on base-files (>= 4.0.1) speechd-el depends on base-files (>= 4.0.1) speechd-el-doc-cs depends on base-files (>= 4.0.1) trn4 depends on base-files (>= 2.2.0) vera depends on base-files (>= 4.0.0) weechat-scripts depends on base-files (>= 4.0.1) So what can we conclude from this list? 1) Most of these dependencies are not relevant anymore and indeed should be deleted. 2) There are still useful dependencies on base-files such as kup-server and libpam-mount. 3) Getting rid of the dependencies which are no longer supported will take time. I'd expect that we still have at least half of them in a year. So I think the request to implement Multi-Arch for base-files is useful. Should I file wishlist bugs to remove dependencies lower than squeeze? Helmut -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
