Am 25.04.2013 14:26, schrieb Julien Cristau: > Thoughts in random order... > 1) this is terrible
A dangling /usr/bin/python symlink is indeed bad. At least that convinced me to never use anything pythonish in the maintainer scripts in the future. > 2) in the absence of loops, dpkg ought to unpack python2.7-minimal > before python-minimal I'm not sure what's going on here, but in the log [1] from anbe python2.7-minimal is not unpacked *at all*. > 3) adding pre-depends at this stage makes me extremely wary > 4) rewriting gconf-schemas only helps if gconf-schemas is upgraded > before its reverse deps All packages shipping gconf schemas have a Depends on gconf2. It is generated via dh_gconf which adds "gconf2 (>= 2.28.1-2)" to misc:Depends. This version is already satisfied by gconf2 from squeeze. We could bump that and binNMU the rdeps. But then we'd have some arch:all rdeps (which would need sourceful uploads) and I'm actually not sure if a more tightened Depends actually ensures that the new gconf2 is unpacked before its rdeps. Does anyone know? > 5) do we know why this only shows up now, and if there are real world > cases where this issue occurs? I've never seen this before. Maybe anbe can tell us a bit more about the test setup he used. Michael [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=5;filename=science-typesetting_1.0.log.gz;att=1;bug=706110 -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
signature.asc
Description: OpenPGP digital signature
