Sources required by Built-Using break multiple assumptions?
Hi!
I was mightily confused when I saw that supposedly dpkg 1.17.12, which
had an RC bug filed and had not spent enough time to transition had
“migrated” [M], w/o any sign of MIGRATION mail, nor any release team
hint directive.
This is what rmadison has to say about this:
,---
dpkg | 1.17.10 | jessie | source, amd64, armel, armhf, i386,
kfreebsd-amd64, kfreebsd-i386, mips, mipsel,
powerpc, s390x
dpkg | 1.17.13 | sid | source, amd64, arm64, armel, armhf, hurd-i386,
i386, kfreebsd-amd64, kfreebsd-i386, mips,
mipsel, powerpc, ppc64el, s390x, sparc
`---
The Sources for jessie do contain both versions, 1.17.10 and 1.17.12
with an Extra-Source-Only:yes field, due to at least deets and the
recently introduced Built-Using in debsig-verify.
The following are probably issues with the QA codebase, it lists as
migrated when that's not true:
(thinks that testing is 1.17.12)
<https://qa.debian.org/developer.php?login=guillem> [M]
<https://packages.qa.debian.org/d/dpkg.html>
<https://packages.debian.org/search?searchon=sourcenames&keywords=dpkg>
(thinks that source is 1.17.12 and binaries 1.17.10, half-true I guess)
<https://packages.debian.org/jessie/dpkg> (1.17.10)
<https://packages.debian.org/source/jessie/dpkg> (1.17.12)
This is part of the release team, which looks wrong:
(thinks it's migrating from 1.17.12 to 1.17.13)
<https://release.debian.org/migration/testing.pl?package=dpkg>
But I've not checked if there are other similar assumptions in other
places, for example britney.
If this analysis seems right I can probably file some bug reports, and
I guess the way to fix this would be to ignore any source that is marked
Extra-Source-Only:yes? Or maybe exposing such packages in the same
Sources file breaks too many assumptions and might need to be moved to
a different Sources file? dunno.
Thanks,
Guillem
Reply to: