[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Mandos rebuild required for libavahi-core soname bump

Hi Mandos Maintainers,

I've uploaded a new avahi release a few days ago, which bumped the soname of
libavahi-core from 6 to 7.
This makes it necessary to recompile mandos against the new libavahi-core.

Usually, this is done via binNMUs and doesn't need any action from the maintainer.

Unfortunately, binNMUs for mango failed [1].
Apparently this is a bug in sbuild, which currently doesn't cope with comments
after the urgency field ( in your case (HIGH on mips and mipsel)).

Here is the relevant discussion from irc:
> <mbiebl> mehdi: hi
> <mbiebl> any idea what's going wrong here: https://buildd.debian.org/fetch.cgi?pkg=mandos&arch=amd64&ver=1.0.14-1%2Bb1&stamp=1278683999&file=log&as=raw
> <mbiebl> looks like the mandos binNMUs failes
> <mehdi> saw it
> <mehdi> I'll have to reschedule them
> <adsb> Do all of them end up with broken changelogs?
> <mehdi> yes
> <mbiebl> got the package name stripped of or what is the problem?
> <adsb> Yeah, it looks like wanna-build didn't create the changelog entry correctly, but the data looks ok afaics
> <adsb> aba: HE: Q_: Any ideas what might cause that? (the changelog entry generated by a binNMU just has " (version) unstable; urgency=low" by the look of the logs)
> <mbiebl> mehdi: thanks for you quick response, btw
> <jwilk> "mandos (1.0.14-1) unstable; urgency=low (HIGH on mips and mipsel)" - maybe that's the culprit.
> <mehdi> mbiebl, thank me when mandos will built :/
> <mehdi> s/t /d /
> <adsb> jwilk: Hrm, maybe
> <adsb> That comment is policy compliant, if rarely used, though
> <adsb> Yep, I think jwilk gets the cookie
> <adsb> The version of sbuild in the archive does:
> <adsb>             $firstline =~ /^(\S+)\s+\((\S+)\)\s+([^;]+)\s*;\s*urgency=(\S+)\s*$/;
> <adsb>             my ($name, $version, $dists, $urgent) = ($1, $2, $3, $4);
> <jcristau> bad sbuild
> <adsb> and then updates $version and $dists from other values and hard-codes urgency to low
> <adsb> So, yeah, sbuild fail :/
> <mehdi> but how the package built in the first place? (the first upload)
> <jcristau> since it's in perl it could just get Dpkg::Changelog::Parse to do that job, no?
> <adsb> because that code is in the binNMU changelog generation
> <adsb> and therefore wasn't run for the initial build
> <adsb> jcristau: Crazy idea :)
> <jcristau> i know..
> <mehdi> :/
> <mehdi> did someone fired up a reportbug already? :)
> <adsb> Not yet. I had checked the BTS index for sbuild though and not spotted anything obvious
> <jcristau> mbiebl: so i guess you get to nmu mandos instead
> <mbiebl> jcristau, adsb, mehdi: do you agree, that I should NMU?
> <mbiebl> do I need to bump the libavahi-core-dev b-dep or can you handle that via dep-waits, too?
> <mbiebl> I'd like to avoid to bump the b-dep in the NMU if possible
> <mehdi> nmu should be fine, ask the maintainer first… maybe theyre are responsive enough
> <mehdi> I can set dep-waits as well… but that should be coordinated
> <mehdi> better to wait for avahi to be installed everywhere, imo
> <jcristau> worst case you can binnmu the nmu
> <jcristau> since it won't trigger sbuild's brokenness
> <mehdi> where is the win?
> <mehdi> a few hours?
> <jcristau> i mean if the dep wait is set too late

So, I would appreciate if you could upload a new mandos revision as soon as
avahi 0.6.26-1 is successfully installed on all architectures to workaround this
bug in sbuild.
If you are busy I can NMU if you are ok with that.


[1] https://buildd.debian.org/status/package.php?p=mandos&suite=unstable
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

Attachment: signature.asc
Description: OpenPGP digital signature

Reply to: