Re: [Monotone-debian] state of the buildds

[Ludovic Brenta]

Zack, thanks for taking care of this.

Zack Weinberg writes:
> The newly uploaded monotone 0.37 is on its way to failing to autobuild
> on lots of architectures again. :-( The good news is, I don't think
> the problems are as fundamental this time.
>
> Currently, http://buildd.debian.org/pkg.cgi?pkg=monotone says (note that
> the overview screen lies - you have to look at each individual build log
> to know what happened :-P)
>
>  * amd64, i386: successfully built and transmitted to the archive.
>  * arm, m68k, powerpc, s390, and possibly hppa: build in progress.

The build on powerpc failed:

g++ -DLOCALEDIR=\"/usr/share/locale\" -DHAVE_CONFIG_H -I. -I. -I.  \
    -I./lua -I./pcre      -g -Wall -O2 -Wall -W -Wno-unused \
    -fno-strict-aliasing -c -o mtn-netsync.o `test -f 'netsync.cc' \
    || echo './'`netsync.cc
make[2]: *** [all-recursive] Terminated
make[1]: *** [all] Terminated
make[3]: *** [mtn-database.o] Terminated
make[3]: *** [mtn-netsync.o] Terminated
make: *** [debian/stamp-makefile-build] Terminated
Build killed with signal 15 after 150 minutes of inactivity

Overall the build attempt took 11:07:58.  I don't know why.  I will
try to build manually on Xavier's POWER5, see what happens, and upload
if all is fine.

>  * sparc, ia64: dependency problems; looks like fallout from the
>    Ghostscript packaging changes in unstable, buildd admin needs to
>    fix it, nothing we can do

Yes, but we will probably need to re-upload after they fix the problem.

>  * mips, mipsel: testsuite tripped over a quirk of the buildd
>    environment.  These apparently set $HOME to a directory that exists
>    but is completely inaccessible by the buildd uid; I've already
>    committed a fix for this to mainline and .debian-diff.  (To avoid
>    the 0.36-2 glitch, though, I have *not* put anything in
>    debian/changelog for it.  I think we'll want to do a 0.37-2 upload
>    in fairly short order but not just yet.)

Thanks.  I'll wait for hppa to attempt the build and upload 0.37-2
after that.

>  * alpha: genuine failure.  Doesn't appear to be the same problem that
>    we had last time round.  I've gotten back in touch with Lennart
>    Sorenson, who was debugging the alpha failures last time, but this
>    would be much faster to fix if some kind soul would give me an
>    account on an alpha machine.  ;-)

-- 
Ludovic Brenta.