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

Re: Bug#616494: git-annex: FTBFS on various archs: undefined reference to symbol 'sem_close@@GLIBC_…'


Am Freitag, den 04.03.2011, 18:33 -0400 schrieb Joey Hess:
> Cyril Brulebois wrote:
> > your package no longer builds on various archs. Example on armel:
> > | make[1]: Entering directory `/build/buildd-git-annex_0.22-mipsel-Yh1YgB/git-annex-0.22'
> > | ghc -O2 -Wall -ignore-package monads-fd --make configure
> > | [1 of 2] Compiling TestConfig       ( TestConfig.hs, TestConfig.o )
> > | [2 of 2] Compiling Main             ( configure.hs, configure.o )
> > | Linking configure ...
> > | /usr/bin/ld: /usr/lib/ghc-6.12.1/unix- undefined reference to symbol 'sem_post@@GLIBC_2.2'
> > | /usr/bin/ld: note: 'sem_post@@GLIBC_2.2' is defined in DSO //lib/libpthread.so.0 so try adding it to the linker command line
> > | //lib/libpthread.so.0: could not read symbols: Invalid operation
> > | collect2: ld returned 1 exit status
> > | make[1]: *** [SysConfig.hs] Error 1
> This must be some sort of ghc problem on those arches. The configure
> program is some quite simple haskell code with no use of libraries
> outside the System haskell libraries. ghc already has one other open bug
> regarding problems with libpthread, that affected static compilation.

there was no change in the Haskell toolchain. Could this be related to
the linker changes mentioned in
http://lists.debian.org/4D6A7520.2090903@debian.org ?

This seems to affects building any binary, including the next version of
ghc (which causes bootstrapping issues with possible fixes), see

In that mail, Iain proposes a fix involving a porter upload. Maybe
someone more knowledgeable with linker issues than me can comment on
whether that is sane? I also would like to understand why this only
happens on some arches, but not on all.

Ah, this could be it, right:
so it should be in the just-released ghc-7.0.2, which I can upload to
experimental right away. Once the bootstrapping issue is worked around
and this package made its way to unstable, git-annex will build again...

Joey, if you need short term solutions, you might have to change linker
flags during your package build. But no idea which ones, or how to get
them in into the right spots.


Joachim "nomeata" Breitner
Debian Developer
  nomeata@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C
  JID: nomeata@joachim-breitner.de | http://people.debian.org/~nomeata

Attachment: signature.asc
Description: This is a digitally signed message part

Reply to: