Re: Let’s enter experimental
On Tue 16 October 2012 20:24:51 Joachim Breitner wrote: Am Dienstag, den 16.10.2012, 13:33 +1300 schrieb Michael Cree: HC [stage 2] utils/haddock/dist/build/Haddock/GhcUtils.o utils/haddock/src/Haddock/GhcUtils.hs:1:35: lexical error at character '\n' make[2]: *** [utils/haddock/dist/build/Haddock/GhcUtils.o] Error 1 Full log is at: http://buildd.debian-ports.org/status/fetch.php?pkg=ghcarch=alphaver=7. 6.1-2stamp=1349784016 That is certainly strange. The file in question is just fine and builds ok on all other arches. The thing to note seems to be that this is the first file built by the stage 2 compiler. I assume this means that it is broken... You can report a bug at upstream: http://hackage.haskell.org/trac/ghc Thanks, I've now done that; it's ticket #7339. I've also realised that ghc first FTBFS at version 7.4.2 on Alpha. I've got the unstable version (7.4.1-4) currently rebuilding to verify that it still builds successfully, and if it does then we know that the problem was introduced between 7.4.1 and 7.4.2. I also see a lot of haskell packages coming through into the experimental distribution Sorry that was a mistake by me: For some of them my dependency adjusting regex failed, and they built against 7.4.1 by accident (and by aptitude’s preference of unstable over experimental). Yesterday I have hopefull re-uploaded all affected packages. OK, I won't upload the built packages then on Alpha. Presumably when the re- uploaded packages come through all those built ones will revert to BD- uninstallable and can stay in that state until we get a working ghc in experimental. Cheers Michael. -- To UNSUBSCRIBE, email to debian-alpha-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201210172124.11257.mc...@orcon.net.nz
Re: Re: Let’s enter experimental
Hi, Am Dienstag, den 16.10.2012, 13:33 +1300 schrieb Michael Cree: HC [stage 2] utils/haddock/dist/build/Haddock/GhcUtils.o utils/haddock/src/Haddock/GhcUtils.hs:1:35: lexical error at character '\n' make[2]: *** [utils/haddock/dist/build/Haddock/GhcUtils.o] Error 1 Full log is at: http://buildd.debian-ports.org/status/fetch.php?pkg=ghcarch=alphaver=7.6.1-2stamp=1349784016 That is certainly strange. The file in question is just fine and builds ok on all other arches. The thing to note seems to be that this is the first file built by the stage 2 compiler. I assume this means that it is broken... You can report a bug at upstream: http://hackage.haskell.org/trac/ghc but they tend to care not a whole lot about odd architectures. Nevertheless maybe someone has seen the error before and can easily tell what’s wrong. I also see a lot of haskell packages coming through into the experimental distribution and on Alpha some have been built with ghc 7.4.1-4 from unstable. Would it be better if I did not upload the built packages and left them sitting in the built state until we get a working ghc in experimental so that they can built with ghc 7.6 and I don't end up having to spend lots of time scheduling binNMUs? Sorry that was a mistake by me: For some of them my dependency adjusting regex failed, and they built against 7.4.1 by accident (and by aptitude’s preference of unstable over experimental). Yesterday I have hopefull re-uploaded all affected packages. Greetings, Joachim -- Joachim nomeata Breitner Debian Developer nome...@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C JID: nome...@joachim-breitner.de | http://people.debian.org/~nomeata signature.asc Description: This is a digitally signed message part
Re: Re: Let’s enter experimental
well let’s try. I uploaded ghc-7.6 to experimental, along with an updated haskell-devscripts (with a ghc = 7.6 bound) and, as a test package, haskell-transformers (with a bound on haskell-devscripts). This should make the autobuilder build everything with only packages from experimental. I see ghc FTBFS on Alpha in experimental at Debian-Ports. I realise Alpha is not an official architecture so probably falls below your radar, but I wondered, if, nevertheless, you might be happy to take a look at the build log which ends with: echo 'executablename=/«PKGBUILDDIR»/inplace/lib/ghc- stage2' inplace/bin/ghc-stage2 echo 'datadir=/«PKGBUILDDIR»/inplace/lib'inplace/bin/ghc- stage2 echo 'bindir=/«PKGBUILDDIR»/inplace/bin' inplace/bin/ghc- stage2 echo 'topdir=/«PKGBUILDDIR»/inplace/lib' inplace/bin/ghc- stage2 echo 'pgmgcc=/usr/bin/gcc' inplace/bin/ghc-stage2 cat ghc/ghc.wrapperinplace/bin/ghc-stage2 chmod +x inplace/bin/ghc-stage2 HC [stage 2] utils/haddock/dist/build/Haddock/GhcUtils.o utils/haddock/src/Haddock/GhcUtils.hs:1:35: lexical error at character '\n' make[2]: *** [utils/haddock/dist/build/Haddock/GhcUtils.o] Error 1 Full log is at: http://buildd.debian-ports.org/status/fetch.php?pkg=ghcarch=alphaver=7.6.1-2stamp=1349784016 Is this likely to require an easy fix or not? (I have no idea since I don't know anything about haskell.) I also see a lot of haskell packages coming through into the experimental distribution and on Alpha some have been built with ghc 7.4.1-4 from unstable. Would it be better if I did not upload the built packages and left them sitting in the built state until we get a working ghc in experimental so that they can built with ghc 7.6 and I don't end up having to spend lots of time scheduling binNMUs? Cheers Michael. -- To UNSUBSCRIBE, email to debian-alpha-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8f4ef044-61c4-4f29-bbb1-78379a773...@orcon.net.nz
