Processed: Re: [Debian-med-packaging] Bug#741288: seqan: FTBFS on many buildds due to RAM exhaustion
Processing control commands: tags -1 + patch Bug #741288 [src:seqan] seqan: FTBFS on many buildds due to RAM exhaustion Added tag(s) patch. -- 741288: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741288 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#741288: [Debian-med-packaging] Bug#741288: seqan: FTBFS on many buildds due to RAM exhaustion
Control: tags -1 + patch Hello, I was able to avoid this error and successfully build seqan on both mips and mipsel, adding flags -O0 -mxgot. -O0 was added to solve issue: virtual memory exhausted: Cannot allocate memory during compilation of pair_align.cpp. -mxgot was added to lift the usual restrictions on the size of the global offset table on MIPS, and to avoid an error: relocation truncated to fit: R_MIPS_GOT16 during linking of pair_align. Patch that contains these changes is attached. This patch affects only mips and mipsel architectures. Could you please consider to include it? Best Regards, Dejandiff -uNr seqan-1.4.1.orig/debian/rules seqan-1.4.1/debian/rules --- seqan-1.4.1.orig/debian/rules 2014-01-29 01:57:47.0 + +++ seqan-1.4.1/debian/rules 2014-07-30 14:09:36.0 + @@ -8,6 +8,11 @@ DEB_HOST_ARCH ?= $(shell dpkg-architecture -qDEB_HOST_ARCH) +DEB_BUILD_ARCH ?= $(shell dpkg-architecture -qDEB_BUILD_ARCH) +ifneq (,$(filter $(DEB_BUILD_ARCH),mips mipsel)) +export CXXFLAGS:=$(shell dpkg-buildflags --get CXXFLAGS | sed 's/-O[1-9]//') -O0 -mxgot +endif + %: dh $@
Bug#741288: [Debian-med-packaging] Bug#741288: seqan: FTBFS on many buildds due to RAM exhaustion
Hi Charles, On Tue, Mar 11, 2014 at 07:38:51AM +0900, Charles Plessy wrote: Le Mon, Mar 10, 2014 at 10:31:47PM +0100, Andreas Tille a écrit : It seems reducing optimisation in d/rules [...] did not really help (perhaps even droping -O1 could be tried but I doubt this). Hello everybody, I would also recommend to not make the program slower by design, unless you want public researchers to buy more computers with your tax money. This is not the case. The default optimisation is done for architectures capable to build currently. It is only decreased for those that don't. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#741288: seqan: FTBFS on many buildds due to RAM exhaustion
Source: seqan Version: 1.4.1-3 Severity: serious Justification: FTBFS Hi, it seems your package is eating too much RAM for breakfast: | cd /«PKGBUILDDIR»/obj-arm-linux-gnueabihf/core/apps/pair_align /usr/bin/c++ -DSEQAN_ENABLE_TESTING=0 -DSEQAN_HAS_EXECINFO=1 -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -g -O2 --param=ssp-buffer-size=4 -Wformat -Werror=format-security -D_FORTIFY_SOURCE=2 -W -Wall -Wno-long-long -fstrict-aliasing -Wstrict-aliasing -pedantic -Wno-variadic-macros -I/«PKGBUILDDIR»/core/include-o CMakeFiles/pair_align.dir/pair_align.cpp.o -c /«PKGBUILDDIR»/core/apps/pair_align/pair_align.cpp | virtual memory exhausted: Cannot allocate memory | make[4]: *** [core/apps/pair_align/CMakeFiles/pair_align.dir/pair_align.cpp.o] Error 1 Only amd64, powerpc, and s390x builds are in the archive by now: https://buildd.debian.org/status/package.php?p=seqansuite=sid (FWIW, bowtie is missing builds accordingly: https://buildd.debian.org/status/package.php?p=bowtiesuite=sid ) Mraw, KiBi. -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#741288: seqan: FTBFS on many buildds due to RAM exhaustion
Hi KiBi, On Mon, Mar 10, 2014 at 09:06:51PM +0100, Cyril Brulebois wrote: Source: seqan Version: 1.4.1-3 Severity: serious Justification: FTBFS it seems your package is eating too much RAM for breakfast: | cd /«PKGBUILDDIR»/obj-arm-linux-gnueabihf/core/apps/pair_align /usr/bin/c++ -DSEQAN_ENABLE_TESTING=0 -DSEQAN_HAS_EXECINFO=1 -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -g -O2 --param=ssp-buffer-size=4 -Wformat -Werror=format-security -D_FORTIFY_SOURCE=2 -W -Wall -Wno-long-long -fstrict-aliasing -Wstrict-aliasing -pedantic -Wno-variadic-macros -I/«PKGBUILDDIR»/core/include-o CMakeFiles/pair_align.dir/pair_align.cpp.o -c /«PKGBUILDDIR»/core/apps/pair_align/pair_align.cpp | virtual memory exhausted: Cannot allocate memory | make[4]: *** [core/apps/pair_align/CMakeFiles/pair_align.dir/pair_align.cpp.o] Error 1 Only amd64, powerpc, and s390x builds are in the archive by now: https://buildd.debian.org/status/package.php?p=seqansuite=sid This is a known fact admitted by upstream. I even failed to build on not so powerfully RAM equiped amd64 machines. The only solution I see is to exclude all other architectures than the one you are mentioning above. It seems reducing optimisation in d/rules DEB_HOST_ARCH ?= $(shell dpkg-architecture -qDEB_HOST_ARCH) DEB_BUILD_ARCH_BITS := $(shell dpkg-architecture -qDEB_BUILD_ARCH_BITS) ifeq ($(DEB_BUILD_ARCH_BITS),32) CFLAGS=$(shell dpkg-buildflags --get CFLAGS | sed 's/-O[2-9]/-O1/') CXXFLAGS=$(shell dpkg-buildflags --get CXXFLAGS | sed 's/-O[2-9]/-O1/') endif did not really help (perhaps even droping -O1 could be tried but I doubt this). Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#741288: [Debian-med-packaging] Bug#741288: seqan: FTBFS on many buildds due to RAM exhaustion
Le Mon, Mar 10, 2014 at 10:31:47PM +0100, Andreas Tille a écrit : It seems reducing optimisation in d/rules [...] did not really help (perhaps even droping -O1 could be tried but I doubt this). Hello everybody, I would also recommend to not make the program slower by design, unless you want public researchers to buy more computers with your tax money. Have a nice day, Charles -- Charles Plessy Debian Med packaging team, http://www.debian.org/devel/debian-med Tsurumi, Kanagawa, Japan -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
