Your message dated Sun, 17 Jun 2012 11:38:02 +0200
with message-id <87haua1dol....@brouzouf.villemot.name>
and subject line Re: Bug#605370: atlas: FTBFS on alpha, armel, ia64, mips,
powerpc and s390
has caused the Debian Bug report #605370,
regarding src:atlas: FTBFS on alpha, armel, ia64, mips, powerpc, s390
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact ow...@bugs.debian.org
immediately.)
--
605370: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=605370
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems
--- Begin Message ---
Package: src:atlas
Version: 3.8.3-29
Severity: serious
Justification: no longer builds from source
Hi!
Your package fails to build from source on the architecures alpha, armel,
ia64, mips, powerpc and s390:
https://buildd.debian.org/fetch.cgi?pkg=atlas;ver=3.8.3-29;arch=alpha;stamp=1290598328
https://buildd.debian.org/fetch.cgi?pkg=atlas;ver=3.8.3-29;arch=armel;stamp=1290616928
https://buildd.debian.org/fetch.cgi?pkg=atlas;ver=3.8.3-29;arch=s390;stamp=1290592573
https://buildd.debian.org/fetch.cgi?pkg=atlas;ver=3.8.3-29;arch=ia64;stamp=1291971874
https://buildd.debian.org/fetch.cgi?pkg=atlas;ver=3.8.3-29;arch=mips;stamp=1290604568
https://buildd.debian.org/fetch.cgi?pkg=atlas;ver=3.8.3-29;arch=powerpc;stamp=1290588116
I'm not really sure, if it is the same bug for all archs, but from what
I've seen, something is wrong in the CPU detection.
On powerpc I saw:
[..]
FlagCheck.c:1: error: bad value (ultrasparc) for -mcpu= switch
FlagCheck.c:1: error: bad value (ultrasparc) for -mtune= switch
FlagCheck.c:1: error: bad value (ultrasparc) for -mcpu= switch
FlagCheck.c:1: error: bad value (ultrasparc) for -mcpu= switch
FlagCheck.c:1: error: bad value (ultrasparc) for -mcpu= switch
[..]
cc1: error: unrecognized command line option "-mips4"
cc1: error: unrecognized command line option "-mips4"
cc1: error: unrecognized command line option "-mips4"
cc1: error: unrecognized command line option "-mips4"
[..]
On ia64, s390 and mips, we have something similar to:
make[2]: *** [atlas_run] Error 1
make[1]: *** [IRun_asm] Error 2
atlconf_misc.o: In function `CmndResults':
/build/buildd-atlas_3.8.3-29-ia64-1y64R9/atlas-3.8.3/build/atlas-base/../..//CONFIG/src/atlconf_misc.c:306:
warning: the use of `tmpnam' is dangerous, better use `mkstemp'
/build/buildd-atlas_3.8.3-29-ia64-1y64R9/atlas-3.8.3/build/atlas-base/../..//CONFIG/src/backend/cpuid.S:
Assembler messages:
/build/buildd-atlas_3.8.3-29-ia64-1y64R9/atlas-3.8.3/build/atlas-base/../..//CONFIG/src/backend/cpuid.S:30:
Error: Unknown opcode `subl $8,%esp'
/build/buildd-atlas_3.8.3-29-ia64-1y64R9/atlas-3.8.3/build/atlas-base/../..//CONFIG/src/backend/cpuid.S:31:
Error: bad expression
/build/buildd-atlas_3.8.3-29-ia64-1y64R9/atlas-3.8.3/build/atlas-base/../..//CONFIG/src/backend/cpuid.S:31:
Error: Illegal operand separator `e'
[..]
make[3]: *** [xarchinfo_x86] Error 1
[..]
Unable to find usable compiler for ICC; abortingMake sure compilers are in
your path, and specify good compilers to configure
(see INSTALL.txt or 'configure --help' for details)make[2]: *** [atlas_run]
Error 1
On armel:
[..]
res/sger1_1_100 : VARIATION EXCEEDS TOLERENCE, RERUN WITH HIGHER REPS.
ATL_ger1_axpy.c : 0.00
ATL_ger1_4x4_1.c : 0.00
ATL_ger1_1x4_0.c : 0.00
ATL_ger1_8x4_0.c : 0.00
ATL_ger1_SSE.c : 0.00
make[4]: *** [res/sR1RES] Error 255
make[4]: Leaving directory
`/build/buildd-atlas_3.8.3-29-armel-dHAMWZ/atlas-3.8.3/build/atlas-base/tune/blas/ger'
make[3]: ***
[/build/buildd-atlas_3.8.3-29-armel-dHAMWZ/atlas-3.8.3/build/atlas-base/tune/blas/ger/res/sR1RES]
Error 2
[..]
On alpha:
[..]
gcc -Wa,--noexecstack -fPIC -DL2SIZE=4194304
-I/build/buildd-atlas_3.8.3-29-alpha-NDNEIc/atlas-3.8.3/build/atlas-base/include
-I/build/buildd-atlas_3.8.3-29-alpha-NDNEIc/atlas-3.8.3/build/atlas-base/../..//include
-I/build/buildd-atlas_3.8.3-29-alpha-NDNEIc/atlas-3.8.3/build/atlas-base/../..//include/contrib
-DAdd_ -DF77_INTEGER=int -DStringSunStyle -DATL_OS_Linux -DWALL -DATL_NCPU=2
-x assembler-with-cpp -DDREAL -c dset.c
dset.c:80:5: error: #error "This kernel requires a gas x86 assembler!"
make[9]: *** [dsettest] Error 1
[..]
Note, that this also hinders #588418 to be fixed in squeeze.
Best regards,
Alexander
-- System Information:
Debian Release: 5.0.7
APT prefers stable
APT policy: (500, 'stable')
Architecture: i386 (i686)
Kernel: Linux 2.6.26-2-xen-686 (SMP w/1 CPU core)
Locale: LANG=de_DE@euro, LC_CTYPE=de_DE@euro (charmap=ISO-8859-15) (ignored:
LC_ALL set to de_DE@euro)
Shell: /bin/sh linked to /bin/bash
--- End Message ---
--- Begin Message ---
Version: 3.8.4-7
Closing this bug, since version 3.8.4-7 compiled on all archs mentioned
in the original report.
For the records, here is my understanding of what happened:
- the original cause of the FTBFS which triggered this bug report was
external to the package (dependency or compiler issue) and it has
disappeared since
- however, in version 3.8.4-1, a new issue was introduced: the number of
threads was no longer forced to 2, and this led to failures on some
arches where autodetection of threads was not working; this change was
reverted in 3.8.4-7, hence the build success
- additionally, there are still random FTBFS due to the very nature of
upstream build system. These typically occur when the buildd is under
heavy load. This is documented upstream:
http://math-atlas.sourceforge.net/errata.html#tol
A build failing because of this known problem will have this message
near the end:
VARIATION EXCEEDS TOLERENCE, RERUN WITH HIGHER REPS.
A give back generally solves the problem. There seems to be no easy
solution to this issue, since fixing it would require thorough changes
in the build system.
--
Sébastien Villemot
Researcher in Economics & Debian Maintainer
http://www.dynare.org/sebastien
Phone: +33-1-40-77-84-04 - GPG Key: 4096R/381A7594
pgpREcQmisitL.pgp
Description: PGP signature
--- End Message ---
