Am Dienstag, den 14.10.2008, 13:53 -0500 schrieb Dirk Eddelbuettel:
On 14 October 2008 at 12:23, Gary Hennigan wrote:
| A lot of Debian lib*-dev packages also include static libraries. When
| we distribute our application we'll often include statically linked
| binaries so that users don't
Hi Sylvestre,
thanks for taking care of this! But I still have some doubts.
Am Mittwoch, den 15.10.2008, 00:43 +0200 schrieb Sylvestre Ledru:
Here is the full comment from the Makefile:
# This library is linked against various MCA components because all
# shared-memory based components
Hi Jeff,
thanks for taking the time to explain the issue!
Am Mittwoch, den 15.10.2008, 16:13 -0400 schrieb Jeff Squyres:
Static libraries are definitely a Good Thing in some scenarios. We
have a few features in this arena, which we consider separately:
- building libmpi (and friends) as
Hi Sylvestre,
thanks for working on this!
Am Montag, den 20.10.2008, 00:47 +0200 schrieb Sylvestre Ledru:
It is done in the debian/rules
The procedure is now:
# do the configure twice (with and without the option)
# launch the build for each configure (vpath)
# for the normal build, make
Hi Sylvestre!
Am Mittwoch, den 27.08.2008, 23:49 +0200 schrieb Sylvestre Ledru:
If you need help with this package, I am quite familiar with Java
packaging into Debian and we could do it into the debian-science team if
you want.
Thanks for your offer! I think putting it in the Debian Science
Am Dienstag, den 05.08.2008, 17:46 +0200 schrieb Manuel Prinz:
Hi Andreas!
^^^
I meant: Hi Alexander!
It's definitely too hot in the office... Sorry! :/
Best regards
Manuel
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil
Hi Andreas!
Thanks for your bug report!
Am Dienstag, den 05.08.2008, 17:25 +0200 schrieb Alexander Achenbach:
What about adding a symlink
/usr/include/openmpi = ../lib/openmpi/include
similar to the existing symlinks of shared libraries?
I was sure that symlink should exist until I
Hi Alexander!
Am Dienstag, den 05.08.2008, 17:43 +0200 schrieb Alexander Achenbach:
The openmpi-bin package adds an alternative 'mpiexec' that controls
/usr/bin/mpiexec and its manpage.
However, another MPI implementation (lam-runtime) adds a single
alternative class 'mpirun' that includes
Package: sysvconfig
Followup-For: Bug #486058
Hi,
the attached patch should fix the issue.
Best regards
Manuel
-- System Information:
Debian Release: lenny/sid
APT prefers testing
APT policy: (500, 'testing'), (200, 'unstable')
Architecture: amd64 (x86_64)
Kernel: Linux 2.6.25-2-amd64
Hi Michael,
thanks for your interest and sponsoring offer! I packaged PDB2PQR with
the help of Steffen Möller (CC'ed). The package is ready but was
rejected due to license issues. Upstream promised to take care of that
but due to holidays this will take some time, so it will not make it
into
Am Freitag, den 08.02.2008, 10:52 -0600 schrieb Dirk Eddelbuettel:
We could do this manually by editing but a better way is to
a) file a bug with libibverbs so that it provides the correct
restriction
b) maybe try building our package with a local shlibs in debian/
Package: libibverbs
Severity: important
Dear Roland,
please add version information to libibverbs so Build-Depend'ing libraries get
a versioned dependancy to libibverbs via {shlibs:Depends}. (See #464705.)
Best regards
Manuel
-- System Information:
Debian Release: lenny/sid
APT prefers
Am Freitag, den 08.02.2008, 10:52 -0600 schrieb Dirk Eddelbuettel:
a) file a bug with libibverbs so that it provides the correct
restriction
Submitted as #465435.
Best regards
Manuel
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil
reopen 464795 =
thanks
[ Reopening the bug with Ondrej as submitter, so we can close it in the
changelog and Ondrej gets notified when the fix is in sid. Ondrej, hope
that's OK with you! ]
Sorry for jumping in a little late, had no access to my mails
yesterday...
Am Mittwoch, den 13.02.2008,
Am Donnerstag, den 14.02.2008, 07:17 -0600 schrieb Dirk Eddelbuettel:
We may need an explicit removal in .preinst for it, just in case.
Which link was it? Which package?
/etc/alternatives/libopenmpi.so from libopenmpi-dev still pointing to
*.so.0. After purging and installing again, the
Am Donnerstag, den 14.02.2008, 20:52 -0600 schrieb Dirk Eddelbuettel:
... but the good news is that my upgrade was clean:
Good to know. Thanks for the upload!
Question for Manuel: should we also symlink the fortran libs given how we do
it for the c/c++ ones?
No, we shouldn't. They don't
Am Freitag, den 01.02.2008, 08:21 +0100 schrieb Yves-Alexis Perez:
Without version I can't really say. But I guess you didn't try with
evolution-data-server 1.12.3-1:
Pawel wrote to me that it works for him now, didn't notice that it was a
private mail. Updating everything to 1.12.3-1 works.
Am Dienstag, den 05.02.2008, 11:19 +0100 schrieb Clemens Krammer:
$ dpkg -l | grep evolution
ii evolution 2.12.3-1 groupware suite with mail
client and organizer
ii evolution-common 2.12.3-1 architecture independent
files for Evolution
ii
Am Freitag, den 08.02.2008, 16:26 +0100 schrieb Joachim Reichel:
What happens if you update to 1.1.1-1 as the package wants you to?
Then it works (see the second ldd call). But libopenmpi1 does not want
me to update libibverbs1 because there is no versioned dependency.
Seems like the
Am Freitag, den 08.02.2008, 10:52 -0600 schrieb Dirk Eddelbuettel:
Could you possibly do a) and b), pretty pretty please? I just built a hot
new R 2.6.2 that was released this am.
Will do so. Maybe tonight, but tomorrow latest.
Best regards
Manuel
signature.asc
Description: Dies ist ein
Hi all!
Am Freitag, den 04.01.2008, 13:52 +0100 schrieb Ondrej Certik:
On Jan 4, 2008 1:40 PM, Dirk Eddelbuettel [EMAIL PROTECTED] wrote:
Ondrej Certik wrote:
| In the last bug I reported:
|
| http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=457088
|
| I was told, that the mpi.h is
Hi Daniel!
First of all, thanks a lot for your patch to #220044!
Am Freitag, den 04.01.2008, 15:18 +0100 schrieb Daniel Leidert:
FYI: The bug in u-a has been fixed recently.
http://bugs.debian.org/220044
I did not get the message about the bug being closed though subscribed,
so thanks for the
Am Freitag, den 04.01.2008, 16:49 +0100 schrieb Ondrej Certik:
Actually, the mpi.h is in /usr/include/mpi, not /usr/include/openmpi,
as I mistakenly thought.
That happens. Are you fine with me closing the bug?
Best regards
Manuel
signature.asc
Description: Dies ist ein digital signierter
Package: slurm-llnl
Version: 1.2.20-1
Severity: important
Tags: patch
Hi Gennaro,
I have noticed an issue with preinst failing if the user slurm does not
exist on the system while upgrading.
Usually, this should not happen at all because the user is created when the
package is installed. In our
tag 459642 + pending
thanks
Hello Adam!
Am Montag, den 07.01.2008, 15:14 -0500 schrieb Adam C Powell IV:
The attached patch adds an alternavites symlink linking libmpi++.so to
OpenMPI's libmpi_cxx.so.0, matching the C++ library links in mpich and
lam. It works in the one case I've tried.
block 452047 by 388313 220044
thanks
Hi,
this bug (as well as #220044) blocks us (Debian OpenMPI Maintainers) in
fixing RC bug #452047. We have a solution in our trunk but it's not
functional because the alternative links are not removed/deleted
correctly by u-a.
I'd like to help in providing a
Am Dienstag, den 11.12.2007, 10:00 -0600 schrieb Dirk Eddelbuettel:
Maybe we should release a new openmpi to fix the few trivial bugs, and maybe
add a NEWS or README item indicating this open issue with the alternatives --
and how our hands are tied by update-alternatives -- to give this some
Am Dienstag, den 11.12.2007, 10:57 -0600 schrieb Dirk Eddelbuettel:
foo:~ dict -P- TTBOMK
No definitions found for TTBOMK
What's TTBOMK ?
To The Best Of My Knowledge. I'm kinda surprised dict doesn't know
that!?
Can you give it a spin against SVN? If all is well, I can
upload this eve.
Hi Ondrej!
Am Mittwoch, den 12.12.2007, 12:50 +0100 schrieb Ondrej Certik:
libmesh depends on openmpi, it build in pbuilder, however it fails on
buildbots:
http://buildd.debian.org/fetch.cgi?pkg=libmesh;ver=0.6.1.dfsg-1;arch=amd64;stamp=1197413726
I suspect the problem is with the mpicc
Hi Ondrej!
Am Mittwoch, den 12.12.2007, 14:04 +0100 schrieb Ondrej Certik:
Unfortunately, that's not the case, see my closed bug report about that:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=450518
I remember your bug report. At that time, we didn't know about that this
affects
Am Mittwoch, den 12.12.2007, 09:26 -0600 schrieb Dirk Eddelbuettel:
| However, openmi is providing an cxx alternative even when C++ compiler
| is not installed. Is it the right way ?
... even though one could argue that openmpi, if it provides wrappers for gcc
and g++, could also depend on
Am Montag, den 17.12.2007, 14:47 -0600 schrieb Dirk Eddelbuettel:
On 17 December 2007 at 21:13, Manuel Prinz wrote:
| Am Montag, den 17.12.2007, 13:36 -0600 schrieb Dirk Eddelbuettel:
| The reasoning behind that was to fix the breaking of other MPI
| implementations by moving stuff to /usr/lib
Am Montag, den 17.12.2007, 16:16 -0500 schrieb Adam C Powell IV:
As the maintainer of mpich, I do not see any conflicts here.
libmpich1.0ldbl has: libmpich.so.1.0, libfmpich.so.1.0,
libpmpich.so.1.0, libpmpich++.so.1.0, libtvmpich.so.1.0, and
libmpe.so.1.0 . There's no ABI compatibility
Hi Dirk!
Am Montag, den 17.12.2007, 16:24 -0600 schrieb Dirk Eddelbuettel:
Are you sure we need alternatives for something like libmpi_cxx.so.0 which
the 'other' (ie LAM) doesn't have?
No. What I currently try to figure out is where the intersection is and
create links all unique libs and
Hi guys!
Am Montag, den 17.12.2007, 17:53 -0500 schrieb Adam C Powell IV:
That already happens via alternatives slaves. As discussed earlier,
it's inappropriate with ABI-incompatible soname-named files e.g. *.so.0
I think we're going in the right direction: alternatives for *.so and
Hello Adam!
Am Dienstag, den 18.12.2007, 08:50 -0500 schrieb Adam C Powell IV:
A couple of notes:
* The lib*.so.0.0.0 and lib*.so.0 files *must* be in libopenmpi1,
that's the shared lib package which other packages will link to
at runtime. So please move those files and
Dear Ian!
Ian Jackson wrote:
Manuel Prinz writes (Re: Can anybody *please* fix #220044 - broken slave
files (link)):
I spent some time on trying to fix it myself but failed because I lacked
of time to dive into u-a. I'd like to do some documentation or even
several changes to make the u
Hi everyone!
Am Dienstag, den 18.12.2007, 15:31 +0100 schrieb Manuel Prinz:
I already noticed my mistake and am working with a modified version.
Here's my new and modified patch for openmpi. It looks right to me and
first checks show that it's working. I'll have a larger test tomorrow
is
whether we should compile the static libraries and/or (also) include
the .la files. I have to do more reading on that one.
On 19 December 2007 at 00:43, Manuel Prinz wrote:
| If noone has complaints, I will apply it to trunk.
Always apply, we can always fix later. No point in sending patches
Am Dienstag, den 18.12.2007, 16:39 -0800 schrieb Nicholas Breen:
Is it somewhere publically available? I'd be happy to test it as well,
it'll be interesting to see if it also works with GROMACS 3.3.3-beta
packages.
Yes, you can find it in the SVN repository linked at
Am Dienstag, den 18.12.2007, 21:23 -0600 schrieb Dirk Eddelbuettel:
On 19 December 2007 at 01:29, Manuel Prinz wrote:
| I'm not sure about that. I didn't see that on a quick read of chapters 8
| and 10, though policy states in 10.2:
| Packages that use libtool to create shared
Hi Ondrej!
Am Mittwoch, den 19.12.2007, 22:15 +0100 schrieb Ondrej Certik:
Unfortunately I still don't understand how it works. I admit
it can be my fault.
No problem. I'll try to explain the situation and reasoning below.
Let me repeat my question:
Why does openmpi use /usr/lib instead
Dear Sune!
Am Mittwoch, den 19.12.2007, 23:43 +0100 schrieb Sune Vuorela:
I have read the discussion in the bug report. If it is anywhere else, please
point to it instead of playing smart-ass.
That applies to everyone: I don't like the tone of the recent emails and
would be glad if we could
Am Mittwoch, den 19.12.2007, 06:58 -0600 schrieb Dirk Eddelbuettel:
On 19 December 2007 at 13:08, Manuel Prinz wrote:
| if we want to handle it via alternatives (which LAM doesn't) we have
| check the situation in pgapack, so we don't get a problem there. What is
| the advantage to have mpi.h
Hi Adam!
Thanks for your explanations. I have one question still:
Am Mittwoch, den 19.12.2007, 08:40 -0500 schrieb Adam C Powell IV:
I think the confusion is: the .la files are not the static libs, they
are libtool metadata files. The -dev package needs to include the .a
static libs. The
Hi Ondrej!
Am Donnerstag, den 20.12.2007, 08:00 +0100 schrieb Ondrej Certik:
thanks very much for your reply. As you explained in your previous email,
I think the misunderstanding is, that you and Dirk think, that
/usr/include/mpi.h
is symlinked to /usr/lib/openmpi/whatever, right?
No. We
tags 456869 + patch
thanks
Hello Ondrej,
attached you'll find a patch that solved the FTBFS of your package for
me. It patches the source directly, so you have to convert it so it can
be used with your favorite patch system.
The problem is that you can't find the MPI includes, as you already
Am Mittwoch, den 19.12.2007, 00:47 +0100 schrieb Manuel Prinz:
I just finished to develop a patch for our broken OpenMPI package. It
looks quite good in first tests. I'll do more tomorrow and will include
GROMACS. So I hope we'll have a fixed openmpi package in a few days.
I'll keep you
Hi Ondrej!
Am Donnerstag, den 20.12.2007, 20:13 +0100 schrieb Ondrej Certik:
[ Some confusing about /usr/include/mpi/mpi.h not being a symlink ]
No, I think this particular bug is solved.
What do you think about the symlink problem?
/usr/include/mpi/mpi.h is not a symlink because
fixed 455886 1.2.4-5
thanks
Am Donnerstag, den 20.12.2007, 17:24 +0100 schrieb Ondrej Certik:
Yes, you can close this bug. And thanks very much for prompt fixing.
OK, closing now. You're welcome.
Best regards
Manuel
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil
Am Donnerstag, den 20.12.2007, 19:57 +0100 schrieb Ondrej Certik:
Thanks again for the work you and Dirk are doing on the openmpi
package and especially the quick responses. And sorry if I made some
confusion.
No worries.
Best regards
Manuel
signature.asc
Description: Dies ist ein digital
Hi Ondrej!
Am Freitag, den 21.12.2007, 09:34 +0100 schrieb Ondrej Certik:
On Dec 21, 2007 9:10 AM, Manuel Prinz [EMAIL PROTECTED] wrote:
/usr/include/mpi/mpi.h is not a symlink because /usr/include/mpi is a
symlink. The MPI packages place all their header files in a directory
Opps
Am Mittwoch, den 19.12.2007, 00:39 +0100 schrieb Daniel Leidert:
The problem is caused in config_alternatives() near line 680. There
simply the link in /etc/alternatives is created, but it is not checked,
if $slavelinks[$slnum] exists. Now there is some code at line 597 for
the auto action,
Hello!
Am Dienstag, den 29.01.2008, 09:30 + schrieb W. Borgert:
Package: wnpp
Severity: wishlist
Owner: W. Martin Borgert [EMAIL PROTECTED]
Package name: refcard
Thanks for your great refcard! I've recommended your card for some years
now to Debian newbies and they found it really
Am Dienstag, den 29.01.2008, 17:15 +0100 schrieb Lapse of Reason:
I guess something else you did must have solved the problem. What else
did you do? Thanks...
The proposed solution worked for me. The following packages were
installed as well:
libedataserver1.2-9
libcamel1.2-10
libebook1.2-9
Hi Pawel!
Am Donnerstag, den 31.01.2008, 01:47 +0100 schrieb Paweł:
I have evolution in version: 2.12.3-1 also have the fallowing packages:
libedataserver1.2-9
libcamel1.2-10
libebook1.2-9
libecal1.2-7
libedata-book1.2-2
libedata-cal1.2-6
libegroupwise1.2-13
but I still have
Hi Nicholas!
Am Dienstag, den 30.10.2007, 10:33 -0700 schrieb Nicholas Breen:
Great! I'd been leaving it out because it didn't compile the last time
I'd checked it (which was admittedly back on openmpi 1.0.2) and wasn't
expecting to be able to add support until GROMACS 4.0. I'll be happy to
Package: gromacs-openmpi
Version: 3.3.2-2
Severity: grave
Justification: renders package unusable
Hi Nicholas,
we already talked about that bug. It affects the current version in unstable
on amd64 and is reprocudible in a lenny and sid chroot.
I think it's caused by LAM but I have not
Package: gromacs-openmpi
Version: 3.3.2-2
Severity: normal
Hi Nicholas,
gromacs-openmpi depends on lam4c2 which seems to be wrong. I think it's
related to #451991 but I'm not sure, so I file it as a seperate bug. (I'm
not sure about the severity either.)
Best regards
Manuel
-- System
Hi Nicholas!
Am Montag, den 19.11.2007, 19:36 -0800 schrieb Nicholas Breen:
[ Problem description ]
I'm confident that the issue you describe causes the problem. So our gut
feeling was right. I'll have a look at it.
I've set the bug severity at serious because this is a filename overlap
Package: libpam-slurm
Severity: wishlist
Tags: patch
Hello Gennaro,
please add support for uscan by adding the attached watch file. It worked for
me.
Best regards
Manuel
-- System Information:
Debian Release: lenny/sid
APT prefers testing
APT policy: (500, 'testing'), (500, 'stable'),
Package: slurm-llnl
Severity: wishlist
Tags: patch
Hello Gennaro,
please add support for uscan by adding the attached watch file. Those worked
for me.
Best regards
Manuel
-- System Information:
Debian Release: lenny/sid
APT prefers testing
APT policy: (500, 'testing'), (500, 'stable'),
Package: munge
Severity: wishlist
Tags: patch
Hello Gennaro,
please add support for uscan by adding the attached watch file. It worked for
me.
Best regards
Manuel
-- System Information:
Debian Release: lenny/sid
APT prefers testing
APT policy: (500, 'testing'), (500, 'stable'), (200,
tags 452047 + pending
thanks
Hi Nicholas,
I added a fix in trunk a while ago which seems to work and uses
alternatives. Nevertheless, when installing other MPI -dev packages, the
problem is still the same due to a bug in update-alternatives. I thought
of patching but it's not as easy as I
/60_amd64_double.dpatch 1970-01-01 01:00:00.0 +0100
+++ gromacs-3.3.1/debian/patches/60_amd64_double.dpatch 2007-09-18 16:39:48.0 +0200
@@ -0,0 +1,31 @@
+#! /bin/sh /usr/share/dpatch/dpatch-run
+## 60_amd64_double.dpatch by Manuel Prinz [EMAIL PROTECTED]
+##
+## DP: Fixes build issues
tags 39730 pending
thanks
Am Donnerstag, den 30.08.2007, 10:55 +0200 schrieb Manuel Prinz:
Dito. I changed the patch. I added a line to disable InfiniBand as a
comment, so everything that needs to be done is to remove the hash. I
also updated README.Debian.
Using this config without problems
Am Dienstag, den 14.08.2007, 08:55 -0500 schrieb Dirk Eddelbuettel:
I'd be happy to upload a fixed 1.2.3-4 [...]
I applied the patch to our repository before I noticed your email. I'm
not tagging this bug as pending as long as we're not sure if it fixes
the issue. Hope that's fine with everyone.
Hi Daobing,
I just noticed your ITP. I filed one myself a while ago (#416269) and am
working on that. It turned out that there are several issues that have
to been worked out first. I'm also new to packaging, so the progress was
rather slow but I hope I'll finish it soon.
If you're interested,
tags 437839 pending
thanks
Am Freitag, den 17.08.2007, 04:22 +0200 schrieb Uwe Hermann:
Anyway, my patch was incomplete, I missed some other architecture lines,
so here's an additional patch which should fix that.
This patch, plus a fixed libc0.1 (see other bug) now allow me to build
the
Am Freitag, den 17.08.2007, 14:49 +0200 schrieb Adrian Knoth:
On Fri, Aug 17, 2007 at 08:26:50AM -0400, Jeff Squyres wrote:
Ok, --enable-progress-threads and --enable-mpi-threads cause the
segfaults. If you compile without, everything works.
I'll now try if it's mpi-threads or the
Am Freitag, den 17.08.2007, 09:02 -0400 schrieb Jeff Squyres:
I don't think those options are safe on any architecture.
I'll disable them in debian/rules then and document it.
Dirk, are you fine with that?
Best regards
Manuel
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of
Am Freitag, den 17.08.2007, 14:27 -0500 schrieb Dirk Eddelbuettel:
On Fri, Aug 17, 2007 at 03:08:12PM +0200, Manuel Prinz wrote:
I'll disable them in debian/rules then and document it.
Dirk, are you fine with that?
Sure thing. We simply didn't know abput the brokenness re threads
Package: slurm-llnl
Severity: minor
Tags: patch
Hi Gennaro,
I noticed some override disparities on your QA site. The provided patch
should fix these.
Best regards
Manuel
--- slurm-llnl-1.2.11.orig/debian/control 2007-08-16 11:14:12.0 +0200
+++ slurm-llnl-1.2.11/debian/control
Package: gromacs
Version: 3.3.1-7
Severity: wishlist
Dear Nicholas,
please package the new upstream version, released on 2007-09-28. Thanks!
If I can help, feel free to ask!
Best regards
Manuel
-- System Information:
Debian Release: lenny/sid
APT prefers testing
APT policy: (500,
Hi Bernd!
Am Montag, den 15.10.2007, 13:23 +0200 schrieb Bernd Zeimetz:
I can also rename genbox to radgenbox.
I guess your package attracts more people, so it should probably keep
the binaries name. On the other side - if genbox is only rarely used in
gromacs, the better thing would be to
Am Montag, den 15.10.2007, 11:40 -0700 schrieb Nicholas Breen:
Huh, I wonder why that didn't get a mention on gmx-announce?
I missed it, too. It seems it wasn't announced anywhere besides the
homepage. I noticed it today, shortly before I filed this bug. To
announce it, so to speak. ;)
I'll
Package: wnpp
Severity: wishlist
Owner: Manuel Prinz [EMAIL PROTECTED]
* Package name: libpj-java
Version : 20070703
Upstream Author : Alan Kaminsky [EMAIL PROTECTED]
* URL : http://www.cs.rit.edu/~ark/pj.shtml
* License : GPL
Programming Lang: Java
tags 435573 pending
thanks
Hello Uwe,
The attached patch was enough for me to get a successful build on
kFreeBSD i386. It seems the build system copes fine with a missing
libibverbs otherwise, no further fixes seem to be required.
thank you for the patch! I applied it to our SVN repository
Am Donnerstag, den 06.09.2007, 15:33 +0100 schrieb Adam D. Barratt:
A fix for this (with a very similar patch :-) was commited to SVN a few days
ago and will therefore be in the next upload.
Great! Thanks!
Best regards,
Manuel
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject
Am Freitag, den 14.09.2007, 14:56 +0800 schrieb LI Daobing:
On 9/14/07, Nicholas Breen [EMAIL PROTECTED] wrote:
in most simulation in gromacs, gromacs need cpp command to process the
input file. So how about let gromacs depends on or recommends `cpp'.
I'd think a Suggests would be most
Package: gromacs
Version: 3.3.1-5
Severity: normal
Dear Nicholas,
I have trouble compiling GROMACS with DEB_BUILD_OPTIONS=double on amd64. The
buildlog reads:
/bin/sh ../../libtool --tag=CC --mode=link x86_64-linux-gnu-gcc -O3
-fomit-frame-pointer -finline-functions -Wall -Wno-unused
Hi Andreas!
Am Sonntag, den 26.08.2007, 18:37 -0400 schrieb Andreas Klöckner:
I don't have any Infiniband cards, but whenever I run an OpenMPI job,
libibverbs complains about not being able to find IB hardware. Life would be
much easier without this warning (which is printed by each nodein
Am Sonntag, den 26.08.2007, 20:27 -0500 schrieb Dirk Eddelbuettel:
There's a config option to turn it off. That was discussed on the ompi lists
a while back.
I think you refer to [1]?
Another option is to just rebuild and comment that activation out. I did that
with the 1.1* series... and I
Am Montag, den 27.08.2007, 11:50 +0200 schrieb Tilman Koschnick:
The option to turn the warning off during runtime is:
$ mpiexec --mca btl ^openib ...
This can easily be set as an alias so one doesn't have to type it again
and again:
$ alias mpiexec='mpiexec --mca btl ^openib'
This
Am Montag, den 27.08.2007, 08:14 -0500 schrieb Dirk Eddelbuettel:
I'd vote for actually 'Debian patching' the config file to set ^openib, and
adding a line there comment out if you have infinit band and want to use it.
Works for me. What are we going to do for the other supported
Am Montag, den 27.08.2007, 08:14 -0500 schrieb Dirk Eddelbuettel:
I'd vote for actually 'Debian patching' the config file to set ^openib, and
adding a line there comment out if you have infinit band and want to use it.
I checked in a patch that modifies the config and documents the issue in
Am Mittwoch, den 29.08.2007, 07:03 -0500 schrieb Dirk Eddelbuettel:
Empirically thinking, I'd say leave the patch as we have yet to meet actual
infiniband users. But we need the big fat warning somewhere. README.Debian
is a start.
That's true. But I fear it's not enough for a change in
Am Donnerstag, den 30.08.2007, 10:01 +0200 schrieb [EMAIL PROTECTED]:
On Wed, Aug 29, 2007 at 03:12:26PM +0200, Uwe Hermann wrote:
I think the default should be a working Infiniband install, and somebody
who wants to silence a few messages can edit one or two lines in the
config file (no I
Package: devscripts
Version: 2.10.7
Severity: normal
Hi,
I noticed that licensecheck fails parsing license information when single
line comments (//) in Java files are used. The attached patch fixed the
issue for me.
Best regards,
Manuel
-- System Information:
Debian Release: lenny/sid
APT
Package: wnpp
Severity: wishlist
Owner: Manuel Prinz [EMAIL PROTECTED]
X-Debbugs-Cc: [EMAIL PROTECTED]
* Package name: jfftw
Version : 20070725
Upstream Author : Will Hossack [EMAIL PROTECTED]
* URL : http://www.ph.ed.ac.uk/~wjh/teaching/Java/fft/
* License
Am Freitag, den 28.09.2007, 09:36 -0500 schrieb Dirk Eddelbuettel:
So we need Conflicts: and Replaces: for openmpi-bin, and possible also
Provides: -- as I always forget this part, could someone check and
then test the upgrade from 1.2.3-4.
I'll have a look at it! (I broke it in the first
Hi Keith!
Am Sonntag, den 30.09.2007, 01:58 -0600 schrieb Keith Hellman:
Package: openmpi-bin
Version: 1.1-2.5
The 1.1 series of OpenMPI is known to be buggy and no longer supported
upstream.
Is it possible for you to use the package in unstable (1.2.4-1) and
check if the bug still exists
[ CC'ing the BTS for documentation purposes ]
Am Sonntag, den 30.09.2007, 09:29 -0600 schrieb Keith Hellman:
On Sun, Sep 30, 2007 at 01:17:08PM +0200, Manuel Prinz wrote:
Am Sonntag, den 30.09.2007, 01:58 -0600 schrieb Keith Hellman:
Package: openmpi-bin
Version: 1.1-2.5
The 1.1
tags 445230 confirmed
thanks
Hi Andreas!
Am Donnerstag, den 04.10.2007, 00:58 -0700 schrieb Andreas Kabel:
Compiler wrappers don't work. Required data files
are not part of the package, nor are they in any
package libopenmpi-dev depends on.
[EMAIL PROTECTED]:~$ mpicc.openmpi
Cannot open
Am Donnerstag, den 04.10.2007, 17:10 -0500 schrieb Dirk Eddelbuettel:
Manuel, you said that you replicated Andreas' finding.
I did.
What version was that?
1.2.4-2
(My home system is still 1.2.3-2 as on testing, so whatever we may
have broken in the 1.2.3-{3,4} versions doesn't bite me).
tags 445230 + pending
thanks
Am Donnerstag, den 04.10.2007, 17:43 -0500 schrieb Dirk Eddelbuettel:
Ah, yes, have the -dev depend on openmpi-common. I read your mail 'the other
way around'. We don't want every user of -common to also have to have -dev.
Full ACK! It wouldn't solve the problem
Am Freitag, den 05.10.2007, 01:03 +0200 schrieb Manuel Prinz:
I've checked in a modified debian/control. Building and testing has to
be done tomorrow morning.
Done. It built, installed and worked fine. The bug can be closed with
the next upload.
Best regards
Manuel
--
To UNSUBSCRIBE
Package: gromacs
Version: 3.3.2-1
Severity: wishlist
Tags: patch
Hi Nicholas,
as a member of the OpenMPI maintainer group, I'm very interested in OpenMPI
support for GROMACS. Please find the attached patch to add it. It compiles
fine in single and double precision modes on my maschine.
OpenMPI
Am Dienstag, den 30.10.2007, 10:33 -0700 schrieb Nicholas Breen:
as a member of the OpenMPI maintainer group, I'm very interested in OpenMPI
support for GROMACS. Please find the attached patch to add it. It compiles
fine in single and double precision modes on my maschine.
Great! I'd
Hi Roland!
Am Freitag, den 02.11.2007, 14:45 -0700 schrieb Roland Dreier:
As far as I can tell, the dependency on libsysfs-dev should be removed.
Thanks for pointing this out! We probably wouldn't have noticed that.
I checked all Depends: and none references sysfs, so I think it's safe
to
1 - 100 of 268 matches
Mail list logo
