On 11/6/13 4:07 PM, "Ralph Castain" wrote:
>Afraid I couldn't parse that debian link - no idea what any of it meant,
>and there was no identified problem listed there.
There's a bugs column you can click on the find the list of outstanding
bugs.
>My concern here is that we seem to be putting so
On 06/11/2013 18:48, Jeff Squyres (jsquyres) wrote:
> (we can consolidate this down to 1 thread, not multiple)
>
> The Debian build stuff is not maintained by us (and is therefore not in our
> tree); it's maintained by Sylvestre Ledru . I'm not
> sure where the original packaging for that is dev
Afraid I couldn't parse that debian link - no idea what any of it meant, and
there was no identified problem listed there.
My concern here is that we seem to be putting something in the OMPI code base
that has nothing to do with OMPI itself. Debian has had an OMPI package for
awhile now, and in
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 07/11/13 04:40, Mike Dubman wrote:
> I did not find debian packaging files in the OMPI tree, could you please
> point me to it?
As Sylvestre explained Debian (and presumably Ubuntu too) will
automatically delete any /debian/ directory in an upstre
What: I want to add another member to mca_mpool_base_registration_t for mpools
that
need to keep track of extra context data.
Why: I plan to add a new mpool to support Cray's udreg registration cache. This
new
mpool will allow us to turn off the memory hooks on Cray systems.
Unfortunately, in
o
http://anonscm.debian.org/viewvc/pkg-openmpi/openmpi/
svn://svn.debian.org/svn/pkg-openmpi/openmpi/
FWIW, hwloc debian packaging is maintained by one of the upstream devs,
but he didn't have to pollute the upstream hwloc repo with debian stuff.
There's a different repo with only the debian subdire
i see.
It sounds like a good idea to have all OMPI packaging files in one
place/repo to allow easy/unified packaging into vendor distributions (like
OFED, MLNX OFED, etc..)
On Wed, Nov 6, 2013 at 7:48 PM, Jeff Squyres (jsquyres)
wrote:
> (we can consolidate this down to 1 thread, not multiple)
Dave,
this sounds like a good idea, will give it a try.
Thx
M
On Wed, Nov 6, 2013 at 7:35 PM, Dave Goodell (dgoodell)
wrote:
> Mike,
>
> Why not put these packaging files in "/contrib/dist/..." in SVN and then
> symlink them to "/debian" as a step in your build script? Top level names
> are (so
(we can consolidate this down to 1 thread, not multiple)
The Debian build stuff is not maintained by us (and is therefore not in our
tree); it's maintained by Sylvestre Ledru . I'm not sure
where the original packaging for that is developed / maintained -- Sylvestre?
On Nov 6, 2013, at 9:20
hi Jeff,
the debian packaging just requires that debian/ folder should be located in
root dir :(
(did not find any explanation about it)
I did not find debian packaging files in the OMPI tree, could you please
point me to it?
Thanks
On Wed, Nov 6, 2013 at 4:26 PM, Jeff Squyres (jsquyres)
wrote:
Mike,
Why not put these packaging files in "/contrib/dist/..." in SVN and then
symlink them to "/debian" as a step in your build script? Top level names are
(somewhat) precious and should not be added casually.
-Dave
On Nov 6, 2013, at 4:50 AM, svn-commit-mai...@open-mpi.org wrote:
> Author:
unfortunately, debian/ packaging works only if located in rootDir. :(
If you have better ones - can you put it and we will use it?
On Wed, Nov 6, 2013 at 4:38 PM, Sylvestre Ledru wrote:
> Hello,
>
> Le 06/11/2013 15:26, Barrett, Brian W a écrit :
> > Hi all -
> >
> > Mike added a debian/ directo
On Wed, Nov 06, 2013 at 02:06:15AM +, Jeff Squyres (jsquyres) wrote:
> On Nov 5, 2013, at 2:59 PM, George Bosilca wrote:
>
> > I have a question regarding the extension of this concept to multi-BTL
> > runs. Granted we will have to have a local indexing of BTL (I'm not
> > concerned about thi
On Nov 6, 2013, at 6:48 AM, "Barrett, Brian W" wrote:
> Yeah, I noticed that. However, I'm not sure that's the best way to do
> that. First, build-time symlinks are good. Second, I'd prefer always
> installing openmpi/opal/opal_portable_platform.h over having two copies of
> the same file in t
On 11/6/13 7:44 AM, "Jeff Squyres (jsquyres)" wrote:
>I looked at that, too -- I think the only reason to have 2 files is that
>mpi_portable_platform.h is included by mpi.h (its used for knowing how to
>define __mpi_interface_deprecated__ in mpi.h), and is therefore
>installed. opal_portable_pla
I looked at that, too -- I think the only reason to have 2 files is that
mpi_portable_platform.h is included by mpi.h (its used for knowing how to
define __mpi_interface_deprecated__ in mpi.h), and is therefore installed.
opal_portable_platform.h is not installed.
On Nov 6, 2013, at 6:40 AM,
All -
Is there a reason we have both opal/include/opal_portable_platform.h and
ompi/include/mpi_portable_platform.h? If they're always supposed to have
the same content (which appears to be the case), then it seems like
building mpi_portable_platform.h from opal_portable_platform.h at build
time
Hello,
Le 06/11/2013 15:26, Barrett, Brian W a écrit :
> Hi all -
>
> Mike added a debian/ directory to the top-level of the tree this morning,
> which looks to be helping in building a Debian package. While I don't
> mind helping Debian, I'm really against having a debian/ directory in our
> top
Hi all -
Mike added a debian/ directory to the top-level of the tree this morning,
which looks to be helping in building a Debian package. While I don't
mind helping Debian, I'm really against having a debian/ directory in our
top-level tree. We have a place for those things (in contrib/). If
D
Mike --
What is this Debian stuff, and why is it at the root level? Your commit
message didn't say *why* it can't be in the usual contrib/dist/linux location.
I would think that the addition of a new top-level directory warrants some
discussion before committing...
Also, how does this compar
Ok, I changed the name to "btl_usnic" in SVN (and can change it again if a
better suggestion comes by...).
On Nov 5, 2013, at 8:24 PM, Paul Hargrove wrote:
> I stand by my previous "vote"
>
> "btl_usnic" gets 90% of my vote.
> "btl_usnic_enum" gets 10%.
> "btl_usnic_*_enum" gets nada.
>
> Ra
21 matches
Mail list logo
