Re: Any progress with FIS GT.M?
Hi Luis,
On Tue, Apr 24, 2012 at 07:02:00AM -0400, Luis Ibanez wrote:
> Yes, there has been a bit more progress.
Great.
> In Intel 64 bits, it builds the libraries and executables;
> while in Intel 32, it builds the libraries but I'm having
> trouble with a few (4 or 5) missing symbols when
> linking the executables.
I guess you discussed this with upstream, right? If yes it might make
sense to add links to some mailing list archive or so. That way you
might increase the number of eyes looking at the problem.
> The debian scripts are updated to use this version 5.5,
> by invoking cmake for configuration.
Good.
> Although in recent messages in the list, I learned that
> the debian build system may have a more natural way
> of using cmake, than what I put in the 'rules' file:
>
> override_dh_auto_configure:
> cp -a ./$(SRC_EXTRAS_DIR)/* .
> cmake .
> dh_auto_configure
That's OK. I do not know what you mean by "natural way" - we should
just gain for a way that works. (By chance yesterday I failed in
migrating a cmake based package (seqan) from cdbs to debhelper because
cmake was not called with proper options - so there might be some issues
with cmake support when using dh.)
> What lays ahead:
>
> 0) Update the Git branch with some recent security
> patches from fis-gtm upstream.
>
> 1) Fix the missing symbols when building
> executables in Intel 32 bits.
>
> 2) Incorporate post-build steps from fis-gtm scripts
>
> 3) Incorporate installation (final destination)
> from fis-gtm scripts
>
> 4) Put heads together with the upstream team and
> look at how they could carry the CMakeLists.txt
> files in their source tree, as a parallel configuration
> system, as well as vetting the bootstrap-generated
> files. We have exchanged emails on this with
> Bhaskar.
>From my (Debian based) perspective item 1 might be the hardest blocker
and we do not have influence on 4. How do you see the chances to get
1) fixed and are you getting help in it?
> More brains and hands are very welcome,
> to help bring this to completion. :-)
If you want me having a look please give a short step by step recipe how
to reproduce the problem and a summary of the discussion you had with
upstream who might be best qualified to answer questions like this.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: