<div dir="ltr">Go for it! I want to see how it turns out.<div><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Sun, Jan 19, 2014 at 6:41 PM, Sean Farley <span dir="ltr"><<a href="mailto:sean@macports.org" target="_blank">sean@macports.org</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5"><br>
<a href="mailto:sean@macports.org">sean@macports.org</a> writes:<br>
<br>
> <a href="mailto:sean@macports.org">sean@macports.org</a> writes:<br>
><br>
>> <a href="mailto:sean@macports.org">sean@macports.org</a> writes:<br>
>><br>
>>> Hello all and Happy Boxing Day!<br>
>>><br>
>>> I have done a complete rewrite of the compilers and mpi port groups<br>
>>> based on suggestions from previous emails. I will try to keep this email<br>
>>> short and provide links for those that want to know more.<br>
>>><br>
>>> Highlights:<br>
>>><br>
>>> - specify which compilers to set via compilers.choose,<br>
>>> e.g. 'compilers.choose f77 f90 fc'<br>
>>><br>
>>> - functions for testing if fortran has been selected for optional<br>
>>> interfaces<br>
>>><br>
>>> - portfile author can specify if fortran (or mpi) is required<br>
>>><br>
>>> - unified and have made non-conflicting openmpi and mpich ports; also<br>
>>> added an openmpi-devel port [1]<br>
>>><br>
>>> - can now select either mpi port as the default mpi installation<br>
>>><br>
>>> Example of a portfile using the new compilers portgroup:<br>
>>> <a href="https://smf.io/macports/changeset/compilers" target="_blank">https://smf.io/macports/changeset/compilers</a><br>
>>><br>
>>> Example of a portfile that uses the mpi portgroup:<br>
>>> <a href="https://smf.io/macports/files/mpi/dports/science/hdf5-18/Portfile" target="_blank">https://smf.io/macports/files/mpi/dports/science/hdf5-18/Portfile</a><br>
>>><br>
>>> The changes I made to the sparskit and hdf5-18 portfiles are just<br>
>>> examples I did for illustrative purposes. I won't push them unless the<br>
>>> portfile author wants it. The mpich and openmpi portfiles need these<br>
>>> changes so that the mpi portgroup will work.<br>
>>><br>
>>> Anyway, I'd like to push this soon so that I can continue other work<br>
>>> (and close a lot of tickets), so it'd be great if others could take a<br>
>>> look at the proposed changes.<br>
>>><br>
>>> [1] <a href="https://smf.io/macports/changeset/81bb51" target="_blank">https://smf.io/macports/changeset/81bb51</a> or look at the changelog<br>
>>> <a href="https://smf.io/macports/changelog" target="_blank">https://smf.io/macports/changelog</a> to see all the diffs<br>
>><br>
>> Just an update to say that I've gone ahead and finished updating all the<br>
>> ports that use mpi to use this new port group. The only work that needs<br>
>> to be done is to get some review of this patch series and after that,<br>
>> permission to push.<br>
>><br>
>> I've cc'd the people who are the maintainers of the ports I changed,<br>
>> listed below. It'd be great to have you guys look at the changes.<br>
>><br>
>> dstrubbe: sparskit, hpl, octopus<br>
>><br>
>> eborisch: mpich<br>
>><br>
>> mww: openmpi<br>
>><br>
>> raimue: valgrind, valgrind-devel<br>
>><br>
>> takeshi: berkeley_upc, omnixmp, gnudatalanguage, netcdf, netcdf-cxx,<br>
>> netcdf-cxx4, netcdf-fortran<br>
>><br>
>> mmoll: optpp, hdf5-18, arpack<br>
>><br>
>> hum: plda<br>
>><br>
>> howarth: apbs-mpi<br>
>><br>
>> mattoates: raxml<br>
>><br>
>> mk: scotch<br>
>><br>
>> michaelld: SuiteSparse<br>
>><br>
>> Check out <a href="http://smf.io/macports" target="_blank">http://smf.io/macports</a> for the updated changes.<br>
>><br>
>> Having other people look at this would be great and hopefully would<br>
>> catch some of the errors I missed. The common ones I noticed were<br>
>> missing revbumps (though, I think I got all of those now).<br>
>><br>
>> Documentation is very much "read the comments and the code itself or<br>
>> look at an example" so, apologies about that. Also, sometimes I couldn't<br>
>> think of a good name for a proc, so if anyone has a better name, please<br>
>> speak up.<br>
>><br>
>> Some changes since last email are:<br>
>><br>
>> - ensure the same mpi is used via mpi.enforce_variant<br>
>><br>
>> - ensure the same c compiler is used (even when using mpi) via<br>
>> compilers.enforce_c<br>
>><br>
>> - similar for fortran, compilers.enforce_fortran<br>
>><br>
>> - test for avx compatible compiler via avx_variant_isset<br>
>><br>
>> Still yet to be done is to replace all custom recipes done for the<br>
>> compiler variants. By my count, there are no more than 72 of those<br>
>> ports. This is lower priority since they'll still work as is.<br>
><br>
> If nobody has any objections to this, I was hoping to push this in a few days.<br>
<br>
</div></div>Any objections before I push? I guess at this point the errors that are<br>
left are of the type that one finds after trying to get the buildbot to<br>
compile a port.<br>
<div class="HOEnZb"><div class="h5">_______________________________________________<br>
macports-dev mailing list<br>
<a href="mailto:macports-dev@lists.macosforge.org">macports-dev@lists.macosforge.org</a><br>
<a href="https://lists.macosforge.org/mailman/listinfo/macports-dev" target="_blank">https://lists.macosforge.org/mailman/listinfo/macports-dev</a><br>
</div></div></blockquote></div><br></div>