[MOAB-dev] Problems compiling Lasso with CUBIT-linked CGM and MOAB

[Tim Tautges]

On 12/14/2011 10:44 AM, Michael Fladerer wrote:
> Hi Tim,
>
> thanks for your quick response!
>
> On Wed Dec 14, 2011 at 09:12:54 -0600, Tim Tautges wrote:
>> Are you grabbing the nightly source tarballs for CGM, Lasso, and
>> MOAB?  The error you're having below looks like your Lasso is out of
>> date.
>>
> yes, I just grabbed the nightly-lasso.tar.gz[1] linked on the ITAPS
> website.  So, I think Lasso should be up-to-date, shouldn't it?
>

You should probably grab it from the Lasso wiki, at http://trac.mcs.anl.gov/projects/ITAPS/wiki/Lasso.

>> For you MOAB build, if you'll be linking with CGM through Lasso
>> later, I suggest configuring *MOAB* without CGM, though I don't know
>> whether that will always fail (I doubt it, assuming you're using the
>> same version of CGM).
>>
> OK, I'll check that next.  I thought there was a strict installation
> order (like CGM frist, then MOAB and then Lasso), because of the
> available configuration options (--with-igeom --with-imesh --with-irel).
>

MOAB and CGM can be installed independent of each other, in either order.  When you make MOAB dependent on CGM, you 
simply enable a MOAB reader that reads geometric model facets and geometric topology into MOAB and represents them as 
entity sets.  If you do enable that option, then you need to build CGM before MOAB.

>> The only other thing I see that might be an issue is that your path
>> (PATH) env variable should include cubit's bin directory.
>>
> Right, my PATH didn't contain cubit's bin dir.  I thought it was
> unnecessary for building each of the software, but maybe it's not, so
> I'll correct that befor the next build attempt.
>
>> For each of the libraries, make sure to run 'make check' after
>> building, to make sure the build worked (for that to work, though,
>> MOAB needs to be configured with netcdf).
>>
> Netcdf was already installed before I started compiling.  However, I
> didn't run the check for all the packages, as they seemed to compile and
> install without giving a real error.  But, of course, I'll run them now
> and from now on, to make sure everything is alright.

Never a bad idea.  We use our make check pretty frequently, so they're designed to run relatively fast (if anything, we 
probably need more tests...)

>
>> One other simplification is that, in your makefile, you can include
>> iRel-Defs.inc, then use IMESH_INCLUDES, IGEOM_INCLUDES, and
>> IREL_INCLUDES make variables on your compile line, and IXXX_LIBS on
>> your link line (XXX=GEOM, MESH, REL); that'll pull in the proper
>> include/lib search paths and other options.
>>
> Ah, good to know, thanks!  I already did that when I compiled mesquite
> as it was already documented in the readme file.  I'll give it try for
> Lasso, too.
>
> Again, thank you very much!

You bet, let us know if you have any problems.


- tim

>
> Cheers,
>   Michael
>

-- 
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (tautges at mcs.anl.gov)      (telecommuting from UW-Madison)
  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
             fax: (608) 263-4499      Madison, WI 53706