GitLab

o Move logic for generating fortran protos to separate script
o Generate MBCN_protos.h automatically


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3039 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3038 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3037 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

* "install" actually works
* "clean" works
* More useful options to install, thanks to Jason


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3036 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

 - allow user-specified python interpreter with PYTHON variable
 - check that python interpreter works
 - if user doesn't specify directory for python C includes, guess
 - check that python includes dir contains Python.h
 - check that python has NumPy module
 - if user doesn't specify directory for NumPy C includes, guess
 - check that NumPy include dir contains arrayobject.h 
 - error if --enable-pytaps w/out working python iterpreter
 - error if --enable-pytaps w/out Python.h
 - error if --enable-pytaps w/out numpy/arrayobject.h
 - error if --enable-pytaps w/out --enable-imesh
 - fix non-portable syntax: test operator is '=', not '==', for string compare
 
 
 


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3035 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

Add new version of MBCN::SubEntityVertexIndices that is a little faster than 
 the old one (avoids coping data) as a compromise for removing the
 direct access to MBCN internals that was broken in the skinner.

Add unit tests for skinning edges, faces, and regions (three separate tests.)



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3034 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

Add new version of MBCN::SubEntityVertexIndices that is a little faster than 
 the old one (avoids coping data) as a compromise for removing the
 direct access to MBCN internals that was broken in the skinner.

Add unit tests for skinning edges, faces, and regions (three separate tests.)



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3033 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

  merged vertices when vertices are contained in tracking sets.

o Fix above bug

o Fix incorrect error handling in MBCore::merge_entities: if updating
  of references fails, don't delete the entity.



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3031 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

autotools-processed version...


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3030 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3029 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

a partially-used array, don't use the total size as the offset.  That is,

   std::fill(ps+num_exist, ps+MAX_SHARING_PROCS, -1)

is REALLY BAD for ps[MAX_SHARING_PROCS] and num_exist > 0.

Added some more assert checks and a local check that makes sure the intersection
of procs sharing vertices of an entity is a superset of those sharing the
entity.



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3028 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3027 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

* Add support for installing Python interface to alternate locations via 
  autotools


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3026 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

* Minor adjustment to Python's lookup of iMesh-Defs.inc (see build.rst 
  for info)


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3025 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3023 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3022 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3021 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3020 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

when reading in parallel, rather than scanning file on every proc and to read summary, read only on root and bcast to others

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3019 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

  to read.

o Add a new function to MBReaderIface that reads only the values of 
  a tag from the file.
  
o Update all readers for above changes (most return MB_NOT_IMPLEMENTED
  for this functionality.)

o Implement read of only tag data for HDF5 and Exodus file formats,
  and add simple tests.



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3018 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

infer which processors may share higher-dimensional (non-vertex)
entities based on which processors share the vertices in that entity.
In some cases, that results in a wrong guess, i.e. in reality some of
those processors may not share that entity.  That appears in the
implementation as the other processors having zero remote handles for
the processor(s) wrongly predicted to share the entity.  This commit
fixes up the sharing data for those entities, without requiring
sending extra messages.

This fixes the problem reported by Lukasz, at least on 4 procs.

There is a remaining issue in parallel sharing, where
higher-dimensional entities on the interface aren't marked with
ownership data (interface sets might not be marked either, can't
remember).

Parallel hdf5 write still fails too.




git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3017 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3016 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

  If non-null, file readers store file IDs in designated tag.
o Clean up any created entities if read of SMS file fails
o Check geometry dimension value read from SMS file before using it
  as an array index.
o Add error handling to ReadIDEAS
o Fix bug in ReadIDEAS: wrong vertices uses in element connectivity
  if database is not empty when reading file.



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3015 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

Added handling of PRINT_PARALLEL read option; this option tells ReadParallel
to print information about shared entities after parallel read is done.

Added -p option to mbparallelcomm_test to turn on above option.




git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3014 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3013 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3012 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

o use HDF5 MPIO for true parallel read


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3011 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3010 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3009 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3007 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

* replace cgm2moab function with ReadCGM in both DagMC & cgm2moab app
* allow default values for tags in get_tag
* separate load_file_and_init into two functions in preparation for
  parallel version
    * load_file (now based on ReadCGM)
    * init_OBBTree
* separate tight connection with MCNP in code base
    * rename and cleanup DagMC settings
    * separate write_log into two functions
        * parse_metadata
        * write_mcnp (this can be replicated for other codes)
* change parsing of group names
    * change some keywords to include '.' instead of '_' so that 
      '_' can be used as delimiter for tokenizer



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3006 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

* Prepare for being able to read CUB files


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3005 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

functions in the top-level namespace: 'intersect' and 'subtract'.  For
exaple, the rather ambiguous:
  diff = from_this.subtract( some_range );
is now:
  diff = subtract( from_this, some_range );


git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3004 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3003 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3002 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

- in MeshTopoUtil::get_bridge_entities, if to_dim is -1, then all
entities (of all dimensions) adjacent to bridge entities are returned,
rather than just entities of to_dim.  This is a hack for now; probably
need extra arg requesting aentities too

- fixed MBCore::SideNumber to work correctly with structured mesh

Passes make check.



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3001 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

idea (send size in first message, which indicates if and how big 2nd
message needs to be).

Fix bug reported by Lukasz Kaczmarczyk regarding ghost exchange with
tet mesh.

Passes all tests except parallel_hdf5_test, which was failing before.

Specific changes:

mbparallelcomm_test: changing default for with_ghosts back to 1, so we
get ghosts by default.

MBParallelComm: 
- changing INITIAL_BUFF_SIZE to unsigned
- added MPE states, #ifdef'd by DEBUG_MPE
- fixed bug where during interface entity exchange, was creating
new entities when existing one wasn't found; proper behavior is to not
create those entities
- changed sending of messages during ghost exchange to always include
message size as first datum in message; for large messages, this tells
receiving processor to always post a 2nd receive, and tells how big
that receive should be; allows us to use Isend on first message too; uses
different message type for 2nd message



git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@3000 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@2999 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@2998 6162379e-bd28-0410-9a7d-b7f4dcfcad3c

git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@2997 6162379e-bd28-0410-9a7d-b7f4dcfcad3c