Arlie Capps

@cyrush , @aaroncblack and I talked about this routine. It seems this takes most of the time in a run he is interested in, so this was a first pass...

@cyrush Thanks for the fixes and for merging it in.

@cyrush , thanks for the discussion Friday March 25.

You wrote, "One thing I want to make sure we capture. The expected shape of the vert-assod and ele-assod fields is determined by the topology (that's a relationship we are...

> `offsets` and `strides` still need `shape` to be interpreted. > > (the topo's `shape` can imply default `offsets` and `strides`, however `offsets` and `strides` are not useful for indexing...

OK, that opening comment was off-target. `Group::load()` and `save()` wrap Conduit's `relay::io::load()` and `save()`. Those functions use exceptions to report problems. If it didn't throw an exception, it went OK....

hmmm, it isn't resolved. I would like to discuss it with the group. The fact that we don't have a clear status report bugs me, but it is non-trivial to...

To excerpt the log file, - test spio_parallel died in conduit::relay, in hdf5_open_file_for_read(), with message "HDF5 Error code-1 Error opening HDF5 file for read only access: out_spio_parallel_write_read.root" - test quest_point_in_cell_mfem...

Edit: I reproduced the failures reported by @yurivict , using v0.6.1. Additionally, if you enable examples, a few other problems pop up. - hdf5_open_file_for_read_write or hdf5_open_file_for_read is the site of...

Moving from v0.6.1 to develop, the HDF5 problems persist (hdf5_open_file_for_read_write and hdf5_open_file_for_read crash in test spio_parallel and examples sidre_createdatastore_parallel, sidre_generateindex_parallel) but the MFEM crash seems to have been fixed. This...