Patch looks ok, but I a) didn't test it and b) don't know how to work these Fortran pointers (otherwise I wouldn't have written the now removed comments :). If this passes the tests (and is indeed tested), please commit. Leaving as 'review' since you might want to find someone with more Fortran knowledge to review this.
It passes all existing EOSOmni table reader tests -- of which there are none :-). Seriously though I added a tiny function to dump the read in data into an ascii file and the ascii file compares identical before and after the patch. This is not urgent though and we can use to hunt around for a EOS Omni table reader test suite.
In this case ok to commit - but I think this is not important enough to go into the next release, so please hold back until after it.