h5py 2.6 bug

[Paul Romano]

Hi OpenMC devs,

Just wanted to give you all a heads up that I made the unusual move of making a direct commit to the develop branch. Evidently there is a bug in h5py 2.6 (which conda just picked up within the last few days) that has resulted in any recent CI runs failing. It has to do with compound datatypes. If you have a recent PR that has failed, try re-running it now and it should be good.

Best,

Paul