Handle I/O failures with HDF5

[jtebert]

There are examples of using try/catch for this in the HDF5 tutorial examples.

These might also be the cause of some intermittent segfaults?

[jtebert]

(Segfaults might also be caused by unclosed files, or for that occasional segfault at the end, by something with shared pointers)

[jtebert]

Program segfaults at end if the H5 file did not exist before running the simulation. Everything seems to work and the data is correctly saved, though. I suspect it's somewhere in a destructor.

[jtebert]

Another one: if the folder for the logger doesn't exist, the programs fails with an error message and exists gracefully, but the error message isn't informative. (Generic HDF IO error, I think.)