CellML files as submodule or with cmake_fetch about chaste

[mirams]

I think I didn't realise that the ontology submodule had disappeared into chaste_codegen.

[mirams]

For reproducibility we probably need to pin to a given commit of Chaste/cellml, otherwise we won't know exactly what files were being tested without lots of looking at timestamps etc.

[MauriceHendrix]

Just a couple of things to add:

• Changing the commit is straightforward, here's an example where I've used it for a different project
• It only gets fetched once per configuration, not per build, so it won't get endlessly getting it each time you re-build

The downside, in fairness, is that the data will be retrieved more often than if it were a submodule as you may well work with multiple configurations.

@MauriceHendrix how much data are we talking about?

about 19MB

[fcooper8472]

Given it's not much data, my personal opinion would be to get this through CMake FetchContent.

[mirams]

And add lots of documentation at the top of all the relevant tests to say how to change that commit!

[fcooper8472]

Just a couple of things to add:

• Changing the commit is straightforward, here's an example where I've used it for a different project
• It only gets fetched once per configuration, not per build, so it won't get endlessly getting it each time you re-build

The downside, in fairness, is that the data will be retrieved more often than if it were a submodule as you may well work with multiple configurations.

@MauriceHendrix how much data are we talking about?