fix: properly handle provision of optional configuration files

[phellipecouto]

Enabled optional_config_files provision through config.yaml. Prior to this change, ROMS varinfo file was being considered a mandatory model config file and mistakenly hard coded into config_files list, leading to model initialisation failure when file name differed from varinfo_seacofs.yaml.

[coveralls]

coverage: 59.162% (+0.08%) from 59.084%
when pulling 270ea77 on ACCESS-NRI:feature/roms-driver
into e6cef2d on payu-org:master.

[aidanheerdegen]

Given this is a fairly significant departure from how optional config files are handled it would be good to have this documented.

We already have an open issue for the lack of documentation of drivers

#470

How is your RST @phellipecouto? Would you be willing to be the first to add driver docs?

[aidanheerdegen]

LGTM. We can deal with docs later ...

[phellipecouto]

Given this is a fairly significant departure from how optional config files are handled it would be good to have this documented.

We already have an open issue for the lack of documentation of drivers

#470

How is your RST @phellipecouto? Would you be willing to be the first to add driver docs?

OK good to hear this! I had a slight impression I could be going on a different direction as I could not find a working example of this on other drivers. Can you point me to the right direction and I will assess if this would still work for the ROMS case?

[aidanheerdegen]

Can you point me to the right direction and I will assess if this would still work for the ROMS case?

I think what you've done is fine. Off the top of my head I can't think of any other drivers that have dynamic model config files, so I think it's fine.

I'll start a separate PR to add driver docs for more of the models, and get your input for ROMS then.