Fix checkpoint value parsing

[timofeymukha]

Fix issue #2301.

Additionally, fixes some default values to be the same as in the documentation (no defaults).

[timfelle]

The checkpoint options are kinda funky now.

If a user does not specify case.output_checkpoints at all, then we expect they intend to save checkpoints. Therefore it is mandatory to have case.checkpoint_control in the case file, since we now removed the default value for that. Is that really intentional?Or would it not be more intuitive for the default of case.output_checkpoints to be false, since if a user wants checkpoints, they probably know how they want them to be spaced.

This currently cause the rayleigh_benard case to fail.

[timofeymukha]

The checkpoint options are kinda funky now.

If a user does not specify case.output_checkpoints at all, then we expect they intend to save checkpoints. Therefore it is mandatory to have case.checkpoint_control in the case file, since we now removed the default value for that. Is that really intentional? Or would it not be more intuitive for the default of case.output_checkpoints to be false, since if a user wants checkpoints, they probably know how they want them to be spaced.

This currently cause the rayleigh_benard case to fail.

Ugh, yeah, good catch! I guess both defaults are defensible; we can discuss at the meeting today!

[timfelle]

Also missing changelog update