Further improve coverage

[sloede]

I modified the positivity tests such that the limiters in 1D and 2D are actually triggered.

[ranocha]

Did you have a look at the numerical solutions or CFL magic?

[sloede]

Did you have a look at the numerical solutions or CFL magic?

Sorry, no, not yet. Do you think you can do it, or should we just merge now and check it later?

[ranocha]

We can merge now and check later 👍

[sloede]

If we're to achieve 97% coverage again, we need to cover 30 more lines. I am looking at it right now...

[sloede]

If this doesn't push coverage beyond 97%, I'll give up on this goal before v0.3... Note that some checks now already make use of the Taal way: Instead of doing only integral tests, I am adding a bunch of unit tests for many methods such as entropy or pressure. I hope that's ok - if not, I propose to proceed anyway and open an issue for it (or, of course, any of you should feel free to add integral checks instead).

[ranocha]

Thanks!

Do you think it would be helpful to add some comments like the following one to test_manual.jl?

Constructing indicators/controllers using the parameters below doesn't make sense. It's just useful to run basic tests of show methods.

[sloede]

Do you think it would be helpful to add some comments like the following one to test_manual.jl?

Constructing indicators/controllers using the parameters below doesn't make sense. It's just useful to run basic tests of show methods.

Yes, good idea 👍. I'll do so once the tests here have passed and before the merge.