Const-correct ArgumentParser

[Bedzior]

Hi!

I wished to provide a small contribution to your fine project.

In another project I am currently working on, I stumbled upon a minor inconvenience in argparse's interface. Normally after parsing the parameters one would expect the parser to be immutable (quite understandably so in simple cases). However, I realized a crucial function checking for a value being actually passed by the user is not, though could be, const!

I am concerned the test I wrote might actually be meaningless. What one would expect there is for the code to actually compile, nothing more, to validate this change. I could also offer rewriting some tests to try to better comply with ArgumentParser's immutability at some point in their bodies, if you deem that viable or useful in any way.

Cheers!

[skrobinson]

@Bedzior

I don't have any say whether this PR is accepted, but I offer a little feedback...

It looks like ::present may have been missed in commit 3734fc6 for #74 and should be const qualified.

Think of your added test as a failing test case that your change to ::present causes to pass. TDD at its best. The test failure is during compilation rather than execution, but it is still a test case.

Are your tests really related to value semantics? I think it would be more fitting to put your tests in a new file (e.g. test_const_correct.cpp).

[p-ranav]

Agreed with @skrobinson, probably best as a separate test file test_const_correct.cpp. Outside of that, would be happy to merge.

[Bedzior]

Hi @skrobinson, thanks for throwing an eye!

Think of your added test as a failing test case that your change to ::present causes to pass. TDD at its best. The test failure is during compilation rather than execution, but it is still a test case.

This was my actual thinking when writing this code. However, all checks done under THEN seem excessive, so I am still undecided.

Are your tests really related to value semantics? I think it would be more fitting to put your tests in a new file (e.g. test_const_correct.cpp).

This is my ignorance in relation to the scope of "value semantics". But from what I understand now,const-qualification has its own rightful spot.

@p-ranav it is done

[Bedzior]

@p-ranav
Is there anything else I could do here?