[FIX, FEATURE] Fix initialising the version_check_option. Add multi-level argument parsing.

[smehringer]

Resolves #1177

[codecov]

Codecov Report

Merging #1185 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1185      +/-   ##
==========================================
+ Coverage    97.3%   97.31%   +<.01%     
==========================================
  Files         219      219              
  Lines        8803     8832      +29     
==========================================
+ Hits         8566     8595      +29     
  Misses        237      237

Impacted Files	Coverage Δ
...lude/seqan3/argument_parser/detail/format_help.hpp	90.45% <100%> (ø)	⬆️
...clude/seqan3/argument_parser/detail/format_man.hpp	97.61% <100%> (+0.05%)	⬆️
...lude/seqan3/argument_parser/detail/format_base.hpp	89.06% <100%> (+0.72%)	⬆️
include/seqan3/argument_parser/argument_parser.hpp	98.54% <100%> (+0.23%)	⬆️
...lude/seqan3/argument_parser/detail/format_html.hpp	92.47% <100%> (+0.08%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c376e88...f3a8f3c. Read the comment docs.

[cpockrandt]

It seems that if you pass multiple special formats (e.g., h and hh), it will overwrite it in the next loop iteration. I think we should either print an error message or always take the first one. Currently it would choose the last argument.

[h-2]

How does that work with nested argument parsers? For nested arg parser the version check argument should refer to the nested inner arg parser if it comes after the keyowrd. I suspect these changes will break that.

We really need test-cases for nested arg parsers as it is a commonly used design.

[rrahn]

Sorry to jump into this but what is a nested argument parser? Do we have a snippet or demo/tutorial for this? I can only think of a tool with different configurations who have their own options, e.g.samtools.

[h-2]

Sorry to jump into this but what is a nested argument parser?

A program with different sub-programs, like git or lambda (equally famous tools ☝️ ). Each sub-program should have their own help/man-page et cetera...

Do we have a snippet or demo/tutorial for this?

@smehringer said she would write a HowTo at some point (or at least we agreed that we want such a HowTo). Currently we have nothing that tests this, that's why I said new changes threaten to break it.

[rrahn]

Ok, this is what I had in mind already.

[smehringer]

Code coverage complains that format man is not tested 🙄

[h-2]

In general: is "multi-level arg parser" the best name? Does this pattern have a name that other projects / arg-parsers use? Maybe "nested" is better?

[smehringer]

@h-2 why did you remove your review? 🙀
(I already sais the code coverage fails because of the untested man format, if this is the case)

EDIT: ok ok the ONE comment

[h-2]

I remove my review request not because I am pissed, just so that I know which of the items in the list is actually waiting for me and not me waiting on the other person 😉

[smehringer]

I know, I just thought it is because of the code coverage that bites me any time I do something on the argument parser 😇