required positional arguments and a vector positionals

[phlptp]

This PR adds a loop on the positional parsing to check for remaining required positionals and prioritizing them if positionals_at_end is activated.

it addresses issue #295

[codecov]

Codecov Report

Merging #306 into master will decrease coverage by 0.06%.
The diff coverage is 91.3%.

@@            Coverage Diff             @@
##           master     #306      +/-   ##
==========================================
- Coverage     100%   99.93%   -0.07%     
==========================================
  Files          12       12              
  Lines        2935     2954      +19     
==========================================
+ Hits         2935     2952      +17     
- Misses          0        2       +2

Impacted Files	Coverage Δ
include/CLI/App.hpp	99.84% <91.3%> (-0.16%)	⬇️

---

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 ba7aac9...e365c78. Read the comment docs.

[codecov]

Codecov Report

Merging #306 into master will not change coverage.
The diff coverage is 100%.

@@          Coverage Diff          @@
##           master   #306   +/-   ##
=====================================
  Coverage     100%   100%           
=====================================
  Files          12     12           
  Lines        2935   2954   +19     
=====================================
+ Hits         2935   2954   +19

Impacted Files	Coverage Δ
include/CLI/App.hpp	100% <100%> (ø)	⬆️

---

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 ba7aac9...f11a9f8. Read the comment docs.

[henryiii]

This looks good to me - I was going to suggest adding a test for two ending positionals but you already have one. 👍 @kelvinhammond, @WenbinHou, does this look good to you? The positionals_at_end setting is already there, this just modifies the way it works with an ending positional.

[phlptp]

I will add that it will ONLY modify current operations if there are required positionals and the positionals_at_end flag is activated.

[henryiii]

Do we need the positionals_at_end and required only behavior? Why would you ever want an unlimited positional followed by limited positionals to not match the limited positionals? Otherwise, they would never match.

[phlptp]

I think the biggest reason was to ensure the behavior worked and reduce potential ambiguities. If you didn't require positionals_at_end you could have regular arguments at the end of the args vector and the option would not trigger.

I think if you removed that, you would add a bit of overhead for regular positionals and might be confusing sometimes why it wouldn't work, but otherwise in the cases tested it should work equivalently. Right now there is a specialized case in which the check occurs, and only a bool check overhead in other cases.

It might be reasonable to expand it further but it might be worth considering clarifying and expanding desired rules for positional argument priority first. This topic also has some relationship with #297.

[WenbinHou]

So why this PR is not merged till now?

[henryiii]

I think we can go ahead and merge now, then address the "worth considering clarifying and expanding desired rules for positional argument priority first" as listed above.