Migrating to Clap

[gierens]

So far the builtin argparser has done it's job well I'd say or at least hasn't prevented us from implementing any features as far as I'm aware. With some fuzzy searching through the code it's also fairly easy to understand how to add a new flag and such. That's why switching to something more mainstream hasn't been a huge priority thus far.

Having said that I do however agree with those raising voices that in the long run it might be beneficial to move to clap since it's sort of the standard nowadays, well understood and documented and very powerful especially with the macros. This will require some heavy re-engineering though, and we got to be careful not to produce a mess but a very clean and easy to review PR. Just looking at src/options I feel like this might require more than one person. Therefore let's discuss the details here.

[daviessm]

I'd be happy to help. So the first discussion point would be - derive or builder?

[cafkafk]

Personally, I prefer builder somewhat. But I have no good reason beyond just liking it more. I'm fine with derive thou.

[MartinFillon]

Derived looks easier to implement and to read

[PThorpe92]

I've always gone with builder myself, but either way it's a major convenience upgrade for the code base. I feel like whoever makes the first solid PR can choose lmao.. it's not going to be simple from the looks of it.

[MartinFillon]

There is one small issue that I just encountered, we can't ovewrite -h option as they reserve it for help but we need it for --header option

[PThorpe92]

That actually is probably more of an issue than it may seem at first... Myself, and probably many many others have the -h flag in my alias's on both my computers. Breaking that would be pretty tragic and probably upset a few people.

Looking into this

[MartinFillon]

#197 -- POC of the parser

DO NOT MERGE IT

[sbatial]

This would also make #37 much easier/possible at all I think which would be nice!