c-blake

Sorry, but "what no shell uses" != "useless." The doc for `parseCmdLine` never said "it works like a shell" whatever your personal expectations might be. Also, shells are ***far*** from...

The new exception raising is the real problem, as immediately identified by @Araq and already explained in detail, not any qualms you have with my example that produces such raising.

It has been amply explained why `parseCmdLine` need never be deprecated. The inverse relation you seem fond of will never be true in reverse { quote(parse())=identity } since parsing erases...

> a new shell command parser can start its life as a Nimble package... As mentioned there is one already: https://github.com/SolitudeSF/shlex even named similarly and which already does "future work"...

Incidentally, what I mean by cursor-oriented design would look like https://forum.nim-lang.org/t/5697#35810 where a cursor is a tree ``Path``.

Something I don't think I have ever mentioned yet though also relevant here is that separation of seek & edit (add or del) is also a good way to provide...

Of course, these (seek,edit) things could be the "expert" interface and considered an "implementation detail" and pairs could be bundled up into the "all-in-one" operators like `[]=` and friends which...

Well, I have an impl over in https://github.com/c-blake/adix of this. The usual B-Tree invariants are violated in the course of this initial phase and require a final touch up at...

I found it compiling a big chunk of The Nimbleverse to test https://github.com/c-blake/nimp. Nimble uses latest tagged releases while nimp uses #head all the time, really..(or whenever the installer last...

EDIT: @arkanoid87 wrote (but then deleted, I guess): >I'm not a fan of cligen and I use cliche for exactly this kind of issues [..]The cligen code feels hacked in...