cli: add support for passing command-line arguments to the REPL and script environments

[attie]

As discussed in #glasgow, it would be nice to make command line arguments available to the REPL and script environments (especially the latter).

This patch adds support for this, with a small example script.

[whitequark]

I'm not sure that's what the discussion asked for? They were talking about reusing the pixel format arguments, which are probably already available from iface.

Also this silently ignores excess arguments for scripts that don't parse args.

[attie]

They were talking about reusing the pixel format arguments

Huh, somehow I missed that detail ... the risks of quickly reading something on mobile and thinking about it lots 🤦‍♀️
This leads me to suggest passing all arguments to these environments, and they can pluck out the remainder if desirable.

Also this silently ignores excess arguments for scripts that don't parse args.

Granted, but only after a -- terminator...

glasgow repl uart -V 3.3    -m 5 # presents Glasgow's normal error
glasgow repl uart -V 3.3 -- -m 5 # passes in [ '-m', '5' ]

[whitequark]

This leads me to suggest passing all arguments to these environments, and they can pluck out the remainder if desirable.

Ew, no.

[whitequark]

Granted, but only after a -- terminator...

Ah, I missed that detail. That's fine then to merge for script; I'm not sure what point this serves in repl mode?

[attie]

This leads me to suggest passing all arguments to these environments, and they can pluck out the remainder if desirable.

Ew, no.

Where I've done this:

future = eval(code, {"iface":iface, "device":device, "args":args.remaining[1:]})

... I was suggesting doing this:

future = eval(code, {"iface":iface, "device":device, "args":args})

... giving the script access to Glasgow's args (e.g: pixel format from the WS2812 applet), as well as args.remaining if they want to deal with them.

Still ew?

---

Granted, but only after a -- terminator...

Ah, I missed that detail. That's fine then to merge for script; I'm not sure what point this serves in repl mode?

I think I agree re REPL mode as it's so heavily interactive... but perhaps changes in light of above?

[whitequark]

... giving the script access to Glasgow's args (e.g: pixel format from the WS2812 applet), as well as args.remaining if they want to deal with them.

Oh, no, this is fine. (That's actually what I would suggest!)

[whitequark]

but perhaps changes in light of above?

Yeah, that's clearly fine.