search

meejah / fowl

Forward over Wormhole: streams over magic-wormhole Dilation connections
MIT License
28 stars 2 forks source link

cli: allow specifying custom Wormhole code length #7

Closed balejk closed 1 year ago

balejk commented 1 year ago

It might be nice to use the _Config default for the CLI default automatically, but I have never used attrs and frozen classes before so I do not know what the best way to do that would be (is it possible without instantiating the class?) and so I followed the example of --clearnet and set the CLI default to the same value manually.

meejah commented 1 year ago

Perfect -- yes, you should be able to add it to _Config (with or without a default). I'd be fine with making it required on the _Config class (so that there's only one place the "2" is specified -- in the Click/cli defaults).

Thanks!

balejk commented 1 year ago

Thank you for merging this!

Perfect -- yes, you should be able to add it to _Config (with or without a default). I'd be fine with making it required on the _Config class (so that there's only one place the "2" is specified -- in the Click/cli defaults).

Sorry, I am not sure that I follow: are you suggesting that the code length would be None in _Config and that it would always get set to the Click value (defaulting to 2 if unspecified on the command line) using evolve (which is really the case now already, except it's already 2 in _Config by default - so just changing this to None is what you mean?)?

meejah commented 1 year ago

Sorry, I hadn't read the code properly when I wrote that -- I was thinking that it could just be a required argument for _Config (so no default at all).

However, since --code-length is only on the "invite" subcommand the evolve call is probably right.

That said, if you're willing it might be nicer to make that a global option (and hence a required _Config option). I haven't done it here (yet?) but there's no technical reason why the fowl invite has to "create" a code and fowl accept "consumes" the code. (Basically I chose that because this seems easier to understand, and "usually" makes sense for the way I would use it -- but the protocol could allow the "invite" side to consume a code created by the "accept" side).

What I'm saying is: --code-length could make (technical) sense as a global option, probably.

(Note that I've sort of "cheated" with the _Config class a little since it's holding options that aren't strictly global ....)

Anyway, thanks for the PR and the evolve() call is fine too :)