access argv/command from coerce

[brewster1134]

i have a global option defined that all commands share, and need to use coerce for some additional manipulation. the issue is i need to access the argv object, or at least the command called, however only the option is passed in. this is likely a feature request, but maybe there is another way to accomplish what i am trying to do. here is a contrived example...

yargs
  .option 'foo',
    coerce: (foo) ->
      # need to access `build` here

  .command 'build', 'build', ->
    # i want to avoid having to declare a `foo` option here, as well as for every other command

[brewster1134]

i am also noticing that coerce is called twice for options defined globally... possible bug?

[aorinevo]

Hi @brewster1134, middleware gets passed a reference to argv. Perhaps that’s an option?

[BigBlueHat]

@brewster1134 your contrived example is exactly what I'm attempting to do. 😄 It mostly works as the option is set and available later in the command's handlers.

i am also noticing that coerce is called twice for options defined globally... possible bug?

It looks like yargs runs coerce on each alias also--which means if you have configDir as an option, it will run once for that and for config-dir.

However, I'm also noticing lots of excess coercion attempts where its value gets set a second time (with multiple additional runs) when it's used alongside a positional arg... Still trying to sort that out, though...

[BigBlueHat]

However, I'm also noticing lots of excess coercion attempts where its value gets set a second time (with multiple additional runs) when it's used alongside a positional arg... Still trying to sort that out, though...

This problem seemed to have been because I was using [] instead of <> for a positional, and also had added the --configDir [configDir] text to the usage line. Seems there's more magic hiding in the "command" text than I'd realized...