How to skip the required arguments of the subcommand?

[chungquantin]

use external_crates::ExternalSubcommand;

#[derive(Args)]
#[command(args_conflicts_with_subcommands = true)]
pub struct Cli {
	#[command(subcommand)]
	pub command: Command,
}

#[derive(Subcommand)]
pub enum Command {
        /// The arguments in the `ExternalSubcommand` subcommand are required. 
	#[clap(alias = "p")]
	Subcommand(ExternalSubcommand),
}

Hi, I wonder if there are anyway to set a flag for that subcommand to avoid parsing checks? Let's say I want to check the arguments passed to the subcommand on the higher command level and handle the logic there, I don't want the error thrown for required arguments when I use the subcommand.

Note: I can't update the subcommand as it is from external crates.

One approach I found is to use ignore_errors but somehow the flag fails the --help command which I don't find the root cause yet.

[epage]

I don't see a general way of doing this. The big problem with this is with the definition of ExternalSubcommand: how do we fill in fields that errored out? For example, say we are working with this definition:

#[derive(Parser)]
struct ExternalSubcommand {
    #[arg(long)]
    name: String
}

What do we put in for name?

[chungquantin]

I see. Agree that it breaks the design of the overall program due to the missing field values. Alternative way could be handling the parsing checks on the runtime from the library user's side. Right now, with clap-derive, it panicks right away when the subcommand is run with missing arguments, I expect library user can check which is missing first before the main parsing check run. Are there any way to do that?

[epage]

I assume you mean "error" and not "panic"?

You can turn the struct into a Command, parse arguments, introspect ArgMatches, and then instantiate the struct as desired.

[epage]

I should add, you can enable ignore_errors on the struct to deal with required arguments or walk into the Command and mark the args as not-required.

[chungquantin]

Thank you. Yes but enabling ignore_errors right now causes an issue with --help as mentioned here #5919. And because it is from external crates, I can't update anything related to the command without forking.

But I agree the ignore_errors is the perfect solution here. While you're here, may I ask are there anyway to print the help output manually to bypass the current issue?

[epage]

Command has functions for printing help.