Num args

[Mehrbod2002]

Fixed #4507

PR will check argument which allow to have one value on argument and check if it's render before and pending list is empty then evaluate the arguments .
One tests fixes , which doesn't contain num_args on it too .

[Mehrbod2002]

Appreciate if help to solve Lint problem on commit .
Thanks

[epage]

Its not clear to me what this PR is trying to accomplish or how its related to #4507.

Besides clarifying in the PR, what would help is to have the first commit that adds a test that demonstrates the problem and then the second commit makes this change and adjusts the test to show the correct behavior. #4567 serves as an example of this pattern.

[Mehrbod2002]

Its not clear to me what this PR is trying to accomplish or how its related to #4507.

Besides clarifying in the PR, what would help is to have the first commit that adds a test that demonstrates the problem and then the second commit makes this change and adjusts the test to show the correct behavior. #4567 serves as an example of this pattern.

The problem is arg with capacity of 1 , Parser can't check the num_args attribute on struct or .num_args(1) , because it's defaults and Parser don't let loop to gather Pending on length of One and Render it with react .
So while checking the arg we can detect which one has num_args by 1 and id of arg rendered before or not , then allow to add it in matcher .

The tests configured before for it but we add new condition on it .

[epage]

Something is still not quite clicking with that explanation.

If this is not related to #4507 (which I am not seeing how this change would be), maybe we should start with an issue being created. The template has places for an example program, reproduction steps, and expected vs actual behavior. I think this would greatly help clarify the point you are trying to make.

[Mehrbod2002]

use clap::Parser;

#[derive(Clone, Debug, Parser)]
pub struct CliArgs {
    #[arg(num_args = 1)]
    multi: Vec<String>,
}

fn main() {
    let CliArgs { multi } = CliArgs::parse();
    dbg!(multi);
}

Reference to #4507 , when we define #[arg(num_args = 1)] , the multi argument can even get more than 1 args .
Because loop which defined on Parser trying to catch all args then evaluate by pending , not one by one.
So if we try cargo run -- -- a b c d works without error now . (we except error) .

by this PR , now it will check if it for example multi value render and evaluate before , then it will be error with second argument on example ( cargo run -- -- a b c d on b ) and pass error . Argument with limit value to One not pass filter on loop of 'Parser' .
Max Value which indicate by num_args or #[arg(num_args = 1)] it's default so we check FlatMap too to watch out on rendered arguments .

Can check this example with current commit and this PR . just test it .

[epage]

I'm sorry, you are right. I was doing too superficial of a read of the Issue and was assuming it was a different one.

First, we need to decide if that issue is a bug or not. I'm not certain. The second step is deciding what to do. Solving this at the parser level is ignoring what the caller specified. This likely would change at the caller level but we were going with a fairly specific behavior that we got through combining ArgAction::Append with num_args(1..) and we'd need to find an alternative (which circles back to my "unsure I'd consider this a bug").

[Mehrbod2002]

I'm sorry, you are right. I was doing too superficial of a read of the Issue and was assuming it was a different one.

First, we need to decide if that issue is a bug or not. I'm not certain. The second step is deciding what to do. Solving this at the parser level is ignoring what the caller specified. This likely would change at the caller level but we were going with a fairly specific behavior that we got through combining ArgAction::Append with num_args(1..) and we'd need to find an alternative (which circles back to my "unsure I'd consider this a bug").

Thanks
I don't think this is a bug . the shortage in code we have I think and if we have this only condition to observe what user declare and respect to it , and I am ensure about this it's the best method of parsing but leak of this condition and this argument only we have .
Changing method for this can change a lot thing as I am observing code.

[Mehrbod2002]

@epage Please consider to look at this PR.
Thanks a lot .

[Mehrbod2002]

@epage
Thanks for your attention