enable/disable usage section on user errors

[sebhoss]

I noticed in https://docs.rs/clap/latest/clap/_derive/_tutorial/chapter_3/index.html that errors raised through a custom value parser do not show a usage section whereas errors raised with a custom validator do show a usage section.

Is there any way to enable showing the usage section for errors raised by custom parsers or disable showing the usage section for errors raised by custom validators? My intention here is to make the UX consistent across error messages and I sometimes need to use custom validators and sometimes use value parsers to throw errors.

[epage]

The reason we skip usage in value parser errors is it seems like its unlikely to provide useful context to solve the problem.

As for custom errors, we don't know what is being done and whether a usage is useful context or not. So the question is how we can make cmd.error() not include usage.

Constraints

• We need access to a Command to add the usage
• We're directly injecting the usage into the message

Potential paths forward, requiring a breaking change

• Add a Error::usage(mut self, cmd: &Command) and stop including usage with `Command::error, requiring people to manually add it
• Track usage separately in a Context , allowing people to add/remove it as needed.

[sebhoss]

Thanks!

Would another approach be to check whether the usage string is empty and only print it if it's not? Then we could use command.override_usage("") to disable showing the usage section for custom errors.

[epage]

That seems out-of-the-ordinary use of that API that I would prefer not to do it.