An option that allows unknown flags to be included in a varargs?

[henderea]

I have a script I'm converting from ruby that allows me to run arbitrary shell commands in a special mode. I define the command as something along the lines of $0 <cmd> [args...], but if i pass an unknown flag as part of args, it gets taken out and treated like a regular flag. Is there any way I can set my script to skip parsing and validation of unknown flags?

Basically, I want a mode that treats unknown flags as regular arguments. If I run script.js ls -la ., it should have argv.cmd be 'ls' and argv.args be ['-la', '.'], rather than putting something in argv.l and argv.a. I really like the functionality of yargs, but the fact that it parses out unknown flags is making it unusable for my purposes.

I haven't managed to find anything for this by looking through the documentation and code, so does yargs not support this? And if not, can support be added?

In case you would like to see my code, it can be found at https://github.com/henderea/jse, specifically the sudo subcommand.

[henderea]

Has anyone had a chance to look at this issue? I should be able to have a varargs include undefined flags as-is. If the argument -l falls inside a variadic argument and isn't defined in my code, it shouldn't fail the strict check and should instead just stay in the variadic argument.

Perhaps something similar to .strict() to enable this on demand instead of by default would be the best option. Maybe call it .literal()?

[sastraxi]

Running into this too -- trying to implement the following and there's no way to do it without this feature as far as I can tell. sastraxi/pgsh#36

[bcoe]

@sastraxi @henderea I'd be open to adding this as an opt-in feature, something you configuring using .parserOptions.

I don't have too many cycles for feature work lately, but would happily accept a PR if you're interested in giving it a shot.

I can be found in chat here when I'm doing OSS on weekends:

http://devtoolscommunity.herokuapp.com/

[henderea]

@bcoe I've submitted a PR to the yargs/yargs-parser library (yargs/yargs-parser#181). My new code has 100% test coverage, and all tests are passing. After looking into the code, I realized that the change needed to be put in the parser library, rather than the user-facing library.