Use Node.js builtin

[Richienb]

Could util.parseArgs be used instead of yargs-parser when targeting Node.js 18?

[sindresorhus]

Probably not, at least not right now. First of all, the API is experimental, but even after it becomes stable, it would be good to wait a few years to let it stabilize. Second, yarns-parser has a lot of options and is more configurable.

I'll keep this open for discussion.

[voxpelli]

Second question would be:

Should util.parseArgs be used instead of yargs-parser when targeting Node.js 18?

The fact that Node.js decides to bundle a module doesn't make it automatically better than all the other modules, so I think one still would need to evaluate it against the alternatives.

If util.parseArgs proves to be the better choice, then when Node.js 18 is the oldest supported LTS that could be a sensible choice to switch to.

Before that and one would need to po(n/l)yfill it

[shadowspawn]

Update: parseArgs now also available in Node.js 16. Still experimental.

Stand-alone package (ponyfill?) at https://github.com/pkgjs/parseargs but only one dependent on npmjs so far and not much interest expressed to date. Longterm maintenance not certain but is being actively maintained at the moment.

Second, yargs-parser has a lot of options and is more configurable.

Yes! By design, parseArgs does not have much configuration. The "big switch" is strict: true to throw for possible cli user error, or strict: false to auto-detect options and leave it up to calling code to decide what to do.

[bcoe]

👋 I have been absent from OSS these days, but wanted to provide my feedback (as the maintainer of yargs-parser, and a contributor to util.parseArgs).

I would love if util.parseArgs could eventually be a replacement to yargs-parser, I think to get there what would be needed would be a shim layer that fills in some of the missing behaviour (probably this shim layer would be a better replacement in meow than util.parseArgs directly).