feat(no-tag-publish): add proposal for a no-tag publish

[wraithgar]

This is a suggestion that grew out of a discussion surrounding #317.

It will afford a way for end-users to publish new versions of packages without having to set a dist-tag for the version being published.

The name of the parameter is still tbd in the Bikeshedding section.

npm publish --skip-tag

References

Related to #319
Related to #317

[ljharb]

--no-tag still seems like the best option; i doubt anyone's relying on that producing a false dist-tag (are there any packages with one in the registry?)

[wraithgar]

--no-tag still seems like the best option; i doubt anyone's relying on that producing a false dist-tag (are there any packages with one in the registry?)

We can't audit private packages for this anyways.

[ljharb]

Surely private packages on the public registry can be audited for this?

[wesleytodd]

Even if they are, I would consider that behavior a bug not a feature. If they want to have a string value of false they should be using --tag="false" IMO.

[ljharb]

@wraithgar this is still a very valuable feature, can this be reopened?

[wraithgar]

The npm registry itself doesn't support it.

[wraithgar]

Also this is in line with our new rfc approach that it's something someone owns and is actively going to implement.

[ljharb]

and there's no opportunity to get it onto the registry's roadmap?