Introduce Review and Release Guidelines

[ryanio]

Description

This PR introduces REVIEW.md and RELEASE.md, two new documents that specify review and release guidelines and procedures, adapted from @nivida's documents Review_Guidelines.md and Release_Guidelines.md.

Closes #3358

Checklist:

• I have selected the correct base branch.
• I have performed a self-review of my own code.
• My changes generate no new warnings.
• I have updated the CHANGELOG.md file in the root folder.

[coveralls]

Coverage remained the same at 88.596% when pulling ec23c67 on reviewAndReleaseGuidelines into 3684fc3 on 1.x.

[cgewecke]

This looks great 🎉

A technical detail that might be helpful to document is the publishing command for both the rc and formal releases - like "how to tag the rc on npm" etc.

Should the 1 week waiting period be considered flexible for emergency patches or releases?

FWIW there's precedent for this in the 1.2.6 release (see #3351). The consensus view there was to make the smallest change necessary to address the emergency while waiving the rc process (meaning many existing additions to master were excluded).

That's pretty conservative and might not be appropriate for every emergency though.

[ryanio]

A technical detail that might be helpful to document is the publishing command for both the rc and formal releases - like "how to tag the rc on npm" etc.

@cgewecke thanks sounds good. Creating the rc track in npm is still on my TODO list.

re: publishing command, how does this look?
npm dist-tag add <package-name>@<version> [<tag>]

[cgewecke]

@ryanio

npm dist-tag add @ []

Yes that LGTM ... do you have a view about whether lerna should be involved here at all? I think they have an npm tagging option as well.

[ryanio]

Yes that LGTM ... do you have a view about whether lerna should be involved here at all? I think they have an npm tagging option as well.

Seems reasonable, I'll have to investigate further. @nivida said lerna has been not-so-reliable in the past:

"mostly it works.. but there are a few times it doesn't"
-@nivida

so we can try it but we may need some backup options if the lerna command doesn't work as planned. I've commonly heard of others (@ricmoo, @alcuadrado) rolling their own lerna-esque solutions. I think if it fails partially through is where the headaches come in, since some packages would have been run and others not.

[holgerd77]

@ryanio Great stuff! 😄 👍

[nivida]

I think this could get changed to a normal process aligned with semver. I and Chris have done the big work already without any breaking change and plenty of new test cases :-)

[nivida]

We discussed the LTS and EOL of 2.x already with Embark, Travis, etc... If I'm correct did we said 2.0 stable release plus 12months == EOL 1.0

[nivida]

This was a list I wrote down quickly in some minutes.. I think this could get worked out to a real document.. idk :-)

btw..: Thanks Ryan for taking those drafts to the next step! 💪

[nivida]

overall and probably off-topic: My personal wish and the wish from the community was to have automated releases.. is this still an idea? or do we get at least a nightly version on npm?

[ryanio]

overall and probably off-topic: My personal wish and the wish from the community was to have automated releases.. is this still an idea?

I'm personally in favor of automated releases.

or do we get at least a nightly version on npm?

We can build a nightly version pretty easily now that we are running on GH Actions.

[cgewecke]

🎉