Regenerate the README files.

[samreid]

While I was working on phetsims/phet-info#120 I tested regenerating a README file, and it showed me that the README files are now missing this line:

git clone https://github.com/phetsims/utterance-queue.git

[zepumph]

When I did the following commands, the utterance-queue line was added back in.

cd gravity-and-orbits
grunt published-readme

@samreid how would you recommend gathering a list of published repos to call grunt published-README on?

[samreid]

How about changing grunt update so that it automatically updates the README, based on looking at the versions in package.json?

[zepumph]

This should be done soon. I should have already gotten to it as all published readmes are broken. Elevating priority.

[samreid]

@zepumph this looks great, thanks! I made a few improvements in the commits. For dealing with the outliers in #5 (comment), I recommend adding a new package.json entry like "readmeTemplate:'...'" which is the relative path to the readme template to use. That way some repos can always use the "unpublished" template, and other sims can define their own template, which will be filled in with the template variables. Does that sound OK to you?

[zepumph]

That way some repos can always use the "unpublished" template, and other sims can define their own template, which will be filled in with the template variables. Does that sound OK to you?

I think this is halfway there. To me, when specifying something manually, you are opting out of both published and unpublished readmes, so I think instead there should be a package.json entry that tells grunt update to not try to auto generate the README.md file, since it is manual. I got to this solution via yours because in the outlier repos, it will involve having a template readme that is the exact same as the readme itself. I'll try a first pass and see what you think.

[zepumph]

The above solution is working really well so far, and will cover permanent cases like blast, but doesn't cover the case where a brand new sim is in RC but not published. Is it alright to just say that it is ok to have a published readme for that case (upwards of a month or two of inconsistency once per new sim)?

[zepumph]

Summary of work:

• Created readmeCreatedManually flag for package.json, and have grunt update respect it.
• Add it to all of the above list save number line integers, and add chains to it.
• run grunt update for all active sims and find that ISLC sims, though all have been published, had unpublished readmes (yay automation to find bugs).
• Committed the change from unpublished to published for number line integers. @samreid please especially let me know what you think about this one.

One cleanup question I had was questioning if we have a central place where the schema for a sim (or general) package.json is defined. Should we make this? Otherwise there is no place to document readmeCreatedManually but to look at the usage. What do you think?
@samreid please review.

[samreid]

A note for when I review this: in phetsims/tasks#1018 we discussed improving the readme files. I wonder if we will want them to be templates so we can fill in the git clone dependencies?

[samreid]

Looks great, thanks! I moved phetsims/chipper#913 to a new issue. If we decide to make READMEs' based on templates, that can be done in phetsims/tasks#1018, closing.