Third party license changes for automated maintenance releases?

[jonathanolson]

Whenever I deploy automated maintenance releases, it looks like it is spamming commits to https://github.com/phetsims/sherpa/commits/master. Can this step be postponed until the end of the deployments, so it is one commit?

Additionally, it looks like whether a version is reported for a sim is inconsistent in the report:

This report is for the following simulations:

‪Acid-Base Solutions‬ 1.2.14
‪Area Builder‬ 1.1.9
‪Area Model Algebra‬
‪Area Model Decimals‬
‪Area Model Introduction‬
‪Area Model Multiplication‬
‪Arithmetic‬ 1.0.13

[pixelzoom]

I recently published RPAL 1.2.0 and Molarity 1.4.0 and 1.4.1 (not automated maintenance releases). I see a commit to third-party-license.json for Molarity 1.4.0, but not for the other two. What performs this commit? What went wrong?

[jonathanolson]

Hmm, it should run at the end of every production deploy (after sending the build to the server).

Maybe it errored (was an error printed? do you have console history still of the deploy)? It's also possible to CTRL-C (or equivalent) out of the deploy at this point that would interrupt it.

[pixelzoom]

I received a "Build Succeeded" email from the build server. I don't recall seeing any console error, and I no longer have console history. And I didn't abort the build before it completed -- grunt production explicitly warns against that, as I recall.

[samreid]

+1 for batching sherpa commits. Or maybe we can git squash them later, if it is too awkward to batch the tasks. I don't know why some sims have no version noted.

[pixelzoom]

@samreid said:

I don't know why some sims have no version noted.

See phetsims/chipper#593.

[jonathanolson]

I'll modify it so it runs once a day on bayes.

[mattpen]

@jonathanolson - What is the status on this issue?

I recently had to run this manually for Graphing Quadratics to solve #118.

[jonathanolson]

Not yet implemented. Let me or @ariel-phet know if this would be a priority.

[mattpen]

Bumping this issue due to phetsims/wave-interference#339. @ariel-phet - I'v had to run this 3-4 times manually since the chipper 2.0 release. It generally takes me an hour or two. I could pretty easily put this in a script that runs daily on phet, which would take 3-4 hours to implement and test.

[jonathanolson]

This also came up in #132.

[jessegreenberg]

+1 for this, I didn't run into an error but it adds a substantial amount of time when running Maintenance.deployProduction in the final steps of the maintenance release process.

EDIT: I am not sure if this issue just to to postpone/collapse commits to sherpa, but it would be ideal if we only generated the third-party report once all sims have been redeployed in the batch.

[ariel-phet]

It would be good to either polish off this issue or close it if it is no longer a priority.

[jonathanolson]

@zepumph handled this in phetsims/chipper#594, closing.