Closed samreid closed 6 years ago
Discussion in #232 says there was an RC that was close to publication and we were deliberating whether to use RC SHAS or not. @jessegreenberg can you comment on whether the RC included LOL testing? If not, can a quick test or two reveal whether this sim (in the RC SHAS) exhibits touch problems like the ones we saw for build an atom? If the RC SHAS are problematic, is master problematic?
FAMB 2.3.0-rc.2 passed RC testing here: https://github.com/phetsims/QA/issues/11
FAMB 2.3.0-phetiorc.2 has not been completed: https://github.com/phetsims/QA/issues/12
2.3.0-rc.2 was deployed at the same time as 2.3.0-phetiorc.2 as part of the release strategy outlined in https://github.com/phetsims/chipper/issues/587. RC SHAs are not problematic, but to keep both sets of SHAs aligned both versions need to be deployed at the same time, and fixes required for the 2.3-phetio branch will need to be applied to the 2.3 branch.
2.3.0-rc.2 was built Tue Jun 27 2017 and did not include testing for LOL. The fix for the issues in build an atom were committed afterward in https://github.com/phetsims/scenery/commit/e75c1c55981e54cb87e77a430a9fcee66362ddb6 on Jun 29 2017.
To verify, I tested 2.3.0-rc.2 in the LoL test harness, and was able to reproduce the same problems with dragging that we were seeing in build-an-atom.
I am unable to reproduce the problem in master. So I think to support LoL we should rebuild with master and do a new RC test including LoL testing.
@samreid @zepumph do you think that we should coordinate the 2.3 and 2.3-phetio releases again? This would be in line with our decision in phetsims/chipper#587, but it could take quite a bit longer to deploy a LoL version.
If 2.3-phetio doesn't need to pick up scenery changes, we could start fresh with a 2.4 branch and let 2.3-phetio proceed as it is (not coordinating the two).
I think the simplest solution to get both released as quickly and effectively as possible is to go ahead with 2.3-phetio, and take master for 2.4, The solution for the phet-io bug https://github.com/phetsims/forces-and-motion-basics/issues/232 can be cherry-picked into that branch. There is an argument for releasing both as 2.3/ 2.3-phetio from current master, because of less branching confusion.
I don't have a preference. Whatever is easiest for QA.
Thanks @zepumph, I agree. Though I might recommend replacing 2.3 with a 2.4 branch so it is clear that it has separate SHAs from 2.3-phetio.
Edited above, I agree. That was a typo!
Sounds good! @samreid if you agree, should I proceed with an RC for FAMB now?
Cherry picking https://github.com/phetsims/chipper/issues/587 into 2.3-phetio
seems reasonable, but it looks like https://github.com/phetsims/QA/issues/12 still has a lot of testing to do for the phet-io version.
I can see that the non-phet-io version passed RC testing and that's great, but it may be inefficient to fully test the phet-io version now then again in a week or two for the LOL release. On the other hand, getting phet-io through testing now probably means less bugs or easier testing for 2.4 LOL time.
I don't think I can make the call, it seems like it should be up to @kathy-phet or @ariel-phet.
Sounds good @samreid, Ill bring it up for discussion at tomorrow's status meeting.
I emailed @ariel-phet (because I think he may be out tomorrow)
Hi Ariel,
Jesse and I had a question about scheduling testing for Forces and Motion: Basics. Kathy wanted to schedule Legends of Learning for the near future (which requires master), but the previous version which has passed QA for non-phet-io is buggy for Legends of Learning and still needs to go through phet-io QA. We weren't sure whether to test the buggy LOL one for phet-io and then test another version for LOL later, or jump to master and only do one round of phet-io testing.
Please comment here or in https://github.com/phetsims/forces-and-motion-basics/issues/235
Best Regards, Sam
@samreid I say bite the bullet and go to master.
OK thanks @ariel-phet. I'll leave it up to you @jessegreenberg
Thanks @samreid, but just to clarify, the decision is to build 2.3 and 2.3-phetio off of master and deploy both at the same time, synchronizing their release and SHAs?
I think that's been our consensus for deploying sims that have phet-io (though I have to admit there has been a lot of churn in those ideas and I'm not 100% sure of our final decision). It seems the discussion was mostly here? https://github.com/phetsims/chipper/issues/587
Alright, thanks. I will build new RCs for 2.3 and 2.3-phetio off of master once these issues are completed:
A reminder to test for LOL features in the next RC, unassigning self.
Icons are located here: https://drive.google.com/drive/folders/0B3v_ZP9Kq36CQnVYRnZYaDRhVkk
I created an RC for this issue here: https://github.com/phetsims/QA/issues/42
This RC is for the phet-io version first, when that RC passes, we will start a new one off of the same SHAs.
2.3.0-rc.4 passed testing, we are ready for a 2.3.0 RC.
EDIT: this will be 2.3.0-rc.3
We will need the fix in https://github.com/phetsims/joist/issues/447
@jessegreenberg as identified in https://github.com/phetsims/joist/issues/450 we will need a joist change before this can be published for LOL. Can you include https://github.com/phetsims/joist/commit/63046a699e8e3994f9c40f8d93482a497f0316b3 and create a new RC?
Please see also https://github.com/phetsims/joist/issues/449#issuecomment-340263119, we are considering picking up some other commits for this maintenance release.
Sure, sounds good. I read through phetsims/joist#449 and it sounds like there is still some work to do there so I will wait to cherry pick commits related to that issue and create a new RC until that is settled.
Actually, the version of joist and LegendsOfLearning.js did not have sim.boundRunAnimationLoop();
, it was added after the rc for FAMB was created. I will try to verify that phetsims/joist#447 is not a problem for FAMB.
From https://github.com/phetsims/joist/issues/447:
CCKDC stayed at 60 fps through multiple pauses and resumes and didn't climb to 60000 fps.
I pressed pause many times and pressed pause then resume many times and did not see fps ever go above 60, and since sim.boundRunAnimationLoop()
isn't being used on the 'pause' event we shouldn't need any patching for 2.3.
EDIT (for phetsims/joist#447, will still need changes for phetsims/joist#449)
May also need https://github.com/phetsims/joist/issues/458 ?? It's a question.
Thanks! Understood.
It looks like phetsims/joist#447 and phetsims/joist#449 and phetsims/joist#458 are all closed, Ill review and see if I can determine what needs to be picked up.
Grabbed all updates, verified they are in both 2.3 and 2.3-phetio branches of this sim. Once https://github.com/phetsims/QA/issues/57 is complete, we can spot check and then redeploy.
Testing is good to go for this.
yay!
FAMB 2.3.0 was deployed with testing and support for LoL. Closing.
@kathy-phet said
@amanda-phet we will also need icons for this.