Note:
Screen Shot 2017-02-07 at 3.42.50 PM.png is the Org config
Screen Shot 2017-02-07 at 3.43.06 PM.png is a single project's config

stephenconnolly this may be a legit regression - I can't remember previous behavior but I think it was always done at the folder level right?

Can you confirm if the configuration changes are applied when there is an index?

My theory is that the code that detects a configuration change and forces an index is not detecting this change, but an index should apply fix-ups.

I'd like to know what the issue I am chasing is:

• Index / scan not being forced when the configuration has changed; or
• Configuration changes not being applied by index / scan

Marking as incomplete to prod the reporter into responding (and re-opening or agreeing to close)

stephenconnolly It worked! But yeah it was as you guessed, that the trouble with scanning caused the regression. I haven't verified that by reproducing, but I have verified that when scanning works (thanks again for the github org patch/fix) the settings do indeed propagate as expected.

I think this can be considered closed, as when the system functions, it all works as expected.

morgan_goose I may create a secondary issue to investigate proactively applying configuration changes where those configuration changes have not changed the scm criteria (and hence will not need a scan)