Closed (fixed)
Project:
Features
Version:
7.x-2.x-dev
Component:
Code
Priority:
Normal
Category:
Task
Assigned:
Unassigned
Reporter:
Created:
16 Jun 2015 at 13:54 UTC
Updated:
19 Oct 2015 at 14:14 UTC
Jump to comment: Most recent, Most recent file
Comments
Comment #1
heddnThis causes errors for multifield, because the nested fields do not exist yet on features revert.
Comment #2
heddnComment #3
mustanggb commentedDoesn't break anything for me and fixes things that weren't working before.
Comment #4
mpotter commentedSo, the problem with this patch is it causes all existing field base features to be marked as overridden.
This is different from #2419479: Remove 'foreign keys' from field base which was done in response to a change in core that was causing overrides.
We need a different solution to this that is backwards compatible and doesn't cause overrides. People shouldn't need to re-export features for something minor like this, and it's an issue for distributions that would need to be updated to re-export features.
Comment #5
heddnIs there a function in features I could call to check if the indexes already exist in a features export? Then only remove the indexes if they don't already exist.
Comment #6
mpotter commentedThere was also some functionality added recently to add an "ignore" list of keys, so maybe just adding indexes to that list would fix it. Look at the bottom of features.module for the implementation of the current ignore hook.
Comment #7
heddnNo interdiff, as this is an entirely different approach than #2.
Comment #8
mpotter commentedSet the status so this gets tested via the testbot
Comment #9
osopolarPatch from #7 works for me (applied together with patch from multifield-issue #2102265: Features integration, comment #32).
Comment #10
brunodboThe patch from #7, together with patch #32 from #2102265: Features integration, work for me as well. I'm now able to deploy multifields from local to staging.
However, the feature that contains the base multifield and base subfields ends up being overridden, no matter how many times I try to revert. It was in 'Needs review' state at first, and then after reverting, it went into eternal 'Overridden' mode.
Here's the part the feature reports as being overridden:
I suspect that it has something to do with this issue, but not 100% sure. Any ideas how to get the feature to revert?
Comment #11
mustanggb commented#7 didn't work for me for the reason described in #10, however this one does, had to fix a bug in
_features_remove_ignores()as well.Comment #12
heddnCan you provide an interdiff?
Comment #13
badrange commentedPatch in #11 worked for me. Thanks MustangGB!
Comment #15
mpotter commentedThis is looking good now. Especially the bug fix to _features_remove_ignores(). I added an additional unset($value) which is just best practice after any foreach using &. Committed to 7c288b1.