
Problem/Motivation
When we deprecated Place blocks, we added a trigger_error() to the module file itself, but this makes it complicated to handle the deprecation in tests.
see #3084472: Remove Place Blocks module and #3062302: Properly deprecate the entity reference module.
Proposed resolution
Remove the trigger_error() from the .module file, leaving the rest.
Remaining tasks
User interface changes
API changes
Data model changes
Release notes snippet
Comment | File | Size | Author |
---|---|---|---|
#2 | 3116399.patch | 590 bytes | catch |
Comments
Comment #2
catchComment #4
kristen polThanks for the patch.
1) Changes seem fine.
2) Patch applies cleanly to 8.9 but is not relevant to D9 as the module was removed.
3) Searching for deprecation in
block_place
and they are still in:which is very thorough.
4) Tests pass.
5) Moving back to 8.9 and marking RTBC though I'm not sure about manual testing.
Comment #5
xjmI don't have a terribly strong opinion about this, but is un-deprecating the module file this late actually worthwhile?
Comment #8
xjmDiscussed with @catch whether this was still worth backporting -- per @catch, we worked around the problem in core by changing tests.
The fact that it was disruptive for core though means contrib/custom tests might have the same problem, and since this is the only model of a whole-module deprecation so far, maybe it's worth also making that example less disruptive. I filed #3135100: [policy and meta] Provide a proper mechanism for deprecating modules and themes to figure this out properly before D10.
Un-deprecating something isn't really disruptive; it just means one less error in Upgrade Status etc. when there are still many errors. I also backported this to 8.8.x because most modules will be trying to bridge compatibility from 8.8 to 9.0, so un-deprecating only for 8.9 doesn't help them really.
Committed to 8.9.x and 8.8.x. Thanks!