Issue #3349968: Trigger deprecation error if queue or exchange config does not have 'name' property set
Closes #3349968
Merge request reports
Activity
requested review from @cmlara
assigned to @cmlara
@cmlara i see your point but as long as we clean up after ourselves i don’t see a huge drawback unless im missing something?
changed milestone to %3.2.0
Personally I wish we could use personal name spaces, however D.O. has refused the ability.
Main reason to not use the project namespace is to keep it clean. Any code that code that is on the master repo should be a 'final version' not a work in project. While we are not packaging random branches as dev releases, its still 'shipped code' from a project standpoint which has inherent obligations on stability and security associated with it.
Beyond that, it holds a high likelihood of confusing D.O. users, who are all being trained to use the 'get push access' button and use a collaborative issue in the issue-fork namespace. With personal name spaces we could at least selectivity choose to grant 3rd parties access to a branch, however with the main project that would pose an unnecessary security risk.
added 3 commits
-
9c5ffd89...625753d1 - 2 commits from branch
3.x
- cc78effe - Merge remote-tracking branch 'origin/3.x' into 3349968-config-without-name-deprecations
-
9c5ffd89...625753d1 - 2 commits from branch
added 1 commit
- 3830680e - Ensure we have queue config before checking for missing names
added 1 commit
- 99de81fc - Add additional count for mixed supplied for foreach error in QueueBase
Sorry for the delay on this one.
Overall its good. While errors don't occur during saving config in production I wonder if we should backport the config schema in case a site has their own tests for a custom module.
I believe the ConfigSchemaChecker would error out in tests if the config option is set but schema doesn't exist.