Am 24.11.2025 um 03:32 schrieb Marshall Clow:
After looking at the diff, I concur that the change is definitely not trivial.
Since you’re new, I’m going to pontificate here a bit - apologies if you’ve heard this all before.
No worries, better to be safe than sorry :-) I've closely read through the contribution guidelines and hope I understood everything.
The goal of the period between beta and release is to increase stability, fix bugs and polish documentation. Things that decrease stability (including, but not limited to breaking existing code) can usually be delayed to the next release (which is in 4 months).
Is the backMP11 back end new for the 1.90 release?
Yes, 1.90 is the first release containing the new backmp11 backend. This was my motivation for requesting permission to apply this late change.
Does this change affect other (existing) backends? (Since the bug was reported against 1.84, I think the answer is “yes”)
No, the existing backends are not affected by the change: They keep their existing mechanism to evaluate deferred events, and changing them would surely be unfit at this late stage. They are also likely to remain untouched since changing the mechanism there induces a high risk of breaking existing user code.
What’s the status of the test runners (and CI runs)? Are they all green?
Yes, the CI runs are green. I have written additional regression tests that address the fixed behavior in the new backend, all test cases pass.