Raise warnings when for DNS events that are not raised due to p_all_addl #4155
Conversation
…p_all_addl By default, dns_skip_all_addl is set to false. This causes several events to not be raised. This change emits warnings when a user defines event handlers for events that will not be raised. Furthermore, it adds notes about this behavior to the documentation. We also introduce a new BIF, `is_event_handled`, which checks if an event is handled. Fixes GH-4061
testing/btest/Baseline/scripts.base.protocols.dns.event-handler-warning/.stderr
Show resolved
Hide resolved
There was a problem hiding this comment.
Approving. There's one more test failure about ZAM BiF tracking.
Some background how to fix is here: #4105
@vpax - could we treat unknown BiFs as having side-effects by default and keep some maintenance list outside of the btest infrastructure?
I'm also wondering if annotating functions with &pure or &no_side_effect or &idempotent instead would be a bit easier. Not just for us, also for new contributors. In the future, BiFs added by external plugins could also inform ZAM about their attributes/properties this way.
I think that's indeed how they'll be treated. The BTest failure will occur but not the mis-use. Regarding
Likewise sounds fine as long as it's pretty streamlined to figure out when to update it. (That might wind up being equivalent to the current BTest approach, but if you have thoughts for how to do it more readily, great!) |
This commit addresses review feedback for DH-4155. Furthermore it fixes test failures, and adds a new test for the is_event_handled bif.
0xxon
force-pushed
the
topic/johanna/gh-4061
branch
from
d9bcfee to
6bfa559
Compare
By default, dns_skip_all_addl is set to false. This causes several events to not be raised. This change emits warnings when a user defines event handlers for events that will not be raised.
Furthermore, it adds notes about this behavior to the documentation. We also introduce a new BIF,
is_event_handled, which checks if an event is handled.Fixes GH-4061