Allowing multiple assertion failures per test

A mechanism for having multiple failures per test. In other words, an assert failing wouldn’t result in a fail-fast for that test. Any remaining assertions would execute, and they’d all be displayed.

https://github.com/darrenburns/ward/pull/256#issuecomment-869049499

The interface is still undecided, but I would like to see this be configurable down to a per-test, and possibly even a per-assert level.

[darrenburns]

Thanks for the feedback!

I like the ideas around using the assertion message and about agree about making it a plugin.

I agree that this might be a little too “magical” for the core lib. That, combined with the fact that it’s not going to be a use-case everyone needs, means that a plugin is probably the best place for this. In this way, you could “opt in” to the magic where it offers value.

[JoshKarpel]

Good point about the messages!

If it’s part of handling the assertion message, I wonder if this behavior can be deferred to a plugin? Then the core library just needs to expose a general purpose on_assertion_failure hook that can decide whether to end the test or not flexibly.

from ward_fail_slow import FAIL_SLOW

assert foo == bar, FAIL_SLOW