feat(jest-message-util): Add support for Error cause in test and it

[dubzzz]

Summary

For the moment, Jest only supports Error coming with causes if the Error was thrown outside of any test or it. The current PR adds support to test/it cases.

Test plan

Various tests have been added on the PR to show what it will look like on various configurations of Errors coming with causes.

[dubzzz]

Here is an example of report:

For the code snippet:

The code snippet is the one that has been used to build the test cases.

[dubzzz]

And here is something close to the second test case:

[SimenB]

thanks for working on this! can you also add integration tests to failures for errors

• outside of test
• in test
• in hook

?

Also, please add a changelog entry 🙂

[dubzzz]

can you also add integration tests to failures for errors

Is there already a file with such integration tests for normal error cases or should I create a new test file?

[SimenB]

can you also add integration tests to failures for errors

Is there already a file with such integration tests for normal error cases or should I create a new test file?

https://github.com/facebook/jest/blob/main/e2e/__tests__/failures.test.ts

[dubzzz]

I'll need to fix error with cause tests for node 14. I already had the issue in my first set of tests, I'll update the PR tomorrow with a working versions of the tests for node 14.

On a pretty close subject, @SimenB do you think we should add support for error with cause into expect(...).toThrowErrorMatchingSnapshot(). I can possibly attempt to work on it in another PR if you do think it can be useful.

[SimenB]

On a pretty close subject, @SimenB do you think we should add support for error with cause into expect(...).toThrowErrorMatchingSnapshot(). I can possibly attempt to work on it in another PR if you do think it can be useful.

That makes sense to me - toThrowErrorMatchingSnapshot is supposed to stringify an error, and its causes seems a natural part of that to me. Might be a breaking change, but happy to land in Jest 30.

[dubzzz]

Not sure how I should deal with the last error remaining for jest-jasmine. It seems that my last change trying to handle results coming from jasmine and accessing to failure.error somehow changed the behaviour we see in another test 🤔 I'll re-dig a little bit more as I was not expecting such regression and definitely it's linked to my last update attempting to fix part of jasmine runs.

[dubzzz]

That makes sense to me - toThrowErrorMatchingSnapshot is supposed to stringify an error, and its causes seems a natural part of that to me. Might be a breaking change, but happy to land in Jest 30.

Making this PR pass first, then I might spend some time on it 👍
Actually I discovered them yesterday while writing tests using cause on fast-check side.

[SimenB]

I'm happy to exclude jasmine for now if it proves troublesome

[SimenB]

this is hard to read - could you extract to a helper function with returns?

[dubzzz]

Definitely, will do

[dubzzz]

Ok, I'll drop jasmine support for now then. And retry for it later.