assertRejects and assertThrows should return the caught error

[lucacasonato]

It'd be nice if these functions returned the error they caught, so users could do more intricate assertions on the error itself.

[bartlomieju]

Sounds good to me, this should be a simple change.

[mrkldshv]

Hey! I'd like to take this issue.

[bartlomieju]

@mrkldshv please do!

[kt3k]

assertThrows accepts errorCallback as the 2nd arg and you can assert on details about the thrown error in it.

I think we discussed similar suggestion before, but we didn't accept it based on the design principle assertion functions shouldn't return anything ref: #1052 (comment)

[kt3k]

(Note: we added errorCallback overload in #1219)

[mrkldshv]

Right, I've just looked through previous discussion. As pointed by @kt3k, it wasn't added because assertion functions shouldn't return anything. What do you think we should do? @lucacasonato @bartlomieju

[lucacasonato]

I think the "Assertions shouldn't have return values." rule is rather arbitrary. I think in this case returning an error is totally reasonable. the error callback does not really solve the issue for me, because I may need to do some async work with the error

[KyleJune]

In this previous issue I proposed a similar change to having these functions return the values thrown. I believe the fix for this could also be made to be a fix for that issue. Here is a link to my comment there. It would be good to keep error class and message arguments as a shorthand for both asserting that they throw/reject and asserting that it is an Error that was thrown.

#1362 (comment)

[kt3k]

I think the "Assertions shouldn't have return values." rule is rather arbitrary. I think in this case returning an error is totally reasonable. the error callback does not really solve the issue for me, because I may need to do some async work with the error

I'm open to changing that rule now as it's requested/suggested many times and it seems to have some practical use case.

@kitsonk What do you think about this suggested change? Do you have any concerns/objections?

[mrkldshv]

@kt3k Should we close this issue?

[kt3k]

@mrkldshv Yes, thanks!