Preventing undefined callback from being passed into the forgotPassword Method

[hwillson]

This PR is intended to address #5676. Preventing an undefined callback from being passed into the forgotPassword Method helps avoid issues with the audit-argument-checks package (more details in issue). Thanks!

[hwillson]

Hmm - tests failed with:

Action failed: METEOR_PRETTY_OUTPUT=0 ./meteor --get-ready

I don't think this is related to my changes ...

[hwillson]

The build/tests passed using my own CircleCI account, so I pushed a minor formatting change here to force the CI build to run again. It worked!

[abernix]

It looks like CircleCI (and Travis too!) couldn't access the package server when that test ran. Not sure what happened there, but it does appear to be okay again now. Maybe it's just really busy today. You can ping me to trigger retries on the tests too if you need to. :)

Haven't had a chance to review this otherwise yet. ;)

[hwillson]

You can ping me to trigger retries on the tests too if you need to. :)

Great to know @abernix - thanks!

[abernix]

LGTM

[abernix]

My only suggestion on this would be to include issue number being fixed in the commit message and probably the added test description too (i.e. (#5676) on the end of the test name)

[hwillson]

Great, thanks @abernix - I've updated the test description. I'm always a bit nervous when it comes to updating older commit messages (force pushing really scares me!) - am I okay to leave the original commit message without the issue # for now? (I mentioned it in the last commit message). I'll be sure to include an issue # in future PR commit messages. Thanks!

[abernix]

Yeah, it's fine; I wouldn't worry about it this time. But good on adding it to the test name. And force pushing should absolutely scare you! 👻 Although, on PR branches where you're the only contributor, I think it's fine. I've definitely been known to git commit --amend and git push --force on a few PR branches. 😜

[tmeasday]

Thanks for the diligence here guys