Glen Mailer

That wouldn't work deeply - it might be best to fix this upstream in Immutable, Lee has since said he'd be ok with that.

They're similar, but I'm unsure if they have all the methods. I also don't really use records, this only came up because of a PR into my transit lib -...

@renanvalentin This isn't really related to this repo, as you're using `Immutable.is` directly - so its all code in immutable-js itself. The behaviour you're seeing is correct though - normal...

Possibly, I believe `test.chuck` in ClojureScript has implemented this now so exporting it across might be fairly doable.

In case anyone wants to quickly apply this manually, you can try the following - which worked for me on OS X. You'll need to adjust the paths as appropriate...

Hrm, I must have imagined that this worked before - probably another test runner! Do we think it's too much of a breaking change to continue onwards in the face...

Are we in agreement that if a failure in an after hook is identified as an assertion failure, that the testsuite should continue?

@kownacki - that's a reasonable suggestion IMO, do you see problems with the current suggestion of using `AssertionError`?

We can bikeshed on the name, I was more interested in thoughts on an explicit opt-in, vs implicit by exception type

From an implementation POV, i'd have that API create a hook using all the same objects as normal, but with an extra flag. I think both options as as easy...