Joseph Musser

I'd say applying to all methods, whether tests or not, makes sense. Before if a setup took too long and the fixture had a timeout specified, it would time out,...

This ties back to https://github.com/nunit/nunit/issues/3283 and I do think we'll find it's tied to https://github.com/nunit/nunit/issues/3740 as well. This is CallContext (thread-local semantics) not LogicalCallContext (async-local semantics), so I think the...

Yes, I do. That would be welcome!

I like the CancellationToken idea. It would also be idiomatic .NET (and cool) if it could be injected as a test method parameter, but that could be done later.

I think we have wanted to deemphasize the \*Assert classes in the past in favor of the constraint model. Can you give examples of the kind of assertions you would...

Great, thanks for the examples! If the team decides in favor of this, we'll probably have it up for grabs since our bandwidth is needed elsewhere. I think we would...

I do agree with Chris and Charlie. Exact equality of all these types already works, but shall we open a separate issue to enable tolerance to work against this list...

I'm in favor of adding tolerance capability for all of the scalar types. If we could come up with a consistent definition of tolerance for the vectors, I'd be in...

So depending on the use of the vector, someone might want tolerance to mean the vectors refer to points in euclidian space that are within the tolerant distance of each...

@vpenades That makes sense. I'm still not sure we're talking about the same thing for one item: > Is the vector A equal to vector B (with tolerance) For this...