testing: introduce fail function

[43081j]

Introduces these two:

fail('some message'); // equivalent to assert(false, msg)
assert(throws(fn), true);

For now, throws returns a boolean and is expected to be used with assert just like equal.

If we want node compatibility, see #122 (throws does the assertion and throws the assertion error in node rather than returning a bool).

[43081j]

this file changed because i ran prettier across it FYI

[hayd]

I know this is part of #122, but perhaps we could discuss the renaming independently of this.

IMO assertThrows(fn, err?) throwing an assertion error (optionally of type/equal to err) seems desirable.

[43081j]

It depends really on what our aim is with naming of test functions.

If we agree on following the pattern of all assert* functions asserting, and all unprefixed functions returning bools then yes we should have an assertThrows (alongside the boolean unprefixed function).

It'd be three args in that case, (fn, err?, msg?).

If however we did end up following node and making all functions assert, there'd just be the one function with the assertion added inside.

[ry]

I think it makes sense to move asserts to the node-style API. I wrote the asserts lib - so biased.
However for now please rename to assertThrows.

Let's do the reorganization to a Node-style API later - and after others have had a chance to chime in (ping @piscisaureus)

[43081j]

Do note that this pr defines it as a boolean function to be used with the assert function.

So if we rename it, we also want it to throw like assert does, right?

[ry]

Do note that this pr defines it as a boolean function to be used with the assert function.

Remove it please

[kitsonk]

Sorry for the double PRs (#124) but I mention in the other PR that I already had something on my local machine because I was frustrated last night when writing some tests.

I think we should decorate assert with additional functions, as that is more common. My PR also has a few more features that brings it in line with something like Chai for the assert.throws().

[43081j]

The original reason for this PR still exists: a fail function.

Will rebase onto master when i get chance too.

Alsoo @kitsonk see #122.

[dsseng]

Interesting feature!

[43081j]

anyhow, updated this to be just fail in case you still want it.

[ry]

I'm going to close this one since assert.throws was added in 9a3eb20
Thanks @43081j

[43081j]

@ry fail is still in here.. i had rebased and stripped it down to only the fail method.

[ry]

@43081j Oops - can you rebase once more? LGTM

[43081j]

@ry rebased

[ry]

LGTM - thanks : )