Simon Brunning

Not quite that simple - `has_entry()` allows a matcher for the key as well as for the value. I'll take a look when I have some time.

Thinking about it, perhaps that's why `has_entry()` doesn't give better mismatch messages. Imagine that the key was specified as `hamcrest.anything()` - how should the message read then? Still, if the...

Does https://github.com/hamcrest/PyHamcrest/pull/157 look any good, @dz0 and @offbyone?

Ah - doing this in a way which is compatible with Python versions prior to 3.9 is going to be... interesting. FYI, PyHamcrest works with versions 3.6 and up as...

Oh, cool - but _something's_ not working with older versions. FWIW, I usually do `tox -e py36,py39,lint` before committing - I find that's enough to give me a good idea...

Strict mode is new to me, I must say, but I'll take a look at this when I get the chance. The `import *` thing is easily fixed in a...

Any truthy or falsy value will do here - we don't need an actual boolean.

It wouldn't be very pythonic to require a boolean type, and it would break a *lot* of existing tests. If we were going to do anything so drastic, I'd be...

OK, maybe not *break* as such, but there will be loads of warnings issued for (currently) perfectly valid tests.

*Some* of these tests may be broken, but most of them will not be. We should only be issuing warnings where things *need* fixing, not where they *might* need fixing....