Allow ignore to be a Function

[anaisbetts]

This PR allows the ignore option to be a Function as well as a Regexp

[malept]

This should probably be named differently.

[anaisbetts]

hwhoops, agree

[malept]

Thanks for the PR! I have no objections to this new feature, does any other contributor have any constructive criticism?

(@paulcbetts, once there's a consensus, could you please squash your commits?)

[kfranqueiro]

My one concern is that specifying ignore as a function, but then requiring it to return false to ignore, feels kind of backwards. (This is primarily awkward because previously the function was internal whereas now it's directly exposed.) I realize the doc update explains that, but I figure it's still apt to throw people.

Given that the code calls the ignore function within the function it actually passes to filter anyway, could we throw ! in front of that call so that ignore as a function could be implemented to return true to ignore instead?

[kfranqueiro]

Also, it seems as if the added test failed on CI?

[anaisbetts]

(@paulcbetts, once there's a consensus, could you please squash your commits?)

nbd

Given that the code calls the ignore function within the function it actually passes to filter anyway, could we throw ! in front of that call so that ignore as a function could be implemented to return true to ignore instead?

I'm okay with that

Also, it seems as if the added test failed on CI?

Yeah, I gotta fix that up

[max-mapper]

lgtm! only nit is that name is a reserved keyword so I normally don't use it, but it probably doesn't matter here