James Addison

This generally looks good to me! It'll be reassuring to have this in place to maintain test style consistency (and therefore readability / reviewability!). Related to the exception comment/suggestion: it...

> * reorders abstract class to match the discussed order but moved `host` to the top of the list. the current setup of the abstract requires that it be first...

Thanks a lot for working through my suggested changes @jknndy, including some that I think were/are a bit faulty. In future I might be tempted to refactor some of the...

Hi @ThomasHFWright - thanks for the bugreport - sorry it's taken a while to respond! I'm planning to investigate this problem soon, likely within the next few days.

...apologies; I didn't get around to that, but should have time to soon (again, next few days).

@ThomasHFWright can you provide any other example URLs where this problem occurs?

I'm going to close this, because opening the recipe in a web browser displays the text "1 1 tbsp cornflour", along with the other unusually-described ingredient quantities. So I think...

Some initial thoughts: * Test coverage can be expanded/reduced without requiring any changes to the source code it references; so initially, yes it seems that it's valid to report changes...

That.. would make a lot more sense, thank you for catching that.

After re-reading your comment: I'm not sure the adjusted logic I've pushed is correct, either; it could still make sense to provide a non-error-severity output, instead of nothing at all.