Fix Product2's indexing logic

[timvermeulen]

This PR changes the way Product2.endIndex is represented: instead of (base1.endIndex, base2.endIndex), it is now (base1.endIndex, base2.startIndex).

The latter turns out to be a more natural way of representing the "past the end" index because it is exactly the index after the last valid index (lastValidIndexOfBase1, lastValidIndexOfBase2) according to the existing mechanism of advancing any other index of Product2. As a result, the methods index(before:), index(after:), and distance(from:to:) don't need any custom logic to deal with endIndex.

endIndex wasn't handled correctly everywhere so this also ends up fixing bugs in index(before:) and distance(from:to:).

Checklist

• I've added at least one test that validates that my change is working, if appropriate
• I've followed the code style of the rest of the project
• I've read the Contribution Guidelines
• I've updated the documentation if necessary

[kylemacomber]

This looks like a great simplification. Nice work @timvermeulen!

[kylemacomber]

I'm not sure we need this comment to make it into the final patch

[timvermeulen]

Thanks for the feedback, the only reason for putting it there was to indicate to someone reading the code that this is not a typo – any idea how else we could convey this?

[kylemacomber]

I'll defer to @natecook1000 on whether we should keep it, I guess I generally don't assume there are typos in the code, and if I'm surprised I'll read the surrounding code closer.

That being said, I can understand wanting to draw attention to i2 being base.startIndex, to prevent the reader from glossing over it. For example:

// `base2.startIndex` simplifies index calculations.

[timvermeulen]

Thanks, that's indeed what I was going for. I'll change it to that for now and then leave it up to Nate whether to remove it completely.

[natecook1000]

Thanks, @timvermeulen!