TerminalLink endIndex should be inclusive (and it should say that in the documentation)

[alexr00]

Testing #101300

The end position of a range used for a DocumentLink appears to be inclusive. I would expect the endIndex of a TerminalLink to behave the same way. Same code to calculate end position results in different range being used:

[Tyriar]

InputBoxOptions.valueSelection uses exclusive end index, so does String.prototype.substring. Opinions @jrieken?

[jrieken]

Yeah, given these are two indexes into a "flat" string I'd stick with substring and valueSelection logic. I see the inconsistency, we can argue that one is ranges with position objects and the other is offsets...

[alexr00]

How often do you expect people to want to share code between a document link provider and a terminal link provider? The inconsistency is annoying in that case because you have to have an extra +1 or -1 somewhere, and that feels broken.