Remove 'u' suffix support from TextFormat.

[thomasvl]

https://protobuf.dev/reference/protobuf/textformat-spec/#numeric and https://protobuf.dev/reference/protobuf/textformat-spec/#literals don't seem to show any signs that a 'u' suffix was ever valid. I also don't see any references in the C++ code. It also seems like we would parse -12u as valid which seems a little odd since it clearly isn't an unsigned value.

So dropping the 'u' support as it seems out of spec.

[thomasvl]

@tbkka - I haven't been able to figure out where this support came from, but it doesn't seem to be to spec. This isn't an api change, but it is a behavior change, so if you'd rather keep it unless we do want to do breaking changes, we can just add a comment that it isn't to spec and it should be removed in the future. Up to you.

[tbkka]

An early version of this file (commit 0aa5109 from Dec 2016) has a comment:

                        case "f", "u":
                            // proto1 allowed floats to be suffixed with 'f'
                            // and unsigned integers to be suffixed with 'u'
                            // Just ignore it:
                            return .number(s)

If I understand correctly, proto1 was never used outside of Google and I presume is no longer used inside of Google, so I'm fine just removing this.

[thomasvl]

Ok, since the C++ doesn't support it either, then, yea, guess safe to drop.

Did you want to go ahead and stamp this with approval?