klippa integration test fails if the Cmap4Iter bumps range start past end of prior range

[rsheeter]

As noted on #1101, if Cmap4Iter sets the start of the range to current.end + 1 this somehow results in klippa integration tests failing. E.g.

// Happy
self.cur_range = next_range.start.max(self.cur_range.end)
                    ..next_range.end.max(self.cur_range.end);

// Sad
self.cur_range = next_range.start.max(self.cur_range.end + 1)  // +1: no reason to reprocess the end of the prior range
                    ..next_range.end.max(self.cur_range.end);

Maybe I'm missing something obvious, at a glance this seems like a bug.

[dfrg]

The cmap4 code uses exclusive ranges so next_range.start.max(self.cur_range.end + 1) will skip an entry if the two ranges are contiguous.

edit: I mean we convert the inclusive range to an exclusive one since it's just easier to work with:

fontations/read-fonts/src/tables/cmap.rs

Line 118 in d4e29ac

// Use end + 1 here because the range in the table is inclusive

[rsheeter]

Oh my bad, confused myself jumping between spec (inclusive) and code (exclusive end). False alarm!