@@ -263,7 +263,7 @@ back to it, we\'ve seen a whole Unicode character. Transitions not in
263the graph are disallowed; they all lead to state one, which has been
263the graph are disallowed; they all lead to state one, which has been
264omitted for readability.
264omitted for readability.
266−
266+
268### Automaton with character class transitions
268### Automaton with character class transitions
@@ -273,7 +273,7 @@ encoding the ranges directly, the ranges are split such that each byte
273belongs to exactly one character class. Then the transitions go over
273belongs to exactly one character class. Then the transitions go over
274these character classes.
274these character classes.
276−
276+
278### Mapping bytes to character classes
278### Mapping bytes to character classes
@@ -404,7 +404,7 @@ the non-ASCII starter is swapped (approximately a 5% increase), but GCC
404I have experimented with various rearrangements of states and character
404I have experimented with various rearrangements of states and character
405classes. A seemingly promising one is the following:
405classes. A seemingly promising one is the following:
407−
407+
409One of the continuation ranges has been split into two, the other
409One of the continuation ranges has been split into two, the other
410changes are just renamings. This arrangement allows, when a continuation
410changes are just renamings. This arrangement allows, when a continuation