private void End(State start, State end, UTF8Sequence utf8, int upto, bool doAll)
{
if (upto == utf8.len - 1)
{
// Done recursing
start.AddTransition(new Transition(utf8.ByteAt(upto) & (~MASKS[utf8.NumBits(upto) - 1]), utf8.ByteAt(upto), end)); // type=end
}
else
{
int startCode;
if (utf8.NumBits(upto) == 5)
{
// special case -- avoid created unused edges (utf8
// doesn't accept certain byte sequences) -- there
// are other cases we could optimize too:
startCode = 194;
}
else
{
startCode = utf8.ByteAt(upto) & (~MASKS[utf8.NumBits(upto) - 1]);
}
if (doAll && utf8.ByteAt(upto) != startCode)
{
All(start, end, startCode, utf8.ByteAt(upto) - 1, utf8.len - upto - 1);
}
State n = NewUTF8State();
start.AddTransition(new Transition(utf8.ByteAt(upto), n)); // type=end
End(n, end, utf8, 1 + upto, true);
}
}
private void Start(State start, State end, UTF8Sequence utf8, int upto, bool doAll)
{
if (upto == utf8.len - 1)
{
// Done recursing
start.AddTransition(new Transition(utf8.ByteAt(upto), utf8.ByteAt(upto) | MASKS[utf8.NumBits(upto) - 1], end)); // type=start
}
else
{
State n = NewUTF8State();
start.AddTransition(new Transition(utf8.ByteAt(upto), n)); // type=start
Start(n, end, utf8, 1 + upto, true);
int endCode = utf8.ByteAt(upto) | MASKS[utf8.NumBits(upto) - 1];
if (doAll && utf8.ByteAt(upto) != endCode)
{
All(start, end, utf8.ByteAt(upto) + 1, endCode, utf8.len - upto - 1);
}
}
}
private void Build(State start, State end, UTF8Sequence startUTF8, UTF8Sequence endUTF8, int upto)
{
// Break into start, middle, end:
if (startUTF8.ByteAt(upto) == endUTF8.ByteAt(upto))
{
// Degen case: lead with the same byte:
if (upto == startUTF8.len - 1 && upto == endUTF8.len - 1)
{
// Super degen: just single edge, one UTF8 byte:
start.AddTransition(new Transition(startUTF8.ByteAt(upto), endUTF8.ByteAt(upto), end));
return;
}
else
{
if (Debugging.AssertsEnabled)
{
Debugging.Assert(startUTF8.len > upto + 1);
Debugging.Assert(endUTF8.len > upto + 1);
}
State n = NewUTF8State();
// Single value leading edge
start.AddTransition(new Transition(startUTF8.ByteAt(upto), n)); // type=single
// Recurse for the rest
Build(n, end, startUTF8, endUTF8, 1 + upto);
}
}
else if (startUTF8.len == endUTF8.len)
{
if (upto == startUTF8.len - 1)
{
start.AddTransition(new Transition(startUTF8.ByteAt(upto), endUTF8.ByteAt(upto), end)); // type=startend
}
else
{
Start(start, end, startUTF8, upto, false);
if (endUTF8.ByteAt(upto) - startUTF8.ByteAt(upto) > 1)
{
// There is a middle
All(start, end, startUTF8.ByteAt(upto) + 1, endUTF8.ByteAt(upto) - 1, startUTF8.len - upto - 1);
}
End(start, end, endUTF8, upto, false);
}
}
else
{
// start
Start(start, end, startUTF8, upto, true);
// possibly middle, spanning multiple num bytes
int byteCount = 1 + startUTF8.len - upto;
int limit = endUTF8.len - upto;
while (byteCount < limit)
{
// wasteful: we only need first byte, and, we should
// statically encode this first byte:
tmpUTF8a.Set(startCodes[byteCount - 1]);
tmpUTF8b.Set(endCodes[byteCount - 1]);
All(start, end, tmpUTF8a.ByteAt(0), tmpUTF8b.ByteAt(0), tmpUTF8a.len - 1);
byteCount++;
}
// end
End(start, end, endUTF8, upto, true);
}
}
private void End(State start, State end, UTF8Sequence utf8, int upto, bool doAll)
{
if (upto == utf8.Len - 1)
{
// Done recursing
start.AddTransition(new Transition(utf8.ByteAt(upto) & (~MASKS[utf8.NumBits(upto) - 1]), utf8.ByteAt(upto), end)); // type=end
}
else
{
int startCode;
if (utf8.NumBits(upto) == 5)
{
// special case -- avoid created unused edges (utf8
// doesn't accept certain byte sequences) -- there
// are other cases we could optimize too:
startCode = 194;
}
else
{
startCode = utf8.ByteAt(upto) & (~MASKS[utf8.NumBits(upto) - 1]);
}
if (doAll && utf8.ByteAt(upto) != startCode)
{
All(start, end, startCode, utf8.ByteAt(upto) - 1, utf8.Len - upto - 1);
}
State n = NewUTF8State();
start.AddTransition(new Transition(utf8.ByteAt(upto), n)); // type=end
End(n, end, utf8, 1 + upto, true);
}
}
private void Start(State start, State end, UTF8Sequence utf8, int upto, bool doAll)
{
if (upto == utf8.Len - 1)
{
// Done recursing
start.AddTransition(new Transition(utf8.ByteAt(upto), utf8.ByteAt(upto) | MASKS[utf8.NumBits(upto) - 1], end)); // type=start
}
else
{
State n = NewUTF8State();
start.AddTransition(new Transition(utf8.ByteAt(upto), n)); // type=start
Start(n, end, utf8, 1 + upto, true);
int endCode = utf8.ByteAt(upto) | MASKS[utf8.NumBits(upto) - 1];
if (doAll && utf8.ByteAt(upto) != endCode)
{
All(start, end, utf8.ByteAt(upto) + 1, endCode, utf8.Len - upto - 1);
}
}
}
private void Build(State start, State end, UTF8Sequence startUTF8, UTF8Sequence endUTF8, int upto)
{
// Break into start, middle, end:
if (startUTF8.ByteAt(upto) == endUTF8.ByteAt(upto))
{
// Degen case: lead with the same byte:
if (upto == startUTF8.Len - 1 && upto == endUTF8.Len - 1)
{
// Super degen: just single edge, one UTF8 byte:
start.AddTransition(new Transition(startUTF8.ByteAt(upto), endUTF8.ByteAt(upto), end));
return;
}
else
{
Debug.Assert(startUTF8.Len > upto + 1);
Debug.Assert(endUTF8.Len > upto + 1);
State n = NewUTF8State();
// Single value leading edge
start.AddTransition(new Transition(startUTF8.ByteAt(upto), n)); // type=single
// Recurse for the rest
Build(n, end, startUTF8, endUTF8, 1 + upto);
}
}
else if (startUTF8.Len == endUTF8.Len)
{
if (upto == startUTF8.Len - 1)
{
start.AddTransition(new Transition(startUTF8.ByteAt(upto), endUTF8.ByteAt(upto), end)); // type=startend
}
else
{
Start(start, end, startUTF8, upto, false);
if (endUTF8.ByteAt(upto) - startUTF8.ByteAt(upto) > 1)
{
// There is a middle
All(start, end, startUTF8.ByteAt(upto) + 1, endUTF8.ByteAt(upto) - 1, startUTF8.Len - upto - 1);
}
End(start, end, endUTF8, upto, false);
}
}
else
{
// start
Start(start, end, startUTF8, upto, true);
// possibly middle, spanning multiple num bytes
int byteCount = 1 + startUTF8.Len - upto;
int limit = endUTF8.Len - upto;
while (byteCount < limit)
{
// wasteful: we only need first byte, and, we should
// statically encode this first byte:
TmpUTF8a.Set(StartCodes[byteCount - 1]);
TmpUTF8b.Set(EndCodes[byteCount - 1]);
All(start, end, TmpUTF8a.ByteAt(0), TmpUTF8b.ByteAt(0), TmpUTF8a.Len - 1);
byteCount++;
}
// end
End(start, end, endUTF8, upto, true);
}
}
