angle-parse.go

package main

/*
[...] angles [...] are specified as
up to three numbers representing degrees, minutes and seconds.  The
last number may be fractional.  The numbers are either separated by
<,> (commas) or suffixed, in order:

  degrees by <°> (degree sign),
  minutes by <′> (prime)        or <'> (apostrophe),
  seconds by <″> (double prime) or <"> (quotation mark).

The last suffix may be omitted.
*/

// angleStringOK returns ok if the string looks valid and suffixed if it
// contains <">, <'> or multibyte UTF-8 characters.
func angleStringOK(s string) (ok, suffixed bool) {
	// The main goals of this function are:
	//   - to disallow strings with:
	//     - signs (<+> and <->), to avoid validating strings before
	//       passing them to strconv.ParseFloat
	//     - more than 2 commas, to avoid excessive allocation
	//   - to decide whether the numbers are separated by commas or
	//     suffixed with <°> etc.
	// Additionally, it disallows:
	//   - illegal ASCII characters
	//   - more than 3 suffixes or multibyte characters
	//   - strings having both commas and multibyte characters
	// The validation is preliminary, optimised for speed and very crude.
	// In particular, it:
	//   - doesn't check which multibyte characters are present and counts
	//     any such character as a suffix
	//   - doesn't validate syntax (character sequences)
	//   - doesn't even validate UTF-8 (byte sequences and invalid bytes
	//     0xc0, 0xc1, 0xfe, 0xff)
	const (
		invalid = iota
		valid
		comma
		suffix
		tbl20
	)
	var commas, suffixes byte
	// map of 8 byte ranges (top 3 bits)
	hi := [8]byte{
		0x20 >> 5: tbl20,  // 0x20 - 0x3f: see table lo20
		0x80 >> 5: valid,  // 0x80 - 0x9f: 10xxxxxx, UTF-8 continuation
		0xa0 >> 5: valid,  // 0xa0 - 0xbf: 10xxxxxx, UTF-8 continuation
		0xc0 >> 5: suffix, // 0xc0 - 0xdf: 11xxxxxx, UTF-8 start
		0xe0 >> 5: suffix, // 0xe0 - 0xff: 11xxxxxx, UTF-8 start
	}
	// map of valid characters in the 0x20 - 0x3f range (bottom 5 bits)
	lo20 := [32]byte{
		'"' - 0x20:  suffix, // 0x22: count suffixes
		'\'' - 0x20: suffix, // 0x27: count suffixes
		',' - 0x20:  comma,  // 0x2c: count commas
		'.' - 0x20:  valid,  // 0x2e
		'0' - 0x20:  valid,  // 0x30
		'1' - 0x20:  valid,  // 0x31
		'2' - 0x20:  valid,  // 0x32
		'3' - 0x20:  valid,  // 0x33
		'4' - 0x20:  valid,  // 0x34
		'5' - 0x20:  valid,  // 0x35
		'6' - 0x20:  valid,  // 0x36
		'7' - 0x20:  valid,  // 0x37
		'8' - 0x20:  valid,  // 0x38
		'9' - 0x20:  valid,  // 0x39
	}
	for i := 0; i < len(s); i++ {
		c := s[i]
		act := hi[c>>5]
		if act == tbl20 {
			act = lo20[c&0x1f]
		}
		switch act {
		case invalid:
			return false, false
		case valid:
		case comma:
			// up to 2 commas, no suffixes
			if commas>>1|suffixes != 0 {
				return false, false
			}
			commas++
		default:
			// up to 3 suffixes, no commas
			if suffixes++; suffixes>>2|commas != 0 {
				return false, false
			}
		}
	}
	return true, suffixes != 0
}