// selfhost/cmd/w6a/lex.ww — port of cmd/w6a/lex.c.
//
// Character-level helpers for w6a's line-oriented parser. The parser
// itself is in parse.ww; here we keep tokenisers for identifiers and
// numbers so parse.ww stays focused on syntax.

export fn isidstart(c: i32) bool = {
	if (c == 95) { return true; };
	if (c >= 65) { if (c <= 90)  { return true; }; };	// A-Z
	if (c >= 97) { if (c <= 122) { return true; }; };	// a-z
	return false;
};

export fn isidcont(c: i32) bool = {
	if (isidstart(c)) { return true; };
	if (c >= 48) { if (c <= 57) { return true; }; };	// 0-9
	if (c == 46) { return true; };				// .
	return false;
};

// parsenum — read a leading [+-]?[0x|0X|0]?digits from p[0..n-1].
// Returns (value, consumed). Stops at first non-digit.
// Plain Plan 9-style: $123 / $0x1f / $-7. Decimal default; 0x prefix
// for hex; 0 prefix for octal when followed by a digit (else just 0).
export fn parsenum(p: *u8, n: u64) (i64, u64) = {
	let i: u64 = 0u64;
	let neg: bool = false;
	if (i < n) {
		if (p[i] == 45u8) { neg = true; i += 1u64; }
		else { if (p[i] == 43u8) { i += 1u64; }; };
	};
	let base: i64 = 10i64;
	if (i + 1u64 < n) {
		if (p[i] == 48u8) {
			if (p[i + 1u64] == 120u8) { base = 16i64; i += 2u64; }
			else { if (p[i + 1u64] == 88u8) { base = 16i64; i += 2u64; }
			else { if (p[i + 1u64] >= 48u8) { if (p[i + 1u64] <= 55u8) {
				base = 8i64; i += 1u64;
			};};};};
		};
	};
	let v: i64 = 0i64;
	let scan: bool = true;
	for (scan) {
		if (i >= n) { scan = false; }
		else {
			let c: u8 = p[i];
			let d: i64 = -1i64;
			if (c >= 48u8) { if (c <= 57u8) { d = (c - 48u8): i64; }; };
			if (d < 0i64) {
				if (base == 16i64) {
					if (c >= 97u8) { if (c <= 102u8) { d = (c - 97u8): i64 + 10i64; }; };
					if (c >= 65u8) { if (c <= 70u8)  { d = (c - 65u8): i64 + 10i64; }; };
				};
			};
			if (d < 0i64) { scan = false; }
			else { if (d >= base) { scan = false; }
			else {
				v = v * base + d;
				i += 1u64;
			}; };
		};
	};
	if (neg) { v = -v; };
	return v, i;
};