ww/cmd/wcc/tok.c

/*
 * tok.c — token names, keyword lookup, debug printer.
 *
 * One table-of-records keyed by kind. The keyword subset is also
 * scanned linearly during lexing — fewer than 25 entries, a hash
 * isn't worth it.
 */
#include "ww.h"
#include <string.h>

struct kwent {
	const char *s;
	Tkind	kind;
};

/* keep alphabetised, so kwlookup is easy to read. */
static const struct kwent kwtab[] = {
	{ "as",       TK_AS       },
	{ "break",    TK_BREAK    },
	{ "case",     TK_CASE     },
	{ "chan",     TK_CHAN     },
	{ "const",    TK_CONST    },
	{ "continue", TK_CONTINUE },
	{ "def",      TK_DEF      },
	{ "defer",    TK_DEFER    },
	{ "else",     TK_ELSE     },
	{ "export",   TK_EXPORT   },
	{ "false",    TK_FALSE    },
	{ "fn",       TK_FN       },
	{ "for",      TK_FOR      },
	{ "if",       TK_IF       },
	{ "let",      TK_LET      },
	{ "match",    TK_MATCH    },
	{ "nil",      TK_NIL      },
	{ "proc",     TK_PROC     },
	{ "return",   TK_RETURN   },
	{ "static",   TK_STATIC   },
	{ "struct",   TK_STRUCT   },
	{ "switch",   TK_SWITCH   },
	{ "true",     TK_TRUE     },
	{ "type",     TK_TYPE     },
	{ "use",      TK_USE      }
};

Tkind
kwlookup(const char *s, u64 n)
{
	/* linear scan: small N, predictable, branchy fall-through is fine. */
	for (u64 i = 0; i < nelem(kwtab); i++) {
		const char *k = kwtab[i].s;
		if (strlen(k) == n && memcmp(k, s, n) == 0)
			return kwtab[i].kind;
	}
	return TK_NONE;
}

const char *
tokname(Tkind k)
{
	switch (k) {
	case TK_NONE:		return "<none>";
	case TK_EOF:		return "EOF";
	case TK_ERR:		return "ERR";
	case TK_IDENT:		return "IDENT";
	case TK_INT:		return "INT";
	case TK_FLOAT:		return "FLOAT";
	case TK_RUNE:		return "RUNE";
	case TK_STR:		return "STR";

	case TK_FN:		return "fn";
	case TK_LET:		return "let";
	case TK_DEF:		return "def";
	case TK_IF:		return "if";
	case TK_ELSE:		return "else";
	case TK_FOR:		return "for";
	case TK_SWITCH:		return "switch";
	case TK_CASE:		return "case";
	case TK_RETURN:		return "return";
	case TK_USE:		return "use";
	case TK_TYPE:		return "type";
	case TK_STRUCT:		return "struct";
	case TK_DEFER:		return "defer";
	case TK_BREAK:		return "break";
	case TK_CONTINUE:	return "continue";
	case TK_EXPORT:		return "export";
	case TK_PROC:		return "proc";
	case TK_CHAN:		return "chan";
	case TK_NIL:		return "nil";
	case TK_TRUE:		return "true";
	case TK_FALSE:		return "false";
	case TK_AS:		return "as";
	case TK_STATIC:		return "static";
	case TK_MATCH:		return "match";
	case TK_CONST:		return "const";
	case TK_UNDER:		return "_";

	case TK_LPAREN:		return "(";
	case TK_RPAREN:		return ")";
	case TK_LBRACE:		return "{";
	case TK_RBRACE:		return "}";
	case TK_LBRACK:		return "[";
	case TK_RBRACK:		return "]";
	case TK_COMMA:		return ",";
	case TK_SEMI:		return ";";
	case TK_COLON:		return ":";
	case TK_DOT:		return ".";
	case TK_ELLIPSIS:	return "...";
	case TK_DOTDOT:		return "..";
	case TK_AT:		return "@";
	case TK_QUESTION:	return "?";

	case TK_ASSIGN:		return "=";
	case TK_PLUSEQ:		return "+=";
	case TK_MINUSEQ:	return "-=";
	case TK_STAREQ:		return "*=";
	case TK_SLASHEQ:	return "/=";
	case TK_PERCENTEQ:	return "%=";
	case TK_AMPEQ:		return "&=";
	case TK_PIPEEQ:		return "|=";
	case TK_CARETEQ:	return "^=";
	case TK_LSHIFTEQ:	return "<<=";
	case TK_RSHIFTEQ:	return ">>=";

	case TK_PLUS:		return "+";
	case TK_MINUS:		return "-";
	case TK_STAR:		return "*";
	case TK_SLASH:		return "/";
	case TK_PERCENT:	return "%";
	case TK_AMP:		return "&";
	case TK_PIPE:		return "|";
	case TK_CARET:		return "^";
	case TK_TILDE:		return "~";
	case TK_LSHIFT:		return "<<";
	case TK_RSHIFT:		return ">>";

	case TK_EQ:		return "==";
	case TK_NEQ:		return "!=";
	case TK_LT:		return "<";
	case TK_LE:		return "<=";
	case TK_GT:		return ">";
	case TK_GE:		return ">=";

	case TK_AND:		return "&&";
	case TK_OR:		return "||";
	case TK_NOT:		return "!";

	case TK_LARROW:		return "<-";
	case TK_ARROW:		return "->";
	case TK_FATARROW:	return "=>";

	case TK_LAST:		return "<last>";
	}
	return "<?>";
}

static void
fputq(FILE *f, const char *s, u64 n)
{
	fputc('"', f);
	for (u64 i = 0; i < n; i++) {
		unsigned char c = (unsigned char)s[i];
		switch (c) {
		case '\\': fputs("\\\\", f); break;
		case '"':  fputs("\\\"", f); break;
		case '\n': fputs("\\n", f);  break;
		case '\t': fputs("\\t", f);  break;
		case '\r': fputs("\\r", f);  break;
		default:
			if (c < 0x20 || c == 0x7f)
				fprintf(f, "\\x%02x", c);
			else
				fputc(c, f);
		}
	}
	fputc('"', f);
}

void
tokprint(FILE *f, Tok t)
{
	fprintf(f, "%s:%d:%d %s",
	    t.pos.file ? t.pos.file : "<none>", t.pos.line, t.pos.col,
	    tokname(t.kind));
	switch (t.kind) {
	case TK_IDENT:
	case TK_STR:
	case TK_ERR:
		fputc(' ', f);
		fputq(f, t.text, t.tlen);
		break;
	case TK_INT:
	case TK_RUNE:
		fprintf(f, " %llu", (unsigned long long)t.v.uval);
		break;
	case TK_FLOAT:
		fprintf(f, " %g", t.v.fval);
		break;
	default:
		break;
	}
	fputc('\n', f);
}