Store token locations and include them in parsing error messages

4 files changed, 39 insertions, 8 deletions

diff --git a/include/lex.h b/include/lex.h
index ba1596d..51485a5 100644
--- a/include/lex.h
+++ b/include/lex.h
@@ -103,8 +103,14 @@ struct number {
 	} value;
 };
 
+struct location {
+	const char *file;
+	int line, column;
+};
+
 struct token {
 	enum lexical_token token;
+	struct location loc;
 	union {
 		struct number num;
 		const char *str;
@@ -117,11 +123,14 @@ struct lexer {
 	char *buf;
 	size_t bufsz, buflen;
 	struct token un;
+	struct location loc;
 };
 
-void lex_init(struct lexer *lexer, FILE *f);
+void lex_init(struct lexer *lexer, FILE *f, const char *filename);
 void lex_finish(struct lexer *lexer);
 
+struct location lex_loc(struct lexer *lexer);
+
 enum lexical_token lex(struct lexer *lexer, struct token *out);
 void unlex(struct lexer *lexer, const struct token *in);
 bool match(struct lexer *lexer, enum lexical_token token);
diff --git a/src/lex.c b/src/lex.c
index 69a92a7..e616160 100644
--- a/src/lex.c
+++ b/src/lex.c
@@ -46,7 +46,7 @@ const char *tokens[] = {
 };
 
 void
-lex_init(struct lexer *lexer, FILE *f)
+lex_init(struct lexer *lexer, FILE *f, const char *filename)
 {
 	lexer->in = f;
 	lexer->c[0] = 0;
@@ -55,6 +55,9 @@ lex_init(struct lexer *lexer, FILE *f)
 	lexer->bufsz = 2;
 	lexer->buflen = 0;
 	lexer->un.token = T_NONE;
+	lexer->loc.file = filename;
+	lexer->loc.line = 1;
+	lexer->loc.column = 0;
 }
 
 void
@@ -63,6 +66,15 @@ lex_finish(struct lexer *lexer)
 	fclose(lexer->in);
 }
 
+struct location
+lex_loc(struct lexer *lexer)
+{
+	if (lexer->un.token != T_NONE) {
+		return lexer->un.loc;
+	}
+	return lexer->loc;
+}
+
 static uint32_t next(struct lexer *lexer)
 {
 	if (lexer->c[0]) {
@@ -75,6 +87,13 @@ static uint32_t next(struct lexer *lexer)
 		return c;
 	}
 
+	if (c == '\n') {
+		++lexer->loc.line;
+		lexer->loc.column = 0;
+	} else {
+		++lexer->loc.column;
+	}
+
 	lexer->c[1] = c;
 	return c;
 }
@@ -240,6 +259,7 @@ lex(struct lexer *lexer, struct token *out)
 	}
 
 	uint32_t c = wgetc(lexer);
+	out->loc = lexer->loc;
 	if (c == C_EOF) {
 		out->token = T_EOF;
 		return out->token;
diff --git a/src/main.c b/src/main.c
index c0cee11..f318c52 100644
--- a/src/main.c
+++ b/src/main.c
@@ -41,7 +41,7 @@ main(int argc, char *argv[])
 			return EXIT_ABNORMAL;
 		}
 
-		lex_init(&lexer, in);
+		lex_init(&lexer, in, path);
 		parse(&lexer, &ast);
 		lex_finish(&lexer);
 	}
diff --git a/src/parse.c b/src/parse.c
index a047424..2fa3ee4 100644
--- a/src/parse.c
+++ b/src/parse.c
@@ -5,8 +5,10 @@
 #include "util.h"
 
 static noreturn void
-error(const char *fmt, ...)
+error(struct location loc, const char *fmt, ...)
 {
+	fprintf(stderr, "%s:%d:%d ", loc.file, loc.line, loc.column);
+
 	va_list ap;
 	va_start(ap, fmt);
 	vfprintf(stderr, fmt, ap);
@@ -34,7 +36,7 @@ parse_path(struct lexer *lexer, struct ast_path *out)
 	struct token name;
 
 	if (lex(lexer, &name) != T_NAME) {
-		error("syntax error: expected path");
+		error(name.loc, "syntax error: expected path");
 	}
 
 	out->segments = must_calloc(1, sizeof(const char *));
@@ -44,7 +46,7 @@ parse_path(struct lexer *lexer, struct ast_path *out)
 
 	while (match(lexer, T_MODDELIM)) {
 		if (lex(lexer, &name) != T_NAME) {
-			error("syntax error: expected name");
+			error(name.loc, "syntax error: expected name");
 		}
 		append(out, name.info.str);
 	}
@@ -60,7 +62,7 @@ parse_import(struct lexer *lexer, struct ast_import *out)
 	}
 
 	if (!parse_path(lexer, &out->path)) {
-		error("syntax error: expected import path");
+		error(lex_loc(lexer), "syntax error: expected import path");
 	}
 
 	struct token token;
@@ -72,7 +74,7 @@ parse_import(struct lexer *lexer, struct ast_import *out)
 	}
 
 	if (!match(lexer, T_SEMICOLON)) {
-		error("syntax error: expected semicolon");
+		error(lex_loc(lexer), "syntax error: expected semicolon");
 	}
 
 	return true;