harec

commit 5c58caef2eed8e41897e06ce6e264ebffb95ca13
parent 090dd37eb360d5168131bf47fb078debd9fc94e2
Author: Bor Grošelj Simić <bor.groseljsimic@telemach.net>
Date:   Sat, 20 Nov 2021 02:48:42 +0100

error out on codepoints not known to Hare grammar

Unknown codepoints were lexed as T_ERROR. T_ERROR is special-cased in
unlex(), which caused such errors to be ignored.

Signed-off-by: Bor Grošelj Simić <bor.groseljsimic@telemach.net>

Diffstat:
M
src/lex.c
 | 
7
+++++--

1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/src/lex.c b/src/lex.c
@@ -897,6 +897,7 @@ _lex(struct lexer *lexer, struct token *out)
 		return lex_literal(lexer, out);
 	}
 
+	char p[5];
 	switch (c) {
 	case '"':
 	case '\'':
@@ -949,8 +950,10 @@ _lex(struct lexer *lexer, struct token *out)
 		out->token = T_QUESTION;
 		break;
 	default:
-		out->token = T_ERROR;
-		break;
+		p[utf8_encode(p, c)] = '\0';
+		fprintf(stderr, "Error: unexpected code point '%s' at %s:%d:%d\n",
+			p, lexer->loc.path, lexer->loc.lineno, lexer->loc.colno);
+		exit(EXIT_FAILURE);
 	}
 
 	return out->token;