hare

commit 2fc399c5e592dc6c8114e790510d4b232f0d341a
parent 8038d63a4306f4124dc61ccffe677c2db4468a73
Author: Alexey Yerin <yyp@disroot.org>
Date:   Sun, 19 Feb 2023 22:52:38 +0300

hare::lex: propagate syntax errors in nextw

Signed-off-by: Alexey Yerin <yyp@disroot.org>

Diffstat:
M
hare/lex/+test.ha
 | 
10
++++++++++
M
hare/lex/lex.ha
 | 
8
++++----

2 files changed, 14 insertions(+), 4 deletions(-)
diff --git a/hare/lex/+test.ha b/hare/lex/+test.ha
@@ -312,6 +312,16 @@ fn loc(line: uint, col: uint) location = location {
 
 	const s = lex(&lexer) as error as syntax;
 	assert(s.1 == "Source file is not valid UTF-8");
+
+	// Regression: invalid UTF-8 at the beginning of a token used to cause
+	// a crash in nextw
+	const in = [0x80: u8];
+
+	let buf = bufio::fixed(in, mode::READ);
+	let lexer = init(&buf, "<test>");
+
+	const s = lex(&lexer) as error as syntax;
+	assert(s.1 == "Source file is not valid UTF-8");
 };
 
 
diff --git a/hare/lex/lex.ha b/hare/lex/lex.ha
@@ -751,10 +751,10 @@ fn next(lex: *lexer) ((rune, location) | syntax | io::EOF | io::error) = {
 	};
 };
 
-fn nextw(lex: *lexer) ((rune, location) | io::EOF | io::error) = {
-	for (true) match (next(lex)) {
-	case let e: (io::error | io::EOF) =>
-		return e;
+fn nextw(lex: *lexer) ((rune, location) | io::EOF | error) = {
+	for (true) match (next(lex)?) {
+	case io::EOF =>
+		return io::EOF;
 	case let r: (rune, location) =>
 		if (ascii::isspace(r.0)) {
 			if (r.0 == '\n') {