Enforce valid encoding in format strings, anyway

eliaskosunen · Nov 8, 2023 · 0529a8f · 0529a8f
1 parent 877efb9
commit 0529a8f
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 2 deletions.
diff --git a/include/scn/detail/format_string.h b/include/scn/detail/format_string.h
@@ -403,7 +403,11 @@ namespace scn {
                                            const CharT* end) const
             {
                 // TODO: Do we want to validate Unicode in format strings?
-#if 1
+                // We're dealing with text, we probably do.
+                // We could do codeunit-to-codeunit matching,
+                // but that could get messy wrt. whitespace matching.
+                // It's simpler to not allow nonsense.
+#if 0
                 SCN_UNUSED(begin);
                 SCN_UNUSED(end);
 #else

diff --git a/tests/unittests/format_string_test.cpp b/tests/unittests/format_string_test.cpp
@@ -278,5 +278,7 @@ TEST(FormatStringTest, LongFormatString4)
 
 TEST(FormatStringTest, MatchLiteralInvalidEncoding)
 {
-    auto result = scn::scan<>("\xc3\na\xa4", "\xc3\n\a\xa4");
+    auto result = scn::scan<>("\xc3\na\xa4", scn::runtime("\xc3\na\xa4"));
+    ASSERT_FALSE(result);
+    EXPECT_EQ(result.error().code(), scn::scan_error::invalid_format_string);
 }