lib/charset: utf8_get() should return error
utf8_get() should return an error if hitting an illegal UTF-8 sequence and
not silently convert the input to a question mark.
Correct utf_8() and the its unit test.
console_read_unicode() now will ignore illegal UTF-8 sequences.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
diff --git a/test/unicode_ut.c b/test/unicode_ut.c
index 2cc6b5f..154361a 100644
--- a/test/unicode_ut.c
+++ b/test/unicode_ut.c
@@ -52,6 +52,7 @@
static const char j1[] = {0x6a, 0x31, 0xa1, 0x6c, 0x00};
static const char j2[] = {0x6a, 0x32, 0xc3, 0xc3, 0x6c, 0x00};
static const char j3[] = {0x6a, 0x33, 0xf0, 0x90, 0xf0, 0x00};
+static const char j4[] = {0xa1, 0x00};
static int unicode_test_u16_strlen(struct unit_test_state *uts)
{
@@ -165,6 +166,12 @@
ut_asserteq(0x0001048d, code);
ut_asserteq_ptr(s, d4 + 4);
+ /* Check illegal character */
+ s = j4;
+ code = utf8_get((const char **)&s);
+ ut_asserteq(-1, code);
+ ut_asserteq_ptr(j4 + 1, s);
+
return 0;
}
UNICODE_TEST(unicode_test_utf8_get);