summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorMarkus Armbruster <armbru@redhat.com>2013-04-11 18:07:20 +0200
committerBlue Swirl <blauwirbel@gmail.com>2013-04-13 19:40:17 +0000
commit1d50c8e947180174acb02bad9ff95e0aee6249ea (patch)
treeaf7699d04ce94fef3c9967fb3580f0ca5a66bcf2
parentd6244e2ce48b353402eff271d382ee6fd47ce166 (diff)
downloadqemu-1d50c8e947180174acb02bad9ff95e0aee6249ea.tar.gz
check-qjson: Test noncharacters other than U+FFFE, U+FFFF in strings
Test cases cover the two noncharacters in the BMP. Add tests for the other 64 noncharacters. Three existing test cases involve noncharacters U+FFFF and U+10FFFF. Instead of deleting them as now duplicates, adjust them to use U+FFFC and U+10FFFFD. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Laszlo Ersek <lersek@redhat.com> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
-rw-r--r--tests/check-qjson.c96
1 files changed, 84 insertions, 12 deletions
diff --git a/tests/check-qjson.c b/tests/check-qjson.c
index 91b4e5dbf5..54074a9275 100644
--- a/tests/check-qjson.c
+++ b/tests/check-qjson.c
@@ -158,7 +158,7 @@ static void utf8_string(void)
* consider using overlong encoding \xC0\x80 for U+0000 ("modified
* UTF-8").
*
- * Test cases are scraped from Markus Kuhn's UTF-8 decoder
+ * Most test cases are scraped from Markus Kuhn's UTF-8 decoder
* capability and stress test at
* http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt
*/
@@ -256,11 +256,19 @@ static void utf8_string(void)
"\xDF\xBF",
"\"\\u07FF\"",
},
- /* 2.2.3 3 bytes U+FFFF */
+ /*
+ * 2.2.3 3 bytes U+FFFC
+ * The last possible sequence is actually U+FFFF. But that's
+ * a noncharacter, and already covered by its own test case
+ * under 5.3. Same for U+FFFE. U+FFFD is the last character
+ * in the BMP, and covered under 2.3. Because of U+FFFD's
+ * special role as replacement character, it's worth testing
+ * U+FFFC here.
+ */
{
- "\"\xEF\xBF\xBF\"",
- "\xEF\xBF\xBF",
- "\"\\uFFFF\"",
+ "\"\xEF\xBF\xBC\"",
+ "\xEF\xBF\xBC",
+ "\"\\uFFFC\"",
},
/* 2.2.4 4 bytes U+1FFFFF */
{
@@ -303,10 +311,10 @@ static void utf8_string(void)
"\"\\uFFFD\"",
},
{
- /* last one in last plane: U+10FFFF */
- "\"\xF4\x8F\xBF\xBF\"",
- "\xF4\x8F\xBF\xBF",
- "\"\\u43FF\\uFFFF\"", /* bug: want "\"\\uDBFF\\uDFFF\"" */
+ /* last one in last plane: U+10FFFD */
+ "\"\xF4\x8F\xBF\xBD\"",
+ "\xF4\x8F\xBF\xBD",
+ "\"\\u43FF\\uFFFF\"", /* bug: want "\"\\uDBFF\\uDFFD\"" */
},
{
/* first one beyond Unicode range: U+110000 */
@@ -589,9 +597,14 @@ static void utf8_string(void)
"\"\\u07FF\"",
},
{
- /* \U+FFFF */
- "\"\xF0\x8F\xBF\xBF\"",
- "\xF0\x8F\xBF\xBF", /* bug: not corrected */
+ /*
+ * \U+FFFC
+ * The actual maximum would be U+FFFF, but that's a
+ * noncharacter. Testing U+FFFC seems more useful. See
+ * also 2.2.3
+ */
+ "\"\xF0\x8F\xBF\xBC\"",
+ "\xF0\x8F\xBF\xBC", /* bug: not corrected */
"\"\\u03FF\\uFFFF\"", /* bug: want "\"\\uFFFF\"" */
},
{
@@ -736,6 +749,7 @@ static void utf8_string(void)
"\"\\uDBFF\\uDFFF\"", /* bug: want "\"\\uFFFF\\uFFFF\"" */
},
/* 5.3 Other illegal code positions */
+ /* BMP noncharacters */
{
/* \U+FFFE */
"\"\xEF\xBF\xBE\"",
@@ -748,6 +762,64 @@ static void utf8_string(void)
"\xEF\xBF\xBF", /* bug: not corrected */
"\"\\uFFFF\"", /* bug: not corrected */
},
+ {
+ /* U+FDD0 */
+ "\"\xEF\xB7\x90\"",
+ "\xEF\xB7\x90", /* bug: not corrected */
+ "\"\\uFDD0\"", /* bug: not corrected */
+ },
+ {
+ /* U+FDEF */
+ "\"\xEF\xB7\xAF\"",
+ "\xEF\xB7\xAF", /* bug: not corrected */
+ "\"\\uFDEF\"", /* bug: not corrected */
+ },
+ /* Plane 1 .. 16 noncharacters */
+ {
+ /* U+1FFFE U+1FFFF U+2FFFE U+2FFFF ... U+10FFFE U+10FFFF */
+ "\"\xF0\x9F\xBF\xBE\xF0\x9F\xBF\xBF"
+ "\xF0\xAF\xBF\xBE\xF0\xAF\xBF\xBF"
+ "\xF0\xBF\xBF\xBE\xF0\xBF\xBF\xBF"
+ "\xF1\x8F\xBF\xBE\xF1\x8F\xBF\xBF"
+ "\xF1\x9F\xBF\xBE\xF1\x9F\xBF\xBF"
+ "\xF1\xAF\xBF\xBE\xF1\xAF\xBF\xBF"
+ "\xF1\xBF\xBF\xBE\xF1\xBF\xBF\xBF"
+ "\xF2\x8F\xBF\xBE\xF2\x8F\xBF\xBF"
+ "\xF2\x9F\xBF\xBE\xF2\x9F\xBF\xBF"
+ "\xF2\xAF\xBF\xBE\xF2\xAF\xBF\xBF"
+ "\xF2\xBF\xBF\xBE\xF2\xBF\xBF\xBF"
+ "\xF3\x8F\xBF\xBE\xF3\x8F\xBF\xBF"
+ "\xF3\x9F\xBF\xBE\xF3\x9F\xBF\xBF"
+ "\xF3\xAF\xBF\xBE\xF3\xAF\xBF\xBF"
+ "\xF3\xBF\xBF\xBE\xF3\xBF\xBF\xBF"
+ "\xF4\x8F\xBF\xBE\xF4\x8F\xBF\xBF\"",
+ /* bug: not corrected */
+ "\xF0\x9F\xBF\xBE\xF0\x9F\xBF\xBF"
+ "\xF0\xAF\xBF\xBE\xF0\xAF\xBF\xBF"
+ "\xF0\xBF\xBF\xBE\xF0\xBF\xBF\xBF"
+ "\xF1\x8F\xBF\xBE\xF1\x8F\xBF\xBF"
+ "\xF1\x9F\xBF\xBE\xF1\x9F\xBF\xBF"
+ "\xF1\xAF\xBF\xBE\xF1\xAF\xBF\xBF"
+ "\xF1\xBF\xBF\xBE\xF1\xBF\xBF\xBF"
+ "\xF2\x8F\xBF\xBE\xF2\x8F\xBF\xBF"
+ "\xF2\x9F\xBF\xBE\xF2\x9F\xBF\xBF"
+ "\xF2\xAF\xBF\xBE\xF2\xAF\xBF\xBF"
+ "\xF2\xBF\xBF\xBE\xF2\xBF\xBF\xBF"
+ "\xF3\x8F\xBF\xBE\xF3\x8F\xBF\xBF"
+ "\xF3\x9F\xBF\xBE\xF3\x9F\xBF\xBF"
+ "\xF3\xAF\xBF\xBE\xF3\xAF\xBF\xBF"
+ "\xF3\xBF\xBF\xBE\xF3\xBF\xBF\xBF"
+ "\xF4\x8F\xBF\xBE\xF4\x8F\xBF\xBF",
+ /* bug: not corrected */
+ "\"\\u07FF\\uFFFF\\u07FF\\uFFFF\\u0BFF\\uFFFF\\u0BFF\\uFFFF"
+ "\\u0FFF\\uFFFF\\u0FFF\\uFFFF\\u13FF\\uFFFF\\u13FF\\uFFFF"
+ "\\u17FF\\uFFFF\\u17FF\\uFFFF\\u1BFF\\uFFFF\\u1BFF\\uFFFF"
+ "\\u1FFF\\uFFFF\\u1FFF\\uFFFF\\u23FF\\uFFFF\\u23FF\\uFFFF"
+ "\\u27FF\\uFFFF\\u27FF\\uFFFF\\u2BFF\\uFFFF\\u2BFF\\uFFFF"
+ "\\u2FFF\\uFFFF\\u2FFF\\uFFFF\\u33FF\\uFFFF\\u33FF\\uFFFF"
+ "\\u37FF\\uFFFF\\u37FF\\uFFFF\\u3BFF\\uFFFF\\u3BFF\\uFFFF"
+ "\\u3FFF\\uFFFF\\u3FFF\\uFFFF\\u43FF\\uFFFF\\u43FF\\uFFFF\"",
+ },
{}
};
int i;