Add qemu 2.4.0
[kvmfornfv.git] / qemu / tests / check-qjson.c
diff --git a/qemu/tests/check-qjson.c b/qemu/tests/check-qjson.c
new file mode 100644 (file)
index 0000000..1cfffa5
--- /dev/null
@@ -0,0 +1,1524 @@
+/*
+ * Copyright IBM, Corp. 2009
+ * Copyright (c) 2013, 2015 Red Hat Inc.
+ *
+ * Authors:
+ *  Anthony Liguori   <aliguori@us.ibm.com>
+ *  Markus Armbruster <armbru@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2.1 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+#include <glib.h>
+
+#include "qapi/qmp/qstring.h"
+#include "qapi/qmp/qint.h"
+#include "qapi/qmp/qdict.h"
+#include "qapi/qmp/qlist.h"
+#include "qapi/qmp/qfloat.h"
+#include "qapi/qmp/qbool.h"
+#include "qapi/qmp/qjson.h"
+
+#include "qemu-common.h"
+
+static void escaped_string(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        const char *decoded;
+        int skip;
+    } test_cases[] = {
+        { "\"\\b\"", "\b" },
+        { "\"\\f\"", "\f" },
+        { "\"\\n\"", "\n" },
+        { "\"\\r\"", "\r" },
+        { "\"\\t\"", "\t" },
+        { "\"/\"", "/" },
+        { "\"\\/\"", "/", .skip = 1 },
+        { "\"\\\\\"", "\\" },
+        { "\"\\\"\"", "\"" },
+        { "\"hello world \\\"embedded string\\\"\"",
+          "hello world \"embedded string\"" },
+        { "\"hello world\\nwith new line\"", "hello world\nwith new line" },
+        { "\"single byte utf-8 \\u0020\"", "single byte utf-8  ", .skip = 1 },
+        { "\"double byte utf-8 \\u00A2\"", "double byte utf-8 \xc2\xa2" },
+        { "\"triple byte utf-8 \\u20AC\"", "triple byte utf-8 \xe2\x82\xac" },
+        { "'\\b'", "\b", .skip = 1 },
+        { "'\\f'", "\f", .skip = 1 },
+        { "'\\n'", "\n", .skip = 1 },
+        { "'\\r'", "\r", .skip = 1 },
+        { "'\\t'", "\t", .skip = 1 },
+        { "'\\/'", "/", .skip = 1 },
+        { "'\\\\'", "\\", .skip = 1 },
+        {}
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QString *str;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QSTRING);
+        
+        str = qobject_to_qstring(obj);
+        g_assert_cmpstr(qstring_get_str(str), ==, test_cases[i].decoded);
+
+        if (test_cases[i].skip == 0) {
+            str = qobject_to_json(obj);
+            g_assert_cmpstr(qstring_get_str(str), ==, test_cases[i].encoded);
+            qobject_decref(obj);
+        }
+
+        QDECREF(str);
+    }
+}
+
+static void simple_string(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        const char *decoded;
+    } test_cases[] = {
+        { "\"hello world\"", "hello world" },
+        { "\"the quick brown fox jumped over the fence\"",
+          "the quick brown fox jumped over the fence" },
+        {}
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QString *str;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QSTRING);
+        
+        str = qobject_to_qstring(obj);
+        g_assert(strcmp(qstring_get_str(str), test_cases[i].decoded) == 0);
+
+        str = qobject_to_json(obj);
+        g_assert(strcmp(qstring_get_str(str), test_cases[i].encoded) == 0);
+
+        qobject_decref(obj);
+        
+        QDECREF(str);
+    }
+}
+
+static void single_quote_string(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        const char *decoded;
+    } test_cases[] = {
+        { "'hello world'", "hello world" },
+        { "'the quick brown fox \\' jumped over the fence'",
+          "the quick brown fox ' jumped over the fence" },
+        {}
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QString *str;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QSTRING);
+        
+        str = qobject_to_qstring(obj);
+        g_assert(strcmp(qstring_get_str(str), test_cases[i].decoded) == 0);
+
+        QDECREF(str);
+    }
+}
+
+static void utf8_string(void)
+{
+    /*
+     * FIXME Current behavior for invalid UTF-8 sequences is
+     * incorrect.  This test expects current, incorrect results.
+     * They're all marked "bug:" below, and are to be replaced by
+     * correct ones as the bugs get fixed.
+     *
+     * The JSON parser rejects some invalid sequences, but accepts
+     * others without correcting the problem.
+     *
+     * We should either reject all invalid sequences, or minimize
+     * overlong sequences and replace all other invalid sequences by a
+     * suitable replacement character.  A common choice for
+     * replacement is U+FFFD.
+     *
+     * Problem: we can't easily deal with embedded U+0000.  Parsing
+     * the JSON string "this \\u0000" is fun" yields "this \0 is fun",
+     * which gets misinterpreted as NUL-terminated "this ".  We should
+     * consider using overlong encoding \xC0\x80 for U+0000 ("modified
+     * UTF-8").
+     *
+     * 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
+     */
+    static const struct {
+        const char *json_in;
+        const char *utf8_out;
+        const char *json_out;   /* defaults to @json_in */
+        const char *utf8_in;    /* defaults to @utf8_out */
+    } test_cases[] = {
+        /*
+         * Bug markers used here:
+         * - bug: not corrected
+         *   JSON parser fails to correct invalid sequence(s)
+         * - bug: rejected
+         *   JSON parser rejects invalid sequence(s)
+         *   We may choose to define this as feature
+         * - bug: want "..."
+         *   JSON parser produces incorrect result, this is the
+         *   correct one, assuming replacement character U+FFFF
+         *   We may choose to reject instead of replace
+         */
+
+        /* 1  Some correct UTF-8 text */
+        {
+            /* a bit of German */
+            "\"Falsches \xC3\x9C" "ben von Xylophonmusik qu\xC3\xA4lt"
+            " jeden gr\xC3\xB6\xC3\x9F" "eren Zwerg.\"",
+            "Falsches \xC3\x9C" "ben von Xylophonmusik qu\xC3\xA4lt"
+            " jeden gr\xC3\xB6\xC3\x9F" "eren Zwerg.",
+            "\"Falsches \\u00DCben von Xylophonmusik qu\\u00E4lt"
+            " jeden gr\\u00F6\\u00DFeren Zwerg.\"",
+        },
+        {
+            /* a bit of Greek */
+            "\"\xCE\xBA\xE1\xBD\xB9\xCF\x83\xCE\xBC\xCE\xB5\"",
+            "\xCE\xBA\xE1\xBD\xB9\xCF\x83\xCE\xBC\xCE\xB5",
+            "\"\\u03BA\\u1F79\\u03C3\\u03BC\\u03B5\"",
+        },
+        /* 2  Boundary condition test cases */
+        /* 2.1  First possible sequence of a certain length */
+        /* 2.1.1  1 byte U+0000 */
+        {
+            "\"\\u0000\"",
+            "",                 /* bug: want overlong "\xC0\x80" */
+            "\"\\u0000\"",
+            "\xC0\x80",
+        },
+        /* 2.1.2  2 bytes U+0080 */
+        {
+            "\"\xC2\x80\"",
+            "\xC2\x80",
+            "\"\\u0080\"",
+        },
+        /* 2.1.3  3 bytes U+0800 */
+        {
+            "\"\xE0\xA0\x80\"",
+            "\xE0\xA0\x80",
+            "\"\\u0800\"",
+        },
+        /* 2.1.4  4 bytes U+10000 */
+        {
+            "\"\xF0\x90\x80\x80\"",
+            "\xF0\x90\x80\x80",
+            "\"\\uD800\\uDC00\"",
+        },
+        /* 2.1.5  5 bytes U+200000 */
+        {
+            "\"\xF8\x88\x80\x80\x80\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xF8\x88\x80\x80\x80",
+        },
+        /* 2.1.6  6 bytes U+4000000 */
+        {
+            "\"\xFC\x84\x80\x80\x80\x80\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFC\x84\x80\x80\x80\x80",
+        },
+        /* 2.2  Last possible sequence of a certain length */
+        /* 2.2.1  1 byte U+007F */
+        {
+            "\"\x7F\"",
+            "\x7F",
+            "\"\\u007F\"",
+        },
+        /* 2.2.2  2 bytes U+07FF */
+        {
+            "\"\xDF\xBF\"",
+            "\xDF\xBF",
+            "\"\\u07FF\"",
+        },
+        /*
+         * 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\xBC\"",
+            "\xEF\xBF\xBC",
+            "\"\\uFFFC\"",
+        },
+        /* 2.2.4  4 bytes U+1FFFFF */
+        {
+            "\"\xF7\xBF\xBF\xBF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xF7\xBF\xBF\xBF",
+        },
+        /* 2.2.5  5 bytes U+3FFFFFF */
+        {
+            "\"\xFB\xBF\xBF\xBF\xBF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFB\xBF\xBF\xBF\xBF",
+        },
+        /* 2.2.6  6 bytes U+7FFFFFFF */
+        {
+            "\"\xFD\xBF\xBF\xBF\xBF\xBF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFD\xBF\xBF\xBF\xBF\xBF",
+        },
+        /* 2.3  Other boundary conditions */
+        {
+            /* last one before surrogate range: U+D7FF */
+            "\"\xED\x9F\xBF\"",
+            "\xED\x9F\xBF",
+            "\"\\uD7FF\"",
+        },
+        {
+            /* first one after surrogate range: U+E000 */
+            "\"\xEE\x80\x80\"",
+            "\xEE\x80\x80",
+            "\"\\uE000\"",
+        },
+        {
+            /* last one in BMP: U+FFFD */
+            "\"\xEF\xBF\xBD\"",
+            "\xEF\xBF\xBD",
+            "\"\\uFFFD\"",
+        },
+        {
+            /* last one in last plane: U+10FFFD */
+            "\"\xF4\x8F\xBF\xBD\"",
+            "\xF4\x8F\xBF\xBD",
+            "\"\\uDBFF\\uDFFD\""
+        },
+        {
+            /* first one beyond Unicode range: U+110000 */
+            "\"\xF4\x90\x80\x80\"",
+            "\xF4\x90\x80\x80",
+            "\"\\uFFFD\"",
+        },
+        /* 3  Malformed sequences */
+        /* 3.1  Unexpected continuation bytes */
+        /* 3.1.1  First continuation byte */
+        {
+            "\"\x80\"",
+            "\x80",             /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 3.1.2  Last continuation byte */
+        {
+            "\"\xBF\"",
+            "\xBF",             /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 3.1.3  2 continuation bytes */
+        {
+            "\"\x80\xBF\"",
+            "\x80\xBF",         /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        /* 3.1.4  3 continuation bytes */
+        {
+            "\"\x80\xBF\x80\"",
+            "\x80\xBF\x80",     /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\\uFFFD\"",
+        },
+        /* 3.1.5  4 continuation bytes */
+        {
+            "\"\x80\xBF\x80\xBF\"",
+            "\x80\xBF\x80\xBF", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\"",
+        },
+        /* 3.1.6  5 continuation bytes */
+        {
+            "\"\x80\xBF\x80\xBF\x80\"",
+            "\x80\xBF\x80\xBF\x80", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\"",
+        },
+        /* 3.1.7  6 continuation bytes */
+        {
+            "\"\x80\xBF\x80\xBF\x80\xBF\"",
+            "\x80\xBF\x80\xBF\x80\xBF", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\"",
+        },
+        /* 3.1.8  7 continuation bytes */
+        {
+            "\"\x80\xBF\x80\xBF\x80\xBF\x80\"",
+            "\x80\xBF\x80\xBF\x80\xBF\x80", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\"",
+        },
+        /* 3.1.9  Sequence of all 64 possible continuation bytes */
+        {
+            "\"\x80\x81\x82\x83\x84\x85\x86\x87"
+            "\x88\x89\x8A\x8B\x8C\x8D\x8E\x8F"
+            "\x90\x91\x92\x93\x94\x95\x96\x97"
+            "\x98\x99\x9A\x9B\x9C\x9D\x9E\x9F"
+            "\xA0\xA1\xA2\xA3\xA4\xA5\xA6\xA7"
+            "\xA8\xA9\xAA\xAB\xAC\xAD\xAE\xAF"
+            "\xB0\xB1\xB2\xB3\xB4\xB5\xB6\xB7"
+            "\xB8\xB9\xBA\xBB\xBC\xBD\xBE\xBF\"",
+             /* bug: not corrected */
+            "\x80\x81\x82\x83\x84\x85\x86\x87"
+            "\x88\x89\x8A\x8B\x8C\x8D\x8E\x8F"
+            "\x90\x91\x92\x93\x94\x95\x96\x97"
+            "\x98\x99\x9A\x9B\x9C\x9D\x9E\x9F"
+            "\xA0\xA1\xA2\xA3\xA4\xA5\xA6\xA7"
+            "\xA8\xA9\xAA\xAB\xAC\xAD\xAE\xAF"
+            "\xB0\xB1\xB2\xB3\xB4\xB5\xB6\xB7"
+            "\xB8\xB9\xBA\xBB\xBC\xBD\xBE\xBF",
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\""
+        },
+        /* 3.2  Lonely start characters */
+        /* 3.2.1  All 32 first bytes of 2-byte sequences, followed by space */
+        {
+            "\"\xC0 \xC1 \xC2 \xC3 \xC4 \xC5 \xC6 \xC7 "
+            "\xC8 \xC9 \xCA \xCB \xCC \xCD \xCE \xCF "
+            "\xD0 \xD1 \xD2 \xD3 \xD4 \xD5 \xD6 \xD7 "
+            "\xD8 \xD9 \xDA \xDB \xDC \xDD \xDE \xDF \"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD "
+            "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD "
+            "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD "
+            "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \"",
+            "\xC0 \xC1 \xC2 \xC3 \xC4 \xC5 \xC6 \xC7 "
+            "\xC8 \xC9 \xCA \xCB \xCC \xCD \xCE \xCF "
+            "\xD0 \xD1 \xD2 \xD3 \xD4 \xD5 \xD6 \xD7 "
+            "\xD8 \xD9 \xDA \xDB \xDC \xDD \xDE \xDF ",
+        },
+        /* 3.2.2  All 16 first bytes of 3-byte sequences, followed by space */
+        {
+            "\"\xE0 \xE1 \xE2 \xE3 \xE4 \xE5 \xE6 \xE7 "
+            "\xE8 \xE9 \xEA \xEB \xEC \xED \xEE \xEF \"",
+            /* bug: not corrected */
+            "\xE0 \xE1 \xE2 \xE3 \xE4 \xE5 \xE6 \xE7 "
+            "\xE8 \xE9 \xEA \xEB \xEC \xED \xEE \xEF ",
+            "\"\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD "
+            "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \"",
+        },
+        /* 3.2.3  All 8 first bytes of 4-byte sequences, followed by space */
+        {
+            "\"\xF0 \xF1 \xF2 \xF3 \xF4 \xF5 \xF6 \xF7 \"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \"",
+            "\xF0 \xF1 \xF2 \xF3 \xF4 \xF5 \xF6 \xF7 ",
+        },
+        /* 3.2.4  All 4 first bytes of 5-byte sequences, followed by space */
+        {
+            "\"\xF8 \xF9 \xFA \xFB \"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD \\uFFFD \\uFFFD \\uFFFD \"",
+            "\xF8 \xF9 \xFA \xFB ",
+        },
+        /* 3.2.5  All 2 first bytes of 6-byte sequences, followed by space */
+        {
+            "\"\xFC \xFD \"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD \\uFFFD \"",
+            "\xFC \xFD ",
+        },
+        /* 3.3  Sequences with last continuation byte missing */
+        /* 3.3.1  2-byte sequence with last byte missing (U+0000) */
+        {
+            "\"\xC0\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xC0",
+        },
+        /* 3.3.2  3-byte sequence with last byte missing (U+0000) */
+        {
+            "\"\xE0\x80\"",
+            "\xE0\x80",           /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 3.3.3  4-byte sequence with last byte missing (U+0000) */
+        {
+            "\"\xF0\x80\x80\"",
+            "\xF0\x80\x80",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 3.3.4  5-byte sequence with last byte missing (U+0000) */
+        {
+            "\"\xF8\x80\x80\x80\"",
+            NULL,                   /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xF8\x80\x80\x80",
+        },
+        /* 3.3.5  6-byte sequence with last byte missing (U+0000) */
+        {
+            "\"\xFC\x80\x80\x80\x80\"",
+            NULL,                        /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFC\x80\x80\x80\x80",
+        },
+        /* 3.3.6  2-byte sequence with last byte missing (U+07FF) */
+        {
+            "\"\xDF\"",
+            "\xDF",             /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 3.3.7  3-byte sequence with last byte missing (U+FFFF) */
+        {
+            "\"\xEF\xBF\"",
+            "\xEF\xBF",           /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 3.3.8  4-byte sequence with last byte missing (U+1FFFFF) */
+        {
+            "\"\xF7\xBF\xBF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xF7\xBF\xBF",
+        },
+        /* 3.3.9  5-byte sequence with last byte missing (U+3FFFFFF) */
+        {
+            "\"\xFB\xBF\xBF\xBF\"",
+            NULL,                 /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFB\xBF\xBF\xBF",
+        },
+        /* 3.3.10  6-byte sequence with last byte missing (U+7FFFFFFF) */
+        {
+            "\"\xFD\xBF\xBF\xBF\xBF\"",
+            NULL,                        /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFD\xBF\xBF\xBF\xBF",
+        },
+        /* 3.4  Concatenation of incomplete sequences */
+        {
+            "\"\xC0\xE0\x80\xF0\x80\x80\xF8\x80\x80\x80\xFC\x80\x80\x80\x80"
+            "\xDF\xEF\xBF\xF7\xBF\xBF\xFB\xBF\xBF\xBF\xFD\xBF\xBF\xBF\xBF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\"",
+            "\xC0\xE0\x80\xF0\x80\x80\xF8\x80\x80\x80\xFC\x80\x80\x80\x80"
+            "\xDF\xEF\xBF\xF7\xBF\xBF\xFB\xBF\xBF\xBF\xFD\xBF\xBF\xBF\xBF",
+        },
+        /* 3.5  Impossible bytes */
+        {
+            "\"\xFE\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFE",
+        },
+        {
+            "\"\xFF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFF",
+        },
+        {
+            "\"\xFE\xFE\xFF\xFF\"",
+            NULL,                 /* bug: rejected */
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\"",
+            "\xFE\xFE\xFF\xFF",
+        },
+        /* 4  Overlong sequences */
+        /* 4.1  Overlong '/' */
+        {
+            "\"\xC0\xAF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xC0\xAF",
+        },
+        {
+            "\"\xE0\x80\xAF\"",
+            "\xE0\x80\xAF",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            "\"\xF0\x80\x80\xAF\"",
+            "\xF0\x80\x80\xAF",  /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            "\"\xF8\x80\x80\x80\xAF\"",
+            NULL,                        /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xF8\x80\x80\x80\xAF",
+        },
+        {
+            "\"\xFC\x80\x80\x80\x80\xAF\"",
+            NULL,                               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFC\x80\x80\x80\x80\xAF",
+        },
+        /*
+         * 4.2  Maximum overlong sequences
+         * Highest Unicode value that is still resulting in an
+         * overlong sequence if represented with the given number of
+         * bytes.  This is a boundary test for safe UTF-8 decoders.
+         */
+        {
+            /* \U+007F */
+            "\"\xC1\xBF\"",
+            NULL,               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xC1\xBF",
+        },
+        {
+            /* \U+07FF */
+            "\"\xE0\x9F\xBF\"",
+            "\xE0\x9F\xBF",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /*
+             * \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 */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+1FFFFF */
+            "\"\xF8\x87\xBF\xBF\xBF\"",
+            NULL,                        /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xF8\x87\xBF\xBF\xBF",
+        },
+        {
+            /* \U+3FFFFFF */
+            "\"\xFC\x83\xBF\xBF\xBF\xBF\"",
+            NULL,                               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFC\x83\xBF\xBF\xBF\xBF",
+        },
+        /* 4.3  Overlong representation of the NUL character */
+        {
+            /* \U+0000 */
+            "\"\xC0\x80\"",
+            NULL,               /* bug: rejected */
+            "\"\\u0000\"",
+            "\xC0\x80",
+        },
+        {
+            /* \U+0000 */
+            "\"\xE0\x80\x80\"",
+            "\xE0\x80\x80",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+0000 */
+            "\"\xF0\x80\x80\x80\"",
+            "\xF0\x80\x80\x80",   /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+0000 */
+            "\"\xF8\x80\x80\x80\x80\"",
+            NULL,                        /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xF8\x80\x80\x80\x80",
+        },
+        {
+            /* \U+0000 */
+            "\"\xFC\x80\x80\x80\x80\x80\"",
+            NULL,                               /* bug: rejected */
+            "\"\\uFFFD\"",
+            "\xFC\x80\x80\x80\x80\x80",
+        },
+        /* 5  Illegal code positions */
+        /* 5.1  Single UTF-16 surrogates */
+        {
+            /* \U+D800 */
+            "\"\xED\xA0\x80\"",
+            "\xED\xA0\x80",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+DB7F */
+            "\"\xED\xAD\xBF\"",
+            "\xED\xAD\xBF",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+DB80 */
+            "\"\xED\xAE\x80\"",
+            "\xED\xAE\x80",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+DBFF */
+            "\"\xED\xAF\xBF\"",
+            "\xED\xAF\xBF",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+DC00 */
+            "\"\xED\xB0\x80\"",
+            "\xED\xB0\x80",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+DF80 */
+            "\"\xED\xBE\x80\"",
+            "\xED\xBE\x80",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+DFFF */
+            "\"\xED\xBF\xBF\"",
+            "\xED\xBF\xBF",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 5.2  Paired UTF-16 surrogates */
+        {
+            /* \U+D800\U+DC00 */
+            "\"\xED\xA0\x80\xED\xB0\x80\"",
+            "\xED\xA0\x80\xED\xB0\x80", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        {
+            /* \U+D800\U+DFFF */
+            "\"\xED\xA0\x80\xED\xBF\xBF\"",
+            "\xED\xA0\x80\xED\xBF\xBF", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        {
+            /* \U+DB7F\U+DC00 */
+            "\"\xED\xAD\xBF\xED\xB0\x80\"",
+            "\xED\xAD\xBF\xED\xB0\x80", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        {
+            /* \U+DB7F\U+DFFF */
+            "\"\xED\xAD\xBF\xED\xBF\xBF\"",
+            "\xED\xAD\xBF\xED\xBF\xBF", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        {
+            /* \U+DB80\U+DC00 */
+            "\"\xED\xAE\x80\xED\xB0\x80\"",
+            "\xED\xAE\x80\xED\xB0\x80", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        {
+            /* \U+DB80\U+DFFF */
+            "\"\xED\xAE\x80\xED\xBF\xBF\"",
+            "\xED\xAE\x80\xED\xBF\xBF", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        {
+            /* \U+DBFF\U+DC00 */
+            "\"\xED\xAF\xBF\xED\xB0\x80\"",
+            "\xED\xAF\xBF\xED\xB0\x80", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        {
+            /* \U+DBFF\U+DFFF */
+            "\"\xED\xAF\xBF\xED\xBF\xBF\"",
+            "\xED\xAF\xBF\xED\xBF\xBF", /* bug: not corrected */
+            "\"\\uFFFD\\uFFFD\"",
+        },
+        /* 5.3  Other illegal code positions */
+        /* BMP noncharacters */
+        {
+            /* \U+FFFE */
+            "\"\xEF\xBF\xBE\"",
+            "\xEF\xBF\xBE",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* \U+FFFF */
+            "\"\xEF\xBF\xBF\"",
+            "\xEF\xBF\xBF",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* U+FDD0 */
+            "\"\xEF\xB7\x90\"",
+            "\xEF\xB7\x90",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        {
+            /* U+FDEF */
+            "\"\xEF\xB7\xAF\"",
+            "\xEF\xB7\xAF",     /* bug: not corrected */
+            "\"\\uFFFD\"",
+        },
+        /* 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",
+            "\"\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD"
+            "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\"",
+        },
+        {}
+    };
+    int i;
+    QObject *obj;
+    QString *str;
+    const char *json_in, *utf8_out, *utf8_in, *json_out;
+
+    for (i = 0; test_cases[i].json_in; i++) {
+        json_in = test_cases[i].json_in;
+        utf8_out = test_cases[i].utf8_out;
+        utf8_in = test_cases[i].utf8_in ?: test_cases[i].utf8_out;
+        json_out = test_cases[i].json_out ?: test_cases[i].json_in;
+
+        obj = qobject_from_json(json_in);
+        if (utf8_out) {
+            g_assert(obj);
+            g_assert(qobject_type(obj) == QTYPE_QSTRING);
+            str = qobject_to_qstring(obj);
+            g_assert_cmpstr(qstring_get_str(str), ==, utf8_out);
+        } else {
+            g_assert(!obj);
+        }
+        qobject_decref(obj);
+
+        obj = QOBJECT(qstring_from_str(utf8_in));
+        str = qobject_to_json(obj);
+        if (json_out) {
+            g_assert(str);
+            g_assert_cmpstr(qstring_get_str(str), ==, json_out);
+        } else {
+            g_assert(!str);
+        }
+        QDECREF(str);
+        qobject_decref(obj);
+
+        /*
+         * Disabled, because qobject_from_json() is buggy, and I can't
+         * be bothered to add the expected incorrect results.
+         * FIXME Enable once these bugs have been fixed.
+         */
+        if (0 && json_out != json_in) {
+            obj = qobject_from_json(json_out);
+            g_assert(obj);
+            g_assert(qobject_type(obj) == QTYPE_QSTRING);
+            str = qobject_to_qstring(obj);
+            g_assert_cmpstr(qstring_get_str(str), ==, utf8_out);
+        }
+    }
+}
+
+static void vararg_string(void)
+{
+    int i;
+    struct {
+        const char *decoded;
+    } test_cases[] = {
+        { "hello world" },
+        { "the quick brown fox jumped over the fence" },
+        {}
+    };
+
+    for (i = 0; test_cases[i].decoded; i++) {
+        QObject *obj;
+        QString *str;
+
+        obj = qobject_from_jsonf("%s", test_cases[i].decoded);
+
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QSTRING);
+        
+        str = qobject_to_qstring(obj);
+        g_assert(strcmp(qstring_get_str(str), test_cases[i].decoded) == 0);
+
+        QDECREF(str);
+    }
+}
+
+static void simple_number(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        int64_t decoded;
+        int skip;
+    } test_cases[] = {
+        { "0", 0 },
+        { "1234", 1234 },
+        { "1", 1 },
+        { "-32", -32 },
+        { "-0", 0, .skip = 1 },
+        { },
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QInt *qint;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QINT);
+
+        qint = qobject_to_qint(obj);
+        g_assert(qint_get_int(qint) == test_cases[i].decoded);
+        if (test_cases[i].skip == 0) {
+            QString *str;
+
+            str = qobject_to_json(obj);
+            g_assert(strcmp(qstring_get_str(str), test_cases[i].encoded) == 0);
+            QDECREF(str);
+        }
+
+        QDECREF(qint);
+    }
+}
+
+static void float_number(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        double decoded;
+        int skip;
+    } test_cases[] = {
+        { "32.43", 32.43 },
+        { "0.222", 0.222 },
+        { "-32.12313", -32.12313 },
+        { "-32.20e-10", -32.20e-10, .skip = 1 },
+        { },
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QFloat *qfloat;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QFLOAT);
+
+        qfloat = qobject_to_qfloat(obj);
+        g_assert(qfloat_get_double(qfloat) == test_cases[i].decoded);
+
+        if (test_cases[i].skip == 0) {
+            QString *str;
+
+            str = qobject_to_json(obj);
+            g_assert(strcmp(qstring_get_str(str), test_cases[i].encoded) == 0);
+            QDECREF(str);
+        }
+
+        QDECREF(qfloat);
+    }
+}
+
+static void vararg_number(void)
+{
+    QObject *obj;
+    QInt *qint;
+    QFloat *qfloat;
+    int value = 0x2342;
+    int64_t value64 = 0x2342342343LL;
+    double valuef = 2.323423423;
+
+    obj = qobject_from_jsonf("%d", value);
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QINT);
+
+    qint = qobject_to_qint(obj);
+    g_assert(qint_get_int(qint) == value);
+
+    QDECREF(qint);
+
+    obj = qobject_from_jsonf("%" PRId64, value64);
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QINT);
+
+    qint = qobject_to_qint(obj);
+    g_assert(qint_get_int(qint) == value64);
+
+    QDECREF(qint);
+
+    obj = qobject_from_jsonf("%f", valuef);
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QFLOAT);
+
+    qfloat = qobject_to_qfloat(obj);
+    g_assert(qfloat_get_double(qfloat) == valuef);
+
+    QDECREF(qfloat);
+}
+
+static void keyword_literal(void)
+{
+    QObject *obj;
+    QBool *qbool;
+    QObject *null;
+    QString *str;
+
+    obj = qobject_from_json("true");
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QBOOL);
+
+    qbool = qobject_to_qbool(obj);
+    g_assert(qbool_get_bool(qbool) == true);
+
+    str = qobject_to_json(obj);
+    g_assert(strcmp(qstring_get_str(str), "true") == 0);
+    QDECREF(str);
+
+    QDECREF(qbool);
+
+    obj = qobject_from_json("false");
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QBOOL);
+
+    qbool = qobject_to_qbool(obj);
+    g_assert(qbool_get_bool(qbool) == false);
+
+    str = qobject_to_json(obj);
+    g_assert(strcmp(qstring_get_str(str), "false") == 0);
+    QDECREF(str);
+
+    QDECREF(qbool);
+
+    obj = qobject_from_jsonf("%i", false);
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QBOOL);
+
+    qbool = qobject_to_qbool(obj);
+    g_assert(qbool_get_bool(qbool) == false);
+
+    QDECREF(qbool);
+
+    /* Test that non-zero values other than 1 get collapsed to true */
+    obj = qobject_from_jsonf("%i", 2);
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QBOOL);
+
+    qbool = qobject_to_qbool(obj);
+    g_assert(qbool_get_bool(qbool) == true);
+
+    QDECREF(qbool);
+
+    obj = qobject_from_json("null");
+    g_assert(obj != NULL);
+    g_assert(qobject_type(obj) == QTYPE_QNULL);
+
+    null = qnull();
+    g_assert(null == obj);
+
+    qobject_decref(obj);
+    qobject_decref(null);
+}
+
+typedef struct LiteralQDictEntry LiteralQDictEntry;
+typedef struct LiteralQObject LiteralQObject;
+
+struct LiteralQObject
+{
+    int type;
+    union {
+        int64_t qint;
+        const char *qstr;
+        LiteralQDictEntry *qdict;
+        LiteralQObject *qlist;
+    } value;
+};
+
+struct LiteralQDictEntry
+{
+    const char *key;
+    LiteralQObject value;
+};
+
+#define QLIT_QINT(val) (LiteralQObject){.type = QTYPE_QINT, .value.qint = (val)}
+#define QLIT_QSTR(val) (LiteralQObject){.type = QTYPE_QSTRING, .value.qstr = (val)}
+#define QLIT_QDICT(val) (LiteralQObject){.type = QTYPE_QDICT, .value.qdict = (val)}
+#define QLIT_QLIST(val) (LiteralQObject){.type = QTYPE_QLIST, .value.qlist = (val)}
+
+typedef struct QListCompareHelper
+{
+    int index;
+    LiteralQObject *objs;
+    int result;
+} QListCompareHelper;
+
+static int compare_litqobj_to_qobj(LiteralQObject *lhs, QObject *rhs);
+
+static void compare_helper(QObject *obj, void *opaque)
+{
+    QListCompareHelper *helper = opaque;
+
+    if (helper->result == 0) {
+        return;
+    }
+
+    if (helper->objs[helper->index].type == QTYPE_NONE) {
+        helper->result = 0;
+        return;
+    }
+
+    helper->result = compare_litqobj_to_qobj(&helper->objs[helper->index++], obj);
+}
+
+static int compare_litqobj_to_qobj(LiteralQObject *lhs, QObject *rhs)
+{
+    if (lhs->type != qobject_type(rhs)) {
+        return 0;
+    }
+
+    switch (lhs->type) {
+    case QTYPE_QINT:
+        return lhs->value.qint == qint_get_int(qobject_to_qint(rhs));
+    case QTYPE_QSTRING:
+        return (strcmp(lhs->value.qstr, qstring_get_str(qobject_to_qstring(rhs))) == 0);
+    case QTYPE_QDICT: {
+        int i;
+
+        for (i = 0; lhs->value.qdict[i].key; i++) {
+            QObject *obj = qdict_get(qobject_to_qdict(rhs), lhs->value.qdict[i].key);
+
+            if (!compare_litqobj_to_qobj(&lhs->value.qdict[i].value, obj)) {
+                return 0;
+            }
+        }
+
+        return 1;
+    }
+    case QTYPE_QLIST: {
+        QListCompareHelper helper;
+
+        helper.index = 0;
+        helper.objs = lhs->value.qlist;
+        helper.result = 1;
+        
+        qlist_iter(qobject_to_qlist(rhs), compare_helper, &helper);
+
+        return helper.result;
+    }
+    default:
+        break;
+    }
+
+    return 0;
+}
+
+static void simple_dict(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        LiteralQObject decoded;
+    } test_cases[] = {
+        {
+            .encoded = "{\"foo\": 42, \"bar\": \"hello world\"}",
+            .decoded = QLIT_QDICT(((LiteralQDictEntry[]){
+                        { "foo", QLIT_QINT(42) },
+                        { "bar", QLIT_QSTR("hello world") },
+                        { }
+                    })),
+        }, {
+            .encoded = "{}",
+            .decoded = QLIT_QDICT(((LiteralQDictEntry[]){
+                        { }
+                    })),
+        }, {
+            .encoded = "{\"foo\": 43}",
+            .decoded = QLIT_QDICT(((LiteralQDictEntry[]){
+                        { "foo", QLIT_QINT(43) },
+                        { }
+                    })),
+        },
+        { }
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QString *str;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QDICT);
+
+        g_assert(compare_litqobj_to_qobj(&test_cases[i].decoded, obj) == 1);
+
+        str = qobject_to_json(obj);
+        qobject_decref(obj);
+
+        obj = qobject_from_json(qstring_get_str(str));
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QDICT);
+
+        g_assert(compare_litqobj_to_qobj(&test_cases[i].decoded, obj) == 1);
+        qobject_decref(obj);
+        QDECREF(str);
+    }
+}
+
+/*
+ * this generates json of the form:
+ * a(0,m) = [0, 1, ..., m-1]
+ * a(n,m) = {
+ *            'key0': a(0,m),
+ *            'key1': a(1,m),
+ *            ...
+ *            'key(n-1)': a(n-1,m)
+ *          }
+ */
+static void gen_test_json(GString *gstr, int nest_level_max,
+                          int elem_count)
+{
+    int i;
+
+    g_assert(gstr);
+    if (nest_level_max == 0) {
+        g_string_append(gstr, "[");
+        for (i = 0; i < elem_count; i++) {
+            g_string_append_printf(gstr, "%d", i);
+            if (i < elem_count - 1) {
+                g_string_append_printf(gstr, ", ");
+            }
+        }
+        g_string_append(gstr, "]");
+        return;
+    }
+
+    g_string_append(gstr, "{");
+    for (i = 0; i < nest_level_max; i++) {
+        g_string_append_printf(gstr, "'key%d': ", i);
+        gen_test_json(gstr, i, elem_count);
+        if (i < nest_level_max - 1) {
+            g_string_append(gstr, ",");
+        }
+    }
+    g_string_append(gstr, "}");
+}
+
+static void large_dict(void)
+{
+    GString *gstr = g_string_new("");
+    QObject *obj;
+
+    gen_test_json(gstr, 10, 100);
+    obj = qobject_from_json(gstr->str);
+    g_assert(obj != NULL);
+
+    qobject_decref(obj);
+    g_string_free(gstr, true);
+}
+
+static void simple_list(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        LiteralQObject decoded;
+    } test_cases[] = {
+        {
+            .encoded = "[43,42]",
+            .decoded = QLIT_QLIST(((LiteralQObject[]){
+                        QLIT_QINT(43),
+                        QLIT_QINT(42),
+                        { }
+                    })),
+        },
+        {
+            .encoded = "[43]",
+            .decoded = QLIT_QLIST(((LiteralQObject[]){
+                        QLIT_QINT(43),
+                        { }
+                    })),
+        },
+        {
+            .encoded = "[]",
+            .decoded = QLIT_QLIST(((LiteralQObject[]){
+                        { }
+                    })),
+        },
+        {
+            .encoded = "[{}]",
+            .decoded = QLIT_QLIST(((LiteralQObject[]){
+                        QLIT_QDICT(((LiteralQDictEntry[]){
+                                    {},
+                                        })),
+                        {},
+                            })),
+        },
+        { }
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QString *str;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QLIST);
+
+        g_assert(compare_litqobj_to_qobj(&test_cases[i].decoded, obj) == 1);
+
+        str = qobject_to_json(obj);
+        qobject_decref(obj);
+
+        obj = qobject_from_json(qstring_get_str(str));
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QLIST);
+
+        g_assert(compare_litqobj_to_qobj(&test_cases[i].decoded, obj) == 1);
+        qobject_decref(obj);
+        QDECREF(str);
+    }
+}
+
+static void simple_whitespace(void)
+{
+    int i;
+    struct {
+        const char *encoded;
+        LiteralQObject decoded;
+    } test_cases[] = {
+        {
+            .encoded = " [ 43 , 42 ]",
+            .decoded = QLIT_QLIST(((LiteralQObject[]){
+                        QLIT_QINT(43),
+                        QLIT_QINT(42),
+                        { }
+                    })),
+        },
+        {
+            .encoded = " [ 43 , { 'h' : 'b' }, [ ], 42 ]",
+            .decoded = QLIT_QLIST(((LiteralQObject[]){
+                        QLIT_QINT(43),
+                        QLIT_QDICT(((LiteralQDictEntry[]){
+                                    { "h", QLIT_QSTR("b") },
+                                    { }})),
+                        QLIT_QLIST(((LiteralQObject[]){
+                                    { }})),
+                        QLIT_QINT(42),
+                        { }
+                    })),
+        },
+        {
+            .encoded = " [ 43 , { 'h' : 'b' , 'a' : 32 }, [ ], 42 ]",
+            .decoded = QLIT_QLIST(((LiteralQObject[]){
+                        QLIT_QINT(43),
+                        QLIT_QDICT(((LiteralQDictEntry[]){
+                                    { "h", QLIT_QSTR("b") },
+                                    { "a", QLIT_QINT(32) },
+                                    { }})),
+                        QLIT_QLIST(((LiteralQObject[]){
+                                    { }})),
+                        QLIT_QINT(42),
+                        { }
+                    })),
+        },
+        { }
+    };
+
+    for (i = 0; test_cases[i].encoded; i++) {
+        QObject *obj;
+        QString *str;
+
+        obj = qobject_from_json(test_cases[i].encoded);
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QLIST);
+
+        g_assert(compare_litqobj_to_qobj(&test_cases[i].decoded, obj) == 1);
+
+        str = qobject_to_json(obj);
+        qobject_decref(obj);
+
+        obj = qobject_from_json(qstring_get_str(str));
+        g_assert(obj != NULL);
+        g_assert(qobject_type(obj) == QTYPE_QLIST);
+
+        g_assert(compare_litqobj_to_qobj(&test_cases[i].decoded, obj) == 1);
+
+        qobject_decref(obj);
+        QDECREF(str);
+    }
+}
+
+static void simple_varargs(void)
+{
+    QObject *embedded_obj;
+    QObject *obj;
+    LiteralQObject decoded = QLIT_QLIST(((LiteralQObject[]){
+            QLIT_QINT(1),
+            QLIT_QINT(2),
+            QLIT_QLIST(((LiteralQObject[]){
+                        QLIT_QINT(32),
+                        QLIT_QINT(42),
+                        {}})),
+            {}}));
+
+    embedded_obj = qobject_from_json("[32, 42]");
+    g_assert(embedded_obj != NULL);
+
+    obj = qobject_from_jsonf("[%d, 2, %p]", 1, embedded_obj);
+    g_assert(obj != NULL);
+
+    g_assert(compare_litqobj_to_qobj(&decoded, obj) == 1);
+
+    qobject_decref(obj);
+}
+
+static void empty_input(void)
+{
+    const char *empty = "";
+
+    QObject *obj = qobject_from_json(empty);
+    g_assert(obj == NULL);
+}
+
+static void unterminated_string(void)
+{
+    QObject *obj = qobject_from_json("\"abc");
+    g_assert(obj == NULL);
+}
+
+static void unterminated_sq_string(void)
+{
+    QObject *obj = qobject_from_json("'abc");
+    g_assert(obj == NULL);
+}
+
+static void unterminated_escape(void)
+{
+    QObject *obj = qobject_from_json("\"abc\\\"");
+    g_assert(obj == NULL);
+}
+
+static void unterminated_array(void)
+{
+    QObject *obj = qobject_from_json("[32");
+    g_assert(obj == NULL);
+}
+
+static void unterminated_array_comma(void)
+{
+    QObject *obj = qobject_from_json("[32,");
+    g_assert(obj == NULL);
+}
+
+static void invalid_array_comma(void)
+{
+    QObject *obj = qobject_from_json("[32,}");
+    g_assert(obj == NULL);
+}
+
+static void unterminated_dict(void)
+{
+    QObject *obj = qobject_from_json("{'abc':32");
+    g_assert(obj == NULL);
+}
+
+static void unterminated_dict_comma(void)
+{
+    QObject *obj = qobject_from_json("{'abc':32,");
+    g_assert(obj == NULL);
+}
+
+static void invalid_dict_comma(void)
+{
+    QObject *obj = qobject_from_json("{'abc':32,}");
+    g_assert(obj == NULL);
+}
+
+static void unterminated_literal(void)
+{
+    QObject *obj = qobject_from_json("nul");
+    g_assert(obj == NULL);
+}
+
+int main(int argc, char **argv)
+{
+    g_test_init(&argc, &argv, NULL);
+
+    g_test_add_func("/literals/string/simple", simple_string);
+    g_test_add_func("/literals/string/escaped", escaped_string);
+    g_test_add_func("/literals/string/utf8", utf8_string);
+    g_test_add_func("/literals/string/single_quote", single_quote_string);
+    g_test_add_func("/literals/string/vararg", vararg_string);
+
+    g_test_add_func("/literals/number/simple", simple_number);
+    g_test_add_func("/literals/number/float", float_number);
+    g_test_add_func("/literals/number/vararg", vararg_number);
+
+    g_test_add_func("/literals/keyword", keyword_literal);
+
+    g_test_add_func("/dicts/simple_dict", simple_dict);
+    g_test_add_func("/dicts/large_dict", large_dict);
+    g_test_add_func("/lists/simple_list", simple_list);
+
+    g_test_add_func("/whitespace/simple_whitespace", simple_whitespace);
+
+    g_test_add_func("/varargs/simple_varargs", simple_varargs);
+
+    g_test_add_func("/errors/empty_input", empty_input);
+    g_test_add_func("/errors/unterminated/string", unterminated_string);
+    g_test_add_func("/errors/unterminated/escape", unterminated_escape);
+    g_test_add_func("/errors/unterminated/sq_string", unterminated_sq_string);
+    g_test_add_func("/errors/unterminated/array", unterminated_array);
+    g_test_add_func("/errors/unterminated/array_comma", unterminated_array_comma);
+    g_test_add_func("/errors/unterminated/dict", unterminated_dict);
+    g_test_add_func("/errors/unterminated/dict_comma", unterminated_dict_comma);
+    g_test_add_func("/errors/invalid_array_comma", invalid_array_comma);
+    g_test_add_func("/errors/invalid_dict_comma", invalid_dict_comma);
+    g_test_add_func("/errors/unterminated/literal", unterminated_literal);
+
+    return g_test_run();
+}