bson 4.7.1 → 4.8.0

data/spec/spec_tests/data/corpus/double.json

@@ -0,0 +1,87 @@
+{
+    "description": "Double type",
+    "bson_type": "0x01",
+    "test_key": "d",
+    "valid": [
+        {
+            "description": "+1.0",
+            "canonical_bson": "10000000016400000000000000F03F00",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"1.0\"}}",
+            "relaxed_extjson": "{\"d\" : 1.0}"
+        },
+        {
+            "description": "-1.0",
+            "canonical_bson": "10000000016400000000000000F0BF00",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"-1.0\"}}",
+            "relaxed_extjson": "{\"d\" : -1.0}"
+        },
+        {
+            "description": "+1.0001220703125",
+            "canonical_bson": "10000000016400000000008000F03F00",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"1.0001220703125\"}}",
+            "relaxed_extjson": "{\"d\" : 1.0001220703125}"
+        },
+        {
+            "description": "-1.0001220703125",
+            "canonical_bson": "10000000016400000000008000F0BF00",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"-1.0001220703125\"}}",
+            "relaxed_extjson": "{\"d\" : -1.0001220703125}"
+        },
+        {
+            "description": "1.2345678921232E+18",
+            "canonical_bson": "100000000164002a1bf5f41022b14300",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"1.2345678921232E+18\"}}",
+            "relaxed_extjson": "{\"d\" : 1.2345678921232E+18}"
+        },
+        {
+            "description": "-1.2345678921232E+18",
+            "canonical_bson": "100000000164002a1bf5f41022b1c300",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"-1.2345678921232E+18\"}}",
+            "relaxed_extjson": "{\"d\" : -1.2345678921232E+18}"
+        },
+        {
+            "description": "0.0",
+            "canonical_bson": "10000000016400000000000000000000",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"0.0\"}}",
+            "relaxed_extjson": "{\"d\" : 0.0}"
+        },
+        {
+            "description": "-0.0",
+            "canonical_bson": "10000000016400000000000000008000",
+            "canonical_extjson": "{\"d\" : {\"$numberDouble\": \"-0.0\"}}",
+            "relaxed_extjson": "{\"d\" : -0.0}"
+        },
+        {
+            "description": "NaN",
+            "canonical_bson": "10000000016400000000000000F87F00",
+            "canonical_extjson": "{\"d\": {\"$numberDouble\": \"NaN\"}}",
+            "relaxed_extjson": "{\"d\": {\"$numberDouble\": \"NaN\"}}",
+            "lossy": true
+        },
+        {
+            "description": "NaN with payload",
+            "canonical_bson": "10000000016400120000000000F87F00",
+            "canonical_extjson": "{\"d\": {\"$numberDouble\": \"NaN\"}}",
+            "relaxed_extjson": "{\"d\": {\"$numberDouble\": \"NaN\"}}",
+            "lossy": true
+        },
+        {
+            "description": "Inf",
+            "canonical_bson": "10000000016400000000000000F07F00",
+            "canonical_extjson": "{\"d\": {\"$numberDouble\": \"Infinity\"}}",
+            "relaxed_extjson": "{\"d\": {\"$numberDouble\": \"Infinity\"}}"
+        },
+        {
+            "description": "-Inf",
+            "canonical_bson": "10000000016400000000000000F0FF00",
+            "canonical_extjson": "{\"d\": {\"$numberDouble\": \"-Infinity\"}}",
+            "relaxed_extjson": "{\"d\": {\"$numberDouble\": \"-Infinity\"}}"
+        }
+    ],
+    "decodeErrors": [
+        {
+            "description": "double truncated",
+            "bson": "0B0000000164000000F03F00"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/int32.json

@@ -0,0 +1,43 @@
+{
+    "description": "Int32 type",
+    "bson_type": "0x10",
+    "test_key": "i",
+    "valid": [
+        {
+            "description": "MinValue",
+            "canonical_bson": "0C0000001069000000008000",
+            "canonical_extjson": "{\"i\" : {\"$numberInt\": \"-2147483648\"}}",
+            "relaxed_extjson": "{\"i\" : -2147483648}"
+        },
+        {
+            "description": "MaxValue",
+            "canonical_bson": "0C000000106900FFFFFF7F00",
+            "canonical_extjson": "{\"i\" : {\"$numberInt\": \"2147483647\"}}",
+            "relaxed_extjson": "{\"i\" : 2147483647}"
+        },
+        {
+            "description": "-1",
+            "canonical_bson": "0C000000106900FFFFFFFF00",
+            "canonical_extjson": "{\"i\" : {\"$numberInt\": \"-1\"}}",
+            "relaxed_extjson": "{\"i\" : -1}"
+        },
+        {
+            "description": "0",
+            "canonical_bson": "0C0000001069000000000000",
+            "canonical_extjson": "{\"i\" : {\"$numberInt\": \"0\"}}",
+            "relaxed_extjson": "{\"i\" : 0}"
+        },
+        {
+            "description": "1",
+            "canonical_bson": "0C0000001069000100000000",
+            "canonical_extjson": "{\"i\" : {\"$numberInt\": \"1\"}}",
+            "relaxed_extjson": "{\"i\" : 1}"
+        }
+    ],
+    "decodeErrors": [
+        {
+            "description": "Bad int32 field length",
+            "bson": "090000001061000500"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/int64.json

@@ -0,0 +1,43 @@
+{
+    "description": "Int64 type",
+    "bson_type": "0x12",
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "MinValue",
+            "canonical_bson": "10000000126100000000000000008000",
+            "canonical_extjson": "{\"a\" : {\"$numberLong\" : \"-9223372036854775808\"}}",
+            "relaxed_extjson": "{\"a\" : -9223372036854775808}"
+        },
+        {
+            "description": "MaxValue",
+            "canonical_bson": "10000000126100FFFFFFFFFFFFFF7F00",
+            "canonical_extjson": "{\"a\" : {\"$numberLong\" : \"9223372036854775807\"}}",
+            "relaxed_extjson": "{\"a\" : 9223372036854775807}"
+        },
+        {
+            "description": "-1",
+            "canonical_bson": "10000000126100FFFFFFFFFFFFFFFF00",
+            "canonical_extjson": "{\"a\" : {\"$numberLong\" : \"-1\"}}",
+            "relaxed_extjson": "{\"a\" : -1}"
+        },
+        {
+            "description": "0",
+            "canonical_bson": "10000000126100000000000000000000",
+            "canonical_extjson": "{\"a\" : {\"$numberLong\" : \"0\"}}",
+            "relaxed_extjson": "{\"a\" : 0}"
+        },
+        {
+            "description": "1",
+            "canonical_bson": "10000000126100010000000000000000",
+            "canonical_extjson": "{\"a\" : {\"$numberLong\" : \"1\"}}",
+            "relaxed_extjson": "{\"a\" : 1}"
+        }
+    ],
+    "decodeErrors": [
+        {
+            "description": "int64 field truncated",
+            "bson": "0C0000001261001234567800"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/maxkey.json

@@ -0,0 +1,12 @@
+{
+    "description": "Maxkey type",
+    "bson_type": "0x7F",
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "Maxkey",
+            "canonical_bson": "080000007F610000",
+            "canonical_extjson": "{\"a\" : {\"$maxKey\" : 1}}"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/minkey.json

@@ -0,0 +1,12 @@
+{
+    "description": "Minkey type",
+    "bson_type": "0xFF",
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "Minkey",
+            "canonical_bson": "08000000FF610000",
+            "canonical_extjson": "{\"a\" : {\"$minKey\" : 1}}"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/multi-type-deprecated.json

@@ -0,0 +1,15 @@
+{
+    "description": "Multiple types within the same document",
+    "bson_type": "0x00",
+    "deprecated": true,
+    "valid": [
+        {
+            "description": "All BSON types",
+            "canonical_bson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
+            "converted_bson": "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",
+            "canonical_extjson": "{\"_id\": {\"$oid\": \"57e193d7a9cc81b4027498b5\"}, \"Symbol\": {\"$symbol\": \"symbol\"}, \"String\": \"string\", \"Int32\": {\"$numberInt\": \"42\"}, \"Int64\": {\"$numberLong\": \"42\"}, \"Double\": {\"$numberDouble\": \"-1.0\"}, \"Binary\": { \"$binary\" : {\"base64\": \"o0w498Or7cijeBSpkquNtg==\", \"subType\": \"03\"}}, \"BinaryUserDefined\": { \"$binary\" : {\"base64\": \"AQIDBAU=\", \"subType\": \"80\"}}, \"Code\": {\"$code\": \"function() {}\"}, \"CodeWithScope\": {\"$code\": \"function() {}\", \"$scope\": {}}, \"Subdocument\": {\"foo\": \"bar\"}, \"Array\": [{\"$numberInt\": \"1\"}, {\"$numberInt\": \"2\"}, {\"$numberInt\": \"3\"}, {\"$numberInt\": \"4\"}, {\"$numberInt\": \"5\"}], \"Timestamp\": {\"$timestamp\": {\"t\": 42, \"i\": 1}}, \"Regex\": {\"$regularExpression\": {\"pattern\": \"pattern\", \"options\": \"\"}}, \"DatetimeEpoch\": {\"$date\": {\"$numberLong\": \"0\"}}, \"DatetimePositive\": {\"$date\": {\"$numberLong\": \"2147483647\"}}, \"DatetimeNegative\": {\"$date\": {\"$numberLong\": \"-2147483648\"}}, \"True\": true, \"False\": false, \"DBPointer\": {\"$dbPointer\": {\"$ref\": \"collection\", \"$id\": {\"$oid\": \"57e193d7a9cc81b4027498b1\"}}}, \"DBRef\": {\"$ref\": \"collection\", \"$id\": {\"$oid\": \"57fd71e96e32ab4225b723fb\"}, \"$db\": \"database\"}, \"Minkey\": {\"$minKey\": 1}, \"Maxkey\": {\"$maxKey\": 1}, \"Null\": null, \"Undefined\": {\"$undefined\": true}}",
+            "converted_extjson": "{\"_id\": {\"$oid\": \"57e193d7a9cc81b4027498b5\"}, \"Symbol\": \"symbol\", \"String\": \"string\", \"Int32\": {\"$numberInt\": \"42\"}, \"Int64\": {\"$numberLong\": \"42\"}, \"Double\": {\"$numberDouble\": \"-1.0\"}, \"Binary\": { \"$binary\" : {\"base64\": \"o0w498Or7cijeBSpkquNtg==\", \"subType\": \"03\"}}, \"BinaryUserDefined\": { \"$binary\" : {\"base64\": \"AQIDBAU=\", \"subType\": \"80\"}}, \"Code\": {\"$code\": \"function() {}\"}, \"CodeWithScope\": {\"$code\": \"function() {}\", \"$scope\": {}}, \"Subdocument\": {\"foo\": \"bar\"}, \"Array\": [{\"$numberInt\": \"1\"}, {\"$numberInt\": \"2\"}, {\"$numberInt\": \"3\"}, {\"$numberInt\": \"4\"}, {\"$numberInt\": \"5\"}], \"Timestamp\": {\"$timestamp\": {\"t\": 42, \"i\": 1}}, \"Regex\": {\"$regularExpression\": {\"pattern\": \"pattern\", \"options\": \"\"}}, \"DatetimeEpoch\": {\"$date\": {\"$numberLong\": \"0\"}}, \"DatetimePositive\": {\"$date\": {\"$numberLong\": \"2147483647\"}}, \"DatetimeNegative\": {\"$date\": {\"$numberLong\": \"-2147483648\"}}, \"True\": true, \"False\": false, \"DBPointer\": {\"$ref\": \"collection\", \"$id\": {\"$oid\": \"57e193d7a9cc81b4027498b1\"}}, \"DBRef\": {\"$ref\": \"collection\", \"$id\": {\"$oid\": \"57fd71e96e32ab4225b723fb\"}, \"$db\": \"database\"}, \"Minkey\": {\"$minKey\": 1}, \"Maxkey\": {\"$maxKey\": 1}, \"Null\": null, \"Undefined\": null}"
+        }
+    ]
+}
+

data/spec/spec_tests/data/corpus/multi-type.json

@@ -0,0 +1,11 @@
+{
+    "description": "Multiple types within the same document",
+    "bson_type": "0x00",
+    "valid": [
+        {
+            "description": "All BSON types",
+            "canonical_bson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
+            "canonical_extjson": "{\"_id\": {\"$oid\": \"57e193d7a9cc81b4027498b5\"}, \"String\": \"string\", \"Int32\": {\"$numberInt\": \"42\"}, \"Int64\": {\"$numberLong\": \"42\"}, \"Double\": {\"$numberDouble\": \"-1.0\"}, \"Binary\": { \"$binary\" : {\"base64\": \"o0w498Or7cijeBSpkquNtg==\", \"subType\": \"03\"}}, \"BinaryUserDefined\": { \"$binary\" : {\"base64\": \"AQIDBAU=\", \"subType\": \"80\"}}, \"Code\": {\"$code\": \"function() {}\"}, \"CodeWithScope\": {\"$code\": \"function() {}\", \"$scope\": {}}, \"Subdocument\": {\"foo\": \"bar\"}, \"Array\": [{\"$numberInt\": \"1\"}, {\"$numberInt\": \"2\"}, {\"$numberInt\": \"3\"}, {\"$numberInt\": \"4\"}, {\"$numberInt\": \"5\"}], \"Timestamp\": {\"$timestamp\": {\"t\": 42, \"i\": 1}}, \"Regex\": {\"$regularExpression\": {\"pattern\": \"pattern\", \"options\": \"\"}}, \"DatetimeEpoch\": {\"$date\": {\"$numberLong\": \"0\"}}, \"DatetimePositive\": {\"$date\": {\"$numberLong\": \"2147483647\"}}, \"DatetimeNegative\": {\"$date\": {\"$numberLong\": \"-2147483648\"}}, \"True\": true, \"False\": false, \"DBRef\": {\"$ref\": \"collection\", \"$id\": {\"$oid\": \"57fd71e96e32ab4225b723fb\"}, \"$db\": \"database\"}, \"Minkey\": {\"$minKey\": 1}, \"Maxkey\": {\"$maxKey\": 1}, \"Null\": null}"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/null.json

@@ -0,0 +1,12 @@
+{
+    "description": "Null type",
+    "bson_type": "0x0A",
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "Null",
+            "canonical_bson": "080000000A610000",
+            "canonical_extjson": "{\"a\" : null}"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/oid.json

@@ -0,0 +1,28 @@
+{
+    "description": "ObjectId",
+    "bson_type": "0x07",
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "All zeroes",
+            "canonical_bson": "1400000007610000000000000000000000000000",
+            "canonical_extjson": "{\"a\" : {\"$oid\" : \"000000000000000000000000\"}}"
+        },
+        {
+            "description": "All ones",
+            "canonical_bson": "14000000076100FFFFFFFFFFFFFFFFFFFFFFFF00",
+            "canonical_extjson": "{\"a\" : {\"$oid\" : \"ffffffffffffffffffffffff\"}}"
+        },
+        {
+            "description": "Random",
+            "canonical_bson": "1400000007610056E1FC72E0C917E9C471416100",
+            "canonical_extjson": "{\"a\" : {\"$oid\" : \"56e1fc72e0c917e9c4714161\"}}"
+        }
+    ],
+    "decodeErrors": [
+        {
+            "description": "OID truncated",
+            "bson": "1200000007610056E1FC72E0C917E9C471"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/regex.json

@@ -0,0 +1,65 @@
+{
+    "description": "Regular Expression type",
+    "bson_type": "0x0B",
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "empty regex with no options",
+            "canonical_bson": "0A0000000B6100000000",
+            "canonical_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"\", \"options\" : \"\"}}}"
+        },
+        {
+            "description": "regex without options",
+            "canonical_bson": "0D0000000B6100616263000000",
+            "canonical_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"abc\", \"options\" : \"\"}}}"
+        },
+        {
+            "description": "regex with options",
+            "canonical_bson": "0F0000000B610061626300696D0000",
+            "canonical_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"abc\", \"options\" : \"im\"}}}"
+        },
+        {
+            "description": "regex with options (keys reversed)",
+            "canonical_bson": "0F0000000B610061626300696D0000",
+            "canonical_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"abc\", \"options\" : \"im\"}}}",
+            "degenerate_extjson": "{\"a\" : {\"$regularExpression\" : {\"options\" : \"im\", \"pattern\": \"abc\"}}}"
+        },
+        {
+            "description": "regex with slash",
+            "canonical_bson": "110000000B610061622F636400696D0000",
+            "canonical_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"ab/cd\", \"options\" : \"im\"}}}"
+        },
+        {
+            "description": "flags not alphabetized",
+            "degenerate_bson": "100000000B6100616263006D69780000",
+            "canonical_bson": "100000000B610061626300696D780000",
+            "canonical_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"abc\", \"options\" : \"imx\"}}}",
+            "degenerate_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"abc\", \"options\" : \"mix\"}}}"
+        },
+        {
+            "description" : "Required escapes",
+            "canonical_bson" : "100000000B610061625C226162000000",
+            "canonical_extjson": "{\"a\" : {\"$regularExpression\" : { \"pattern\": \"ab\\\\\\\"ab\", \"options\" : \"\"}}}"
+        },
+        {
+            "description" : "Regular expression as value of $regex query operator",
+            "canonical_bson" : "180000000B247265676578007061747465726E0069780000",
+            "canonical_extjson": "{\"$regex\" : {\"$regularExpression\" : { \"pattern\": \"pattern\", \"options\" : \"ix\"}}}"
+        },
+        {
+            "description" : "Regular expression as value of $regex query operator with $options",
+            "canonical_bson" : "270000000B247265676578007061747465726E000002246F7074696F6E73000300000069780000",
+            "canonical_extjson": "{\"$regex\" : {\"$regularExpression\" : { \"pattern\": \"pattern\", \"options\" : \"\"}}, \"$options\" : \"ix\"}"
+        }
+    ],
+    "decodeErrors": [
+        {
+            "description": "embedded null in pattern",
+            "bson": "0F0000000B610061006300696D0000"
+        },
+        {
+            "description": "embedded null in flags",
+            "bson": "100000000B61006162630069006D0000"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/string.json

@@ -0,0 +1,72 @@
+{
+    "description": "String",
+    "bson_type": "0x02",
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "Empty string",
+            "canonical_bson": "0D000000026100010000000000",
+            "canonical_extjson": "{\"a\" : \"\"}"
+        },
+        {
+            "description": "Single character",
+            "canonical_bson": "0E00000002610002000000620000",
+            "canonical_extjson": "{\"a\" : \"b\"}"
+        },
+        {
+            "description": "Multi-character",
+            "canonical_bson": "190000000261000D0000006162616261626162616261620000",
+            "canonical_extjson": "{\"a\" : \"abababababab\"}"
+        },
+        {
+            "description": "two-byte UTF-8 (\u00e9)",
+            "canonical_bson": "190000000261000D000000C3A9C3A9C3A9C3A9C3A9C3A90000",
+            "canonical_extjson": "{\"a\" : \"\\u00e9\\u00e9\\u00e9\\u00e9\\u00e9\\u00e9\"}"
+        },
+        {
+            "description": "three-byte UTF-8 (\u2606)",
+            "canonical_bson": "190000000261000D000000E29886E29886E29886E298860000",
+            "canonical_extjson": "{\"a\" : \"\\u2606\\u2606\\u2606\\u2606\"}"
+        },
+        {
+            "description": "Embedded nulls",
+            "canonical_bson": "190000000261000D0000006162006261620062616261620000",
+            "canonical_extjson": "{\"a\" : \"ab\\u0000bab\\u0000babab\"}"
+        },
+        {
+            "description": "Required escapes",
+            "canonical_bson" : "320000000261002600000061625C220102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F61620000",
+            "canonical_extjson" : "{\"a\":\"ab\\\\\\\"\\u0001\\u0002\\u0003\\u0004\\u0005\\u0006\\u0007\\b\\t\\n\\u000b\\f\\r\\u000e\\u000f\\u0010\\u0011\\u0012\\u0013\\u0014\\u0015\\u0016\\u0017\\u0018\\u0019\\u001a\\u001b\\u001c\\u001d\\u001e\\u001fab\"}"
+        }
+    ],
+    "decodeErrors": [
+        {
+            "description": "bad string length: 0 (but no 0x00 either)",
+            "bson": "0C0000000261000000000000"
+        },
+        {
+            "description": "bad string length: -1",
+            "bson": "0C000000026100FFFFFFFF00"
+        },
+        {
+            "description": "bad string length: eats terminator",
+            "bson": "10000000026100050000006200620000"
+        },
+        {
+            "description": "bad string length: longer than rest of document",
+            "bson": "120000000200FFFFFF00666F6F6261720000"
+        },
+        {
+            "description": "string is not null-terminated",
+            "bson": "1000000002610004000000616263FF00"
+        },
+        {
+            "description": "empty string, but extra null",
+            "bson": "0E00000002610001000000000000"
+        },
+        {
+            "description": "invalid UTF-8",
+            "bson": "0E00000002610002000000E90000"
+        }
+    ]
+}

data/spec/spec_tests/data/corpus/symbol.json

@@ -0,0 +1,80 @@
+{
+    "description": "Symbol",
+    "bson_type": "0x0E",
+    "deprecated": true,
+    "test_key": "a",
+    "valid": [
+        {
+            "description": "Empty string",
+            "canonical_bson": "0D0000000E6100010000000000",
+            "canonical_extjson": "{\"a\": {\"$symbol\": \"\"}}",
+            "converted_bson": "0D000000026100010000000000",
+            "converted_extjson": "{\"a\": \"\"}"
+        },
+        {
+            "description": "Single character",
+            "canonical_bson": "0E0000000E610002000000620000",
+            "canonical_extjson": "{\"a\": {\"$symbol\": \"b\"}}",
+            "converted_bson": "0E00000002610002000000620000",
+            "converted_extjson": "{\"a\": \"b\"}"
+        },
+        {
+            "description": "Multi-character",
+            "canonical_bson": "190000000E61000D0000006162616261626162616261620000",
+            "canonical_extjson": "{\"a\": {\"$symbol\": \"abababababab\"}}",
+            "converted_bson": "190000000261000D0000006162616261626162616261620000",
+            "converted_extjson": "{\"a\": \"abababababab\"}"
+        },
+        {
+            "description": "two-byte UTF-8 (\u00e9)",
+            "canonical_bson": "190000000E61000D000000C3A9C3A9C3A9C3A9C3A9C3A90000",
+            "canonical_extjson": "{\"a\": {\"$symbol\": \"éééééé\"}}",
+            "converted_bson": "190000000261000D000000C3A9C3A9C3A9C3A9C3A9C3A90000",
+            "converted_extjson": "{\"a\": \"éééééé\"}"
+        },
+        {
+            "description": "three-byte UTF-8 (\u2606)",
+            "canonical_bson": "190000000E61000D000000E29886E29886E29886E298860000",
+            "canonical_extjson": "{\"a\": {\"$symbol\": \"☆☆☆☆\"}}",
+            "converted_bson": "190000000261000D000000E29886E29886E29886E298860000",
+            "converted_extjson": "{\"a\": \"☆☆☆☆\"}"
+        },
+        {
+            "description": "Embedded nulls",
+            "canonical_bson": "190000000E61000D0000006162006261620062616261620000",
+            "canonical_extjson": "{\"a\": {\"$symbol\": \"ab\\u0000bab\\u0000babab\"}}",
+            "converted_bson": "190000000261000D0000006162006261620062616261620000",
+            "converted_extjson": "{\"a\": \"ab\\u0000bab\\u0000babab\"}"
+        }
+    ],
+    "decodeErrors": [
+        {
+            "description": "bad symbol length: 0 (but no 0x00 either)",
+            "bson": "0C0000000261000000000000"
+        },
+        {
+            "description": "bad symbol length: -1",
+            "bson": "0C000000026100FFFFFFFF00"
+        },
+        {
+            "description": "bad symbol length: eats terminator",
+            "bson": "10000000026100050000006200620000"
+        },
+        {
+            "description": "bad symbol length: longer than rest of document",
+            "bson": "120000000200FFFFFF00666F6F6261720000"
+        },
+        {
+            "description": "symbol is not null-terminated",
+            "bson": "1000000002610004000000616263FF00"
+        },
+        {
+            "description": "empty symbol, but extra null",
+            "bson": "0E00000002610001000000000000"
+        },
+        {
+            "description": "invalid UTF-8",
+            "bson": "0E00000002610002000000E90000"
+        }
+    ]
+}