From c1ccada2bc0f3eff3687da81876b65d4b7655a68 Mon Sep 17 00:00:00 2001
From: Qingyang Hu <qingyang.hu@mongodb.com>
Date: Mon, 31 Jul 2023 16:30:30 -0400
Subject: [PATCH 1/2] GODRIVER-2106 Sync code and symbol BSON corpus tests.

---
 testdata/bson-corpus/code.json   | 26 +++++++++++++-------------
 testdata/bson-corpus/symbol.json | 14 +++++++-------
 2 files changed, 20 insertions(+), 20 deletions(-)

diff --git a/testdata/bson-corpus/code.json b/testdata/bson-corpus/code.json
index 6f37349ad0..b8482b2541 100644
--- a/testdata/bson-corpus/code.json
+++ b/testdata/bson-corpus/code.json
@@ -20,48 +20,48 @@
         },
         {
             "description": "two-byte UTF-8 (\u00e9)",
-            "canonical_bson": "190000000261000D000000C3A9C3A9C3A9C3A9C3A9C3A90000",
-            "canonical_extjson": "{\"a\" : \"\\u00e9\\u00e9\\u00e9\\u00e9\\u00e9\\u00e9\"}"
+            "canonical_bson": "190000000D61000D000000C3A9C3A9C3A9C3A9C3A9C3A90000",
+            "canonical_extjson": "{\"a\" : {\"$code\" : \"\\u00e9\\u00e9\\u00e9\\u00e9\\u00e9\\u00e9\"}}"
         },
         {
             "description": "three-byte UTF-8 (\u2606)",
-            "canonical_bson": "190000000261000D000000E29886E29886E29886E298860000",
-            "canonical_extjson": "{\"a\" : \"\\u2606\\u2606\\u2606\\u2606\"}"
+            "canonical_bson": "190000000D61000D000000E29886E29886E29886E298860000",
+            "canonical_extjson": "{\"a\" : {\"$code\" : \"\\u2606\\u2606\\u2606\\u2606\"}}"
         },
         {
             "description": "Embedded nulls",
-            "canonical_bson": "190000000261000D0000006162006261620062616261620000",
-            "canonical_extjson": "{\"a\" : \"ab\\u0000bab\\u0000babab\"}"
+            "canonical_bson": "190000000D61000D0000006162006261620062616261620000",
+            "canonical_extjson": "{\"a\" : {\"$code\" : \"ab\\u0000bab\\u0000babab\"}}"
         }
     ],
     "decodeErrors": [
         {
             "description": "bad code string length: 0 (but no 0x00 either)",
-            "bson": "0C0000000261000000000000"
+            "bson": "0C0000000D61000000000000"
         },
         {
             "description": "bad code string length: -1",
-            "bson": "0C000000026100FFFFFFFF00"
+            "bson": "0C0000000D6100FFFFFFFF00"
         },
         {
             "description": "bad code string length: eats terminator",
-            "bson": "10000000026100050000006200620000"
+            "bson": "100000000D6100050000006200620000"
         },
         {
             "description": "bad code string length: longer than rest of document",
-            "bson": "120000000200FFFFFF00666F6F6261720000"
+            "bson": "120000000D00FFFFFF00666F6F6261720000"
         },
         {
             "description": "code string is not null-terminated",
-            "bson": "1000000002610004000000616263FF00"
+            "bson": "100000000D610004000000616263FF00"
         },
         {
             "description": "empty code string, but extra null",
-            "bson": "0E00000002610001000000000000"
+            "bson": "0E0000000D610001000000000000"
         },
         {
             "description": "invalid UTF-8",
-            "bson": "0E00000002610002000000E90000"
+            "bson": "0E0000000D610002000000E90000"
         }
     ]
 }
diff --git a/testdata/bson-corpus/symbol.json b/testdata/bson-corpus/symbol.json
index 4e46cb9511..3dd3577ebd 100644
--- a/testdata/bson-corpus/symbol.json
+++ b/testdata/bson-corpus/symbol.json
@@ -50,31 +50,31 @@
     "decodeErrors": [
         {
             "description": "bad symbol length: 0 (but no 0x00 either)",
-            "bson": "0C0000000261000000000000"
+            "bson": "0C0000000E61000000000000"
         },
         {
             "description": "bad symbol length: -1",
-            "bson": "0C000000026100FFFFFFFF00"
+            "bson": "0C0000000E6100FFFFFFFF00"
         },
         {
             "description": "bad symbol length: eats terminator",
-            "bson": "10000000026100050000006200620000"
+            "bson": "100000000E6100050000006200620000"
         },
         {
             "description": "bad symbol length: longer than rest of document",
-            "bson": "120000000200FFFFFF00666F6F6261720000"
+            "bson": "120000000E00FFFFFF00666F6F6261720000"
         },
         {
             "description": "symbol is not null-terminated",
-            "bson": "1000000002610004000000616263FF00"
+            "bson": "100000000E610004000000616263FF00"
         },
         {
             "description": "empty symbol, but extra null",
-            "bson": "0E00000002610001000000000000"
+            "bson": "0E0000000E610001000000000000"
         },
         {
             "description": "invalid UTF-8",
-            "bson": "0E00000002610002000000E90000"
+            "bson": "0E0000000E610002000000E90000"
         }
     ]
 }

From 8ff8309ff2be29c9177c4febb2766c0cd69e4ee0 Mon Sep 17 00:00:00 2001
From: Qingyang Hu <qingyang.hu@mongodb.com>
Date: Fri, 11 Aug 2023 18:53:00 -0400
Subject: [PATCH 2/2] update test assertion

---
 bson/bson_corpus_spec_test.go | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/bson/bson_corpus_spec_test.go b/bson/bson_corpus_spec_test.go
index 726d015ec8..8a27b06272 100644
--- a/bson/bson_corpus_spec_test.go
+++ b/bson/bson_corpus_spec_test.go
@@ -14,6 +14,7 @@ import (
 	"math"
 	"os"
 	"path"
+	"reflect"
 	"strconv"
 	"strings"
 	"testing"
@@ -407,8 +408,8 @@ func runTest(t *testing.T, file string) {
 					// and assert that there was no error if any of the string or DBPointer values contain invalid UTF-8
 					// characters.
 					for _, elem := range doc {
-						str, ok := elem.Value.(string)
-						invalidString := ok && !utf8.ValidString(str)
+						value := reflect.ValueOf(elem.Value)
+						invalidString := (value.Kind() == reflect.String) && !utf8.ValidString(value.String())
 						dbPtr, ok := elem.Value.(primitive.DBPointer)
 						invalidDBPtr := ok && !utf8.ValidString(dbPtr.DB)