From: drh <>
Date: Wed, 31 Jan 2024 13:46:44 +0000 (+0000)
Subject: Allow control characters in JSON5 string literals.
X-Git-Tag: version-3.46.0~245^2~2
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=c24f53635c3b17f732119a224255c61edc90a089;p=thirdparty%2Fsqlite.git

Allow control characters in JSON5 string literals.
[forum:/forumpost/05182119f69c3a92|Forum thread 05182119f69c3a92].

FossilOrigin-Name: 34709c7cc910539e23a830ad8b589a97a88be25e924a59670c1017fb51447dad
---

diff --git a/manifest b/manifest
index 2604713c72..e88859a0db 100644
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Fix\sa\smemory\sallocation\sbug\sin\sthe\s(debug-use-only)\sjson_parse()\sSQL\sfunction.
-D 2024-01-31T12:15:57.302
+C Allow\scontrol\scharacters\sin\sJSON5\sstring\sliterals.\n[forum:/forumpost/05182119f69c3a92|Forum\sthread\s05182119f69c3a92].
+D 2024-01-31T13:46:44.886
 F .fossil-settings/empty-dirs dbb81e8fc0401ac46a1491ab34a7f2c7c0452f2f06b54ebb845d024ca8283ef1
 F .fossil-settings/ignore-glob 35175cdfcf539b2318cb04a9901442804be81cd677d8b889fcc9149c21f239ea
 F LICENSE.md df5091916dbb40e6e9686186587125e1b2ff51f022cc334e886c19a0e9982724
@@ -698,7 +698,7 @@ F src/hash.h 3340ab6e1d13e725571d7cee6d3e3135f0779a7d8e76a9ce0a85971fa3953c51
 F src/hwtime.h f9c2dfb84dce7acf95ce6d289e46f5f9d3d1afd328e53da8f8e9008e3b3caae6
 F src/in-operator.md 10cd8f4bcd225a32518407c2fb2484089112fd71
 F src/insert.c 3f0a94082d978bbdd33c38fefea15346c6c6bffb70bc645a71dc0f1f87dd3276
-F src/json.c 192c5e5013607b14441a738f0bd73e9df057b13eeac4060b679b32021a25e018
+F src/json.c 366c4502167672d9ded0d45b1a955acb487ddabb531e889a7b8901fd3e46d013
 F src/legacy.c d7874bc885906868cd51e6c2156698f2754f02d9eee1bae2d687323c3ca8e5aa
 F src/loadext.c 7432c944ff197046d67a1207790a1b13eec4548c85a9457eb0896bb3641dfb36
 F src/main.c 438b95162acfa17b7d218f586f5bde11d6ae82bcf030c9611fc537556870ad6b
@@ -1343,7 +1343,7 @@ F test/json104.test 1b844a70cddcfa2e4cd81a5db0657b2e61e7f00868310f24f56a9ba01143
 F test/json105.test 043838b56e68f3252a0dcf5be1689016f6f3f05056f8dcfcdc9d074f4d932988
 F test/json106.test 1d46a9294e2ced35c7f87cebbcb9626d01abab04f1969d7ded7b6f6a1d9be0f2
 F test/json107.test 59054e815c8f6b67d634d44ace421cf975828fb5651c4460aa66015c8e19d562
-F test/json501.test ab168a12eb6eb14d479f8c1cdae3ac062fd5a4679f17f976e96f1af518408330
+F test/json501.test b95e2d14988b682a5cadf079dd6162f0f85fb74cd59c6b1f1624110104a974eb
 F test/json502.test 84634d3dbb521d2814e43624025b760c6198456c8197bbec6c977c0236648f5b
 F test/jsonb01.test f4cdfb4cf5a0c940091b17675ed9583f45add0c938f07d65b0de0e19d3a9a101
 F test/keyword1.test 37ef6bba5d2ed5b07ecdd6810571de2956599dff
@@ -2161,8 +2161,8 @@ F vsixtest/vsixtest.tcl 6a9a6ab600c25a91a7acc6293828957a386a8a93
 F vsixtest/vsixtest.vcxproj.data 2ed517e100c66dc455b492e1a33350c1b20fbcdc
 F vsixtest/vsixtest.vcxproj.filters 37e51ffedcdb064aad6ff33b6148725226cd608e
 F vsixtest/vsixtest_TemporaryKey.pfx e5b1b036facdb453873e7084e1cae9102ccc67a0
-P 72f3e1d0eca2b9ebd7c4b4cdccc6d7a0173b2fdb204cb1aa207917aedab6098a
-R 1df2c3fb322d8908152fe23b99fe1f47
+P 32ce7dacf58bbf35cf70da8b03fa25f97fcea970edbc1f8aa12739ac4c8608fb
+R c7867d5f34f471f1c203d9ecfab6ff98
 U drh
-Z d967a8cc96a8b8a2072c7fb0795b6f21
+Z a250ac8f403a1bdc996f04e6b4453df0
 # Remove this line to create a well-formed Fossil manifest.
diff --git a/manifest.uuid b/manifest.uuid
index 048895706f..aea1dec179 100644
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-32ce7dacf58bbf35cf70da8b03fa25f97fcea970edbc1f8aa12739ac4c8608fb
\ No newline at end of file
+34709c7cc910539e23a830ad8b589a97a88be25e924a59670c1017fb51447dad
\ No newline at end of file
diff --git a/src/json.c b/src/json.c
index 8f8196bd75..d585189902 100644
--- a/src/json.c
+++ b/src/json.c
@@ -621,6 +621,40 @@ static void jsonAppendSeparator(JsonString *p){
   jsonAppendChar(p, ',');
 }
 
+/* c is a control character.  Append the canonical JSON representation
+** of that control character to p.
+**
+** This routine assumes that the output buffer has already been enlarged
+** sufficiently to hold the worst-case encoding plus a nul terminator.
+*/
+static void jsonAppendControlChar(JsonString *p, u8 c){
+  static const char aSpecial[] = {
+     0, 0, 0, 0, 0, 0, 0, 0, 'b', 't', 'n', 0, 'f', 'r', 0, 0,
+     0, 0, 0, 0, 0, 0, 0, 0,   0,   0,   0, 0,   0,   0, 0, 0
+  };
+  assert( sizeof(aSpecial)==32 );
+  assert( aSpecial['\b']=='b' );
+  assert( aSpecial['\f']=='f' );
+  assert( aSpecial['\n']=='n' );
+  assert( aSpecial['\r']=='r' );
+  assert( aSpecial['\t']=='t' );
+  assert( c>=0 && c<sizeof(aSpecial) );
+  assert( p->nUsed+7 <= p->nAlloc );
+  if( aSpecial[c] ){
+    p->zBuf[p->nUsed] = '\\';
+    p->zBuf[p->nUsed+1] = aSpecial[c];
+    p->nUsed += 2;
+  }else{
+    p->zBuf[p->nUsed] = '\\';
+    p->zBuf[p->nUsed+1] = 'u';
+    p->zBuf[p->nUsed+2] = '0';
+    p->zBuf[p->nUsed+3] = '0';
+    p->zBuf[p->nUsed+4] = "0123456789abcdef"[c>>4];
+    p->zBuf[p->nUsed+5] = "0123456789abcdef"[c&0xf];
+    p->nUsed += 6;
+  }
+}
+
 /* Append the N-byte string in zIn to the end of the JsonString string
 ** under construction.  Enclose the string in double-quotes ("...") and
 ** escape any double-quotes or backslash characters contained within the
@@ -680,35 +714,14 @@ static void jsonAppendString(JsonString *p, const char *zIn, u32 N){
     }
     c = z[0];
     if( c=='"' || c=='\\' ){
-      json_simple_escape:
       if( (p->nUsed+N+3 > p->nAlloc) && jsonStringGrow(p,N+3)!=0 ) return;
       p->zBuf[p->nUsed++] = '\\';
       p->zBuf[p->nUsed++] = c;
     }else if( c=='\'' ){
       p->zBuf[p->nUsed++] = c;
     }else{
-      static const char aSpecial[] = {
-         0, 0, 0, 0, 0, 0, 0, 0, 'b', 't', 'n', 0, 'f', 'r', 0, 0,
-         0, 0, 0, 0, 0, 0, 0, 0,   0,   0,   0, 0,   0,   0, 0, 0
-      };
-      assert( sizeof(aSpecial)==32 );
-      assert( aSpecial['\b']=='b' );
-      assert( aSpecial['\f']=='f' );
-      assert( aSpecial['\n']=='n' );
-      assert( aSpecial['\r']=='r' );
-      assert( aSpecial['\t']=='t' );
-      assert( c>=0 && c<sizeof(aSpecial) );
-      if( aSpecial[c] ){
-        c = aSpecial[c];
-        goto json_simple_escape;
-      }
       if( (p->nUsed+N+7 > p->nAlloc) && jsonStringGrow(p,N+7)!=0 ) return;
-      p->zBuf[p->nUsed++] = '\\';
-      p->zBuf[p->nUsed++] = 'u';
-      p->zBuf[p->nUsed++] = '0';
-      p->zBuf[p->nUsed++] = '0';
-      p->zBuf[p->nUsed++] = "0123456789abcdef"[c>>4];
-      p->zBuf[p->nUsed++] = "0123456789abcdef"[c&0xf];
+      jsonAppendControlChar(p, c);
     }
     z++;
     N--;
@@ -1409,6 +1422,9 @@ static u32 jsonbValidityCheck(
         if( !jsonIsOk[z[j]] && z[j]!='\'' ){
           if( z[j]=='"' ){
             if( x==JSONB_TEXTJ ) return j+1;
+          }else if( z[j]<=0x1f ){
+            /* Control characters in JSON5 string literals are ok */
+            if( x==JSONB_TEXTJ ) return j+1;
           }else if( z[j]!='\\' || j+1>=k ){
             return j+1;
           }else if( strchr("\"\\/bfnrt",z[j+1])!=0 ){
@@ -1703,9 +1719,10 @@ json_parse_restart:
           return -1;
         }
       }else if( c<=0x1f ){
-        /* Control characters are not allowed in strings */
-        pParse->iErr = j;
-        return -1;
+        /* Control characters are not allowed in canonical JSON string
+        ** literals, but are allowed in JSON5 string literals. */
+        opcode = JSONB_TEXT5;
+        pParse->hasNonstd = 1;
       }else if( c=='"' ){
         opcode = JSONB_TEXT5;
       }
@@ -2186,7 +2203,7 @@ static u32 jsonTranslateBlobToText(
       zIn = (const char*)&pParse->aBlob[i+n];
       jsonAppendChar(pOut, '"');
       while( sz2>0 ){
-        for(k=0; k<sz2 && zIn[k]!='\\' && zIn[k]!='"'; k++){}
+        for(k=0; k<sz2 && (jsonIsOk[(u8)zIn[k]] || zIn[k]=='\''); k++){}
         if( k>0 ){
           jsonAppendRawNZ(pOut, zIn, k);
           if( k>=sz2 ){
@@ -2201,6 +2218,13 @@ static u32 jsonTranslateBlobToText(
           sz2--;
           continue;
         }
+        if( zIn[0]<=0x1f ){
+          if( pOut->nUsed+7<pOut->nAlloc && jsonStringGrow(pOut,7) ) break;
+          jsonAppendControlChar(pOut, zIn[0]);
+          zIn++;
+          sz2--;
+          continue;
+        }
         assert( zIn[0]=='\\' );
         assert( sz2>=1 );
         if( sz2<2 ){
@@ -4391,7 +4415,7 @@ static void jsonErrorFunc(
         /* Convert byte-offset s.iErr into a character offset */
         u32 k;
         assert( s.zJson!=0 );  /* Because s.oom is false */
-        for(k=0; k<s.iErr && ALWAYS(s.zJson[k]); k++){
+        for(k=0; k<s.iErr && s.zJson[k]; k++){
           if( (s.zJson[k] & 0xc0)!=0x80 ) iErrPos++;
         }
         iErrPos++;
diff --git a/test/json501.test b/test/json501.test
index 40b3b563db..bfd21055a7 100644
--- a/test/json501.test
+++ b/test/json501.test
@@ -306,4 +306,31 @@ do_execsql_test 13.1 {
   SELECT json('{x:''a "b" c''}');
 } {{{"x":"a \"b\" c"}}}
 
+# 2024-01-31
+# Allow control characters within JSON5 string literals.
+#
+for {set c 1} {$c<=0x1f} {incr c} {
+  do_execsql_test 14.$c.1 {
+    SELECT json_valid('"abc' || char($c) || 'xyz"');
+  } {0}
+  do_execsql_test 14.$c.2 {
+    SELECT json_valid('"abc' || char($c) || 'xyz"', 2);
+  } {1}
+  switch $c {
+    8   {set e "\\b"}
+    9   {set e "\\t"}
+    10  {set e "\\n"}
+    12  {set e "\\f"}
+    13  {set e "\\r"}
+    default {set e [format "\\u00%02x" $c]}
+  }
+  do_execsql_test 14.$c.3 {
+    SELECT json('{label:"abc' || char($c) || 'xyz"}');
+  } "{{\"label\":\"abc${e}xyz\"}}"
+  do_execsql_test 14.$c.4 {
+    SELECT jsonb('{label:"abc' || char($c) || 'xyz"}') -> '$';
+  } "{{\"label\":\"abc${e}xyz\"}}"
+}
+
+
 finish_test