Don't require nul-terminated string inputs

Avoid giving '\0' characters special treatment in JSON parsing code. This way the parser will reject input strings like "[1,2,3]\0-extra-nonsense" instead of returning a partially parsed interpretation of the input. It also allows the parser to be called with char pointers that might not be nul terminated (like the char pointers returned by std::string::data()). Fixes following test failures in https://github.com/nst/JSONTestSuite: bitcoin SHOULD_HAVE_FAILED n_multidigit_number_then_00.json
maflcko · Nov 8, 2016 · fd32d1a · fd32d1a
1 parent 3f03bfd
commit fd32d1a
Show file tree

Hide file tree

Showing 8 changed files with 48 additions and 23 deletions.
diff --git a/Makefile.am b/Makefile.am
@@ -20,7 +20,7 @@ libunivalue_la_LDFLAGS = \
 	-no-undefined
 libunivalue_la_CXXFLAGS = -I$(top_srcdir)/include
 
-TESTS = test/unitester
+TESTS = test/unitester test/no_nul
 
 GENBIN = gen/gen$(BUILD_EXEEXT)
 GEN_SRCS = gen/gen.cpp
@@ -42,6 +42,11 @@ test_unitester_LDADD = libunivalue.la
 test_unitester_CXXFLAGS = -I$(top_srcdir)/include -DJSON_TEST_SRC=\"$(srcdir)/$(TEST_DATA_DIR)\"
 test_unitester_LDFLAGS = -static $(LIBTOOL_APP_LDFLAGS)
 
+test_no_nul_SOURCES = test/no_nul.cpp
+test_no_nul_LDADD = libunivalue.la
+test_no_nul_CXXFLAGS = -I$(top_srcdir)/include
+test_no_nul_LDFLAGS = -static $(LIBTOOL_APP_LDFLAGS)
+
 TEST_FILES = \
 	$(TEST_DATA_DIR)/fail10.json \
 	$(TEST_DATA_DIR)/fail11.json \
@@ -77,6 +82,7 @@ TEST_FILES = \
 	$(TEST_DATA_DIR)/fail39.json \
 	$(TEST_DATA_DIR)/fail40.json \
 	$(TEST_DATA_DIR)/fail41.json \
+	$(TEST_DATA_DIR)/fail42.json \
 	$(TEST_DATA_DIR)/fail3.json \
 	$(TEST_DATA_DIR)/fail4.json \
 	$(TEST_DATA_DIR)/fail5.json \

diff --git a/include/univalue.h b/include/univalue.h
@@ -124,9 +124,10 @@ class UniValue {
     std::string write(unsigned int prettyIndent = 0,
                       unsigned int indentLevel = 0) const;
 
+    bool read(const char *raw, size_t len);
     bool read(const char *raw);
     bool read(const std::string& rawStr) {
-        return read(rawStr.c_str());
+        return read(rawStr.data(), rawStr.size());
     }
 
 private:
@@ -240,7 +241,7 @@ enum jtokentype {
 };
 
 extern enum jtokentype getJsonToken(std::string& tokenVal,
-                                    unsigned int& consumed, const char *raw);
+                                    unsigned int& consumed, const char *raw, const char *end);
 extern const char *uvTypeName(UniValue::VType t);
 
 static inline bool jsonTokenIsValue(enum jtokentype jtt)

diff --git a/lib/univalue.cpp b/lib/univalue.cpp
@@ -104,7 +104,7 @@ static bool validNumStr(const string& s)
 {
     string tokenVal;
     unsigned int consumed;
-    enum jtokentype tt = getJsonToken(tokenVal, consumed, s.c_str());
+    enum jtokentype tt = getJsonToken(tokenVal, consumed, s.data(), s.data() + s.size());
     return (tt == JTOK_NUMBER);
 }
 

diff --git a/lib/univalue_read.cpp b/lib/univalue_read.cpp
@@ -43,21 +43,21 @@ static const char *hatoui(const char *first, const char *last,
 }
 
 enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
-                            const char *raw)
+                            const char *raw, const char *end)
 {
     tokenVal.clear();
     consumed = 0;
 
     const char *rawStart = raw;
 
-    while ((*raw) && (json_isspace(*raw)))             // skip whitespace
+    while (raw < end && (json_isspace(*raw)))          // skip whitespace
         raw++;
 
-    switch (*raw) {
-
-    case 0:
+    if (raw >= end)
         return JTOK_NONE;
 
+    switch (*raw) {
+
     case '{':
         raw++;
         consumed = (raw - rawStart);
@@ -127,40 +127,40 @@ enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
         numStr += *raw;                       // copy first char
         raw++;
 
-        if ((*first == '-') && (!json_isdigit(*raw)))
+        if ((*first == '-') && (raw < end) && (!json_isdigit(*raw)))
             return JTOK_ERR;
 
-        while ((*raw) && json_isdigit(*raw)) {     // copy digits
+        while (raw < end && json_isdigit(*raw)) {  // copy digits
             numStr += *raw;
             raw++;
         }
 
         // part 2: frac
-        if (*raw == '.') {
+        if (raw < end && *raw == '.') {
             numStr += *raw;                   // copy .
             raw++;
 
-            if (!json_isdigit(*raw))
+            if (raw >= end || !json_isdigit(*raw))
                 return JTOK_ERR;
-            while ((*raw) && json_isdigit(*raw)) { // copy digits
+            while (raw < end && json_isdigit(*raw)) { // copy digits
                 numStr += *raw;
                 raw++;
             }
         }
 
         // part 3: exp
-        if (*raw == 'e' || *raw == 'E') {
+        if (raw < end && (*raw == 'e' || *raw == 'E')) {
             numStr += *raw;                   // copy E
             raw++;
 
-            if (*raw == '-' || *raw == '+') { // copy +/-
+            if (raw < end && (*raw == '-' || *raw == '+')) { // copy +/-
                 numStr += *raw;
                 raw++;
             }
 
-            if (!json_isdigit(*raw))
+            if (raw >= end || !json_isdigit(*raw))
                 return JTOK_ERR;
-            while ((*raw) && json_isdigit(*raw)) { // copy digits
+            while (raw < end && json_isdigit(*raw)) { // copy digits
                 numStr += *raw;
                 raw++;
             }
@@ -177,13 +177,16 @@ enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
         string valStr;
         JSONUTF8StringFilter writer(valStr);
 
-        while (*raw) {
+        while (raw < end) {
             if ((unsigned char)*raw < 0x20)
                 return JTOK_ERR;
 
             else if (*raw == '\\') {
                 raw++;                        // skip backslash
 
+                if (raw >= end)
+                    return JTOK_ERR;
+
                 switch (*raw) {
                 case '"':  writer.push_back('\"'); break;
                 case '\\': writer.push_back('\\'); break;
@@ -196,7 +199,8 @@ enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
 
                 case 'u': {
                     unsigned int codepoint;
-                    if (hatoui(raw + 1, raw + 1 + 4, codepoint) !=
+                    if (raw + 1 + 4 >= end ||
+                        hatoui(raw + 1, raw + 1 + 4, codepoint) !=
                                raw + 1 + 4)
                         return JTOK_ERR;
                     writer.push_back_u(codepoint);
@@ -246,7 +250,7 @@ enum expect_bits {
 #define setExpect(bit) (expectMask |= EXP_##bit)
 #define clearExpect(bit) (expectMask &= ~EXP_##bit)
 
-bool UniValue::read(const char *raw)
+bool UniValue::read(const char *raw, size_t size)
 {
     clear();
 
@@ -257,10 +261,11 @@ bool UniValue::read(const char *raw)
     unsigned int consumed;
     enum jtokentype tok = JTOK_NONE;
     enum jtokentype last_tok = JTOK_NONE;
+    const char* end = raw + size;
     do {
         last_tok = tok;
 
-        tok = getJsonToken(tokenVal, consumed, raw);
+        tok = getJsonToken(tokenVal, consumed, raw, end);
         if (tok == JTOK_NONE || tok == JTOK_ERR)
             return false;
         raw += consumed;
@@ -432,10 +437,13 @@ bool UniValue::read(const char *raw)
     } while (!stack.empty ());
 
     /* Check that nothing follows the initial construct (parsed above).  */
-    tok = getJsonToken(tokenVal, consumed, raw);
+    tok = getJsonToken(tokenVal, consumed, raw, end);
     if (tok != JTOK_NONE)
         return false;
 
     return true;
 }
 
+bool UniValue::read(const char *raw) {
+    return read(raw, strlen(raw));
+}
diff --git a/test/.gitignore b/test/.gitignore
@@ -1,4 +1,5 @@
 unitester
+no_nul
 
 *.trs
 *.log
diff --git a/test/fail42.json b/test/fail42.json
diff --git a/test/no_nul.cpp b/test/no_nul.cpp
@@ -0,0 +1,8 @@
+#include "univalue.h"
+
+int main (int argc, char *argv[])
+{
+    char buf[] = "___[1,2,3]___";
+    UniValue val;
+    return val.read(buf + 3, 7) ? 0 : 1;
+}
diff --git a/test/unitester.cpp b/test/unitester.cpp
@@ -113,6 +113,7 @@ static const char *filenames[] = {
         "fail39.json",               // invalid unicode: only second half of surrogate pair
         "fail40.json",               // invalid unicode: broken UTF-8
         "fail41.json",               // invalid unicode: unfinished UTF-8
+        "fail42.json",               // valid json with garbage following a nul byte
         "fail3.json",
         "fail4.json",                // extra comma
         "fail5.json",