[idl_parser] Unify parsing of NaN values read from .fbs and .json files (#6296)

This commit unifies parsing of NaN values read
from .fbs and .json files by converting them to unsigned NaN.

7 files changed, 55 insertions, 8 deletions

diff --git a/include/flatbuffers/base.h b/include/flatbuffers/base.h
index 95b788e0..92db55b7 100644
--- a/include/flatbuffers/base.h
+++ b/include/flatbuffers/base.h
@@ -197,12 +197,17 @@ namespace flatbuffers {
 #if (!defined(_MSC_VER) || _MSC_FULL_VER >= 180020827) && \
     (!defined(__GNUC__) || (__GNUC__ * 100 + __GNUC_MINOR__ >= 404)) || \
     defined(__clang__)
-  #define FLATBUFFERS_DEFAULT_DECLARATION
   #define FLATBUFFERS_DELETE_FUNC(func) func = delete;
 #else
   #define FLATBUFFERS_DELETE_FUNC(func) private: func;
 #endif
 
+#if (!defined(_MSC_VER) || _MSC_VER >= 1900) && \
+    (!defined(__GNUC__) || (__GNUC__ * 100 + __GNUC_MINOR__ >= 409)) || \
+    defined(__clang__)
+  #define FLATBUFFERS_DEFAULT_DECLARATION
+#endif
+
 // Check if we can use template aliases
 // Not possible if Microsoft Compiler before 2012
 // Possible is the language feature __cpp_alias_templates is defined well
diff --git a/include/flatbuffers/stl_emulation.h b/include/flatbuffers/stl_emulation.h
index 66bd620d..70e5dc94 100644
--- a/include/flatbuffers/stl_emulation.h
+++ b/include/flatbuffers/stl_emulation.h
@@ -164,6 +164,8 @@ inline void vector_emplace_back(std::vector<T> *vector, V &&data) {
     using conditional = std::conditional<B, T, F>;
     template<class T, T v>
     using integral_constant = std::integral_constant<T, v>;
+    template <bool B>
+    using bool_constant = integral_constant<bool, B>;
   #else
     // Map C++ TR1 templates defined by stlport.
     template <typename T> using is_scalar = std::tr1::is_scalar<T>;
@@ -187,6 +189,8 @@ inline void vector_emplace_back(std::vector<T> *vector, V &&data) {
     using conditional = std::tr1::conditional<B, T, F>;
     template<class T, T v>
     using integral_constant = std::tr1::integral_constant<T, v>;
+    template <bool B>
+    using bool_constant = integral_constant<bool, B>;
   #endif  // !FLATBUFFERS_CPP98_STL
 #else
   // MSVC 2010 doesn't support C++11 aliases.
@@ -201,6 +205,8 @@ inline void vector_emplace_back(std::vector<T> *vector, V &&data) {
   struct conditional : public std::conditional<B, T, F> {};
   template<class T, T v>
   struct integral_constant : public std::integral_constant<T, v> {};
+  template <bool B>
+  struct bool_constant : public integral_constant<bool, B> {};
 #endif  // defined(FLATBUFFERS_TEMPLATES_ALIASES)
 
 #ifndef FLATBUFFERS_CPP98_STL
diff --git a/src/idl_parser.cpp b/src/idl_parser.cpp
index 6d916e5d..bb23dc11 100644
--- a/src/idl_parser.cpp
+++ b/src/idl_parser.cpp
@@ -172,8 +172,21 @@ template<typename T> std::string TypeToIntervalString() {
 
 // atot: template version of atoi/atof: convert a string to an instance of T.
 template<typename T>
-inline CheckedError atot(const char *s, Parser &parser, T *val) {
-  auto done = StringToNumber(s, val);
+bool atot_scalar(const char *s, T *val, bool_constant<false>) {
+  return StringToNumber(s, val);
+}
+
+template<typename T>
+bool atot_scalar(const char *s, T *val, bool_constant<true>) {
+  // Normalize NaN parsed from fbs or json to unsigned NaN.
+  if (false == StringToNumber(s, val)) return false;
+  *val = (*val != *val) ? std::fabs(*val) : *val;
+  return true;
+}
+
+template<typename T>
+CheckedError atot(const char *s, Parser &parser, T *val) {
+  auto done = atot_scalar(s, val, bool_constant<is_floating_point<T>::value>());
   if (done) return NoError();
   if (0 == *val)
     return parser.Error("invalid number: \"" + std::string(s) + "\"");
@@ -1806,7 +1819,6 @@ CheckedError Parser::ParseSingleValue(const std::string *name, Value &e,
       match = true;
     }
     // Parse a float/integer number from the string.
-    if (!match) check_now = true;  // Re-pack if parsed from string literal.
     // A "scalar-in-string" value needs extra checks.
     if (!match && is_tok_string && IsScalar(in_type)) {
       // Strip trailing whitespaces from attribute_.
diff --git a/tests/fuzzer/CMakeLists.txt b/tests/fuzzer/CMakeLists.txt
index 5deba0fd..717bc943 100644
--- a/tests/fuzzer/CMakeLists.txt
+++ b/tests/fuzzer/CMakeLists.txt
@@ -139,6 +139,23 @@ target_link_libraries(monster_fuzzer PRIVATE flatbuffers_fuzzed)
 # Build debugger for weird cases found with fuzzer.
 if(BUILD_DEBUGGER)
   add_library(flatbuffers_nonfuzz STATIC ${FlatBuffers_Library_SRCS})
+  target_compile_options(
+    flatbuffers_nonfuzz
+    PUBLIC
+      $<$<BOOL:${USE_ASAN}>:
+        -fsanitize=undefined,address
+      >
+      -fno-limit-debug-info
+  )
+  
+  target_link_libraries(
+    flatbuffers_nonfuzz
+    PUBLIC
+      $<$<BOOL:${USE_ASAN}>:
+        -fsanitize=undefined,address
+      >
+  )
+
   target_compile_definitions(
     flatbuffers_nonfuzz
     PUBLIC
@@ -147,6 +164,9 @@ if(BUILD_DEBUGGER)
     PRIVATE
       FLATBUFFERS_MAX_PARSING_DEPTH=${FLATBUFFERS_MAX_PARSING_DEPTH}
   )
-  add_executable(scalar_debug flatbuffers_scalar_fuzzer.cc scalar_debug.cpp)
+  add_executable(scalar_debug
+    flatbuffers_scalar_fuzzer.cc
+    scalar_debug.cpp
+  )
   target_link_libraries(scalar_debug PRIVATE flatbuffers_nonfuzz)
 endif(BUILD_DEBUGGER)
diff --git a/tests/fuzzer/flatbuffers_scalar_fuzzer.cc b/tests/fuzzer/flatbuffers_scalar_fuzzer.cc
index 3c711fcb..16728d29 100644
--- a/tests/fuzzer/flatbuffers_scalar_fuzzer.cc
+++ b/tests/fuzzer/flatbuffers_scalar_fuzzer.cc
@@ -288,7 +288,7 @@ extern "C" int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) {
     }
 
     // Parse original input as-is.
-    auto orig_scalar = "{ \"Y\" : " + input + " }";
+    auto orig_scalar = "{\"Y\" : " + input + "}";
     std::string orig_back;
     auto orig_done = Parse(parser, orig_scalar, &orig_back);
 
@@ -326,7 +326,7 @@ extern "C" int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) {
 
     // Test quoted version of the string
     if (!qouted_input.empty()) {
-      auto fix_scalar = "{ \"Y\" : " + qouted_input + " }";
+      auto fix_scalar = "{\"Y\" : " + qouted_input + "}";
       std::string fix_back;
       auto fix_done = Parse(parser, fix_scalar, &fix_back);
 
diff --git a/tests/fuzzer/scalar_debug.cpp b/tests/fuzzer/scalar_debug.cpp
index 9ce9e5e1..d0c9b407 100644
--- a/tests/fuzzer/scalar_debug.cpp
+++ b/tests/fuzzer/scalar_debug.cpp
@@ -23,6 +23,6 @@ int main(int argc, char *argv[]) {
   auto rc = LLVMFuzzerTestOneInput(
       reinterpret_cast<const uint8_t *>(crash_file_data.data()),
       crash_file_data.size());
-  std::cout << "LLVMFuzzerTestOneInput finished with code " << rc;
+  std::cout << "LLVMFuzzerTestOneInput finished with code " << rc << "\n\n";
   return rc;
 }
diff --git a/tests/fuzzer/scalar_json.dict b/tests/fuzzer/scalar_json.dict
index 60689ddf..7558dc3e 100644
--- a/tests/fuzzer/scalar_json.dict
+++ b/tests/fuzzer/scalar_json.dict
@@ -17,7 +17,11 @@
 "e"
 "f"
 "nan"
+"-nan"
+"+nan"
 "inf"
+"+inf"
 "-inf"
 "infinity"
+"+infinity"
 "-infinity"