Small fixes from fuzzing and regression tests.

Author: wcjohns

CMakeLists.txt

@@ -231,7 +231,7 @@ set_target_properties( SpecUtils PROPERTIES PREFIX "lib" OUTPUT_NAME "SpecUtils"
 if( SpecUtils_USE_FROM_CHARS )
   if( NOT (MSVC AND (MSVC_VERSION GREATER_EQUAL 1920)) )
     # MSVC >=2019 supports floating point from_chars, while Xcode 14 still doesnt (havent checked Android Studio)
-    message("Please dopuble check your compiler does actually support std::from_chars for parsing floats")
+    message("Please double check your compiler does actually support std::from_chars for parsing floats")
   endif()
 
   set_target_properties( SpecUtils PROPERTIES

SpecUtils/ParseUtils.h

@@ -125,6 +125,8 @@ namespace  SpecUtils
      ex: "Title: FA-SG-LANL-0-0-8{dx=-155.6579,dy=-262.5} @235cm H=262.5cm V=221.1404cm/s : Det=Ba2"
    
      TODO: does not currently take into account units (e.g., cm)
+   
+   Throws exception if not successful.
   */
   float dx_from_remark( std::string remark );
 
@@ -133,6 +135,8 @@ namespace  SpecUtils
     ex: "Title: FA-SG-LANL-0-0-8{dx=-155.6579,dy=-262.5} @235cm H=262.5cm V=221.1404cm/s : Det=Ba2"
    
     TODO: does not currently take into account units (e.g., cm)
+   
+   Throws exception if not successful.
   */
   float dy_from_remark( std::string remark );
 

fuzz_test/CMakeLists.txt

@@ -5,17 +5,21 @@ cmake_policy(SET CMP0079 NEW) #allow altering SpecUtils linking from this file
 
 add_executable( file_parse_fuzz file_parse_fuzz.cpp fuzz_interface.h fuzz_interface.cpp )
 
+# spirit float parser can assert with inputs like 
+#  ".000000000000000000000000000000000000000000000000000000000000000000000000000" 
+#  when we compile as RelWithDebInfo; so we'll disable this asserting for fuzzing
+target_compile_definitions( SpecUtils PUBLIC BOOST_DISABLE_ASSERTS )
+
 target_compile_options( file_parse_fuzz PRIVATE $<$<CXX_COMPILER_ID:Clang>:-g -O1 -fsanitize=fuzzer,address,undefined -fprofile-instr-generate -fcoverage-mapping> )
 target_compile_options( SpecUtils PRIVATE $<$<CXX_COMPILER_ID:Clang>:-g -O1 -fsanitize=fuzzer,address,undefined -fprofile-instr-generate -fcoverage-mapping> )
 target_link_libraries( SpecUtils PUBLIC $<$<CXX_COMPILER_ID:Clang>:-fsanitize=fuzzer,address,undefined -fprofile-instr-generate -fcoverage-mapping > )
 target_link_libraries( file_parse_fuzz PRIVATE SpecUtils $<$<CXX_COMPILER_ID:Clang>:-fsanitize=fuzzer,address,undefined> )
 set_target_properties( file_parse_fuzz PROPERTIES CXX_STANDARD 14 CXX_STANDARD_REQUIRED YES CXX_EXTENSIONS NO )
 
-
 #add_executable( run_one_file_parse run_one_file_parse.cpp fuzz_interface.h fuzz_interface.cpp )
 #target_link_libraries( run_one_file_parse PRIVATE SpecUtils )
 #set_target_properties( run_one_file_parse PROPERTIES CXX_STANDARD 11 CXX_STANDARD_REQUIRED YES CXX_EXTENSIONS NO )
 
 add_executable( fuzz_str_utils fuzz_str_utils.cpp )
-target_link_libraries( fuzz_str_utils PRIVATE SpecUtils $<$<CXX_COMPILER_ID:Clang>:-fsanitize=fuzzer,address,undefined> )
+target_link_libraries( fuzz_str_utils PRIVATE SpecUtils $<$<CXX_COMPILER_ID:Clang>:-fsanitize=fuzzer,address,undefined -fprofile-instr-generate -fcoverage-mapping> )
 set_target_properties( fuzz_str_utils PROPERTIES CXX_STANDARD 14 CXX_STANDARD_REQUIRED YES CXX_EXTENSIONS NO )

fuzz_test/fuzz_interface.cpp

@@ -275,6 +275,20 @@ int run_file_parse_fuzz( const uint8_t *data, size_t size )
       test_write_output( spec );
   }
 
+  {
+    SpecUtils::SpecFile spec;
+    stringstream strm( datastr, ios_base::in );
+    if( spec.load_from_json( strm ) )
+      test_write_output( spec );
+  }
+
+  {
+    SpecUtils::SpecFile spec;
+    stringstream strm( datastr, ios_base::in );
+    if( spec.load_from_caen_gxml( strm ) )
+      test_write_output( spec );
+  }
+
 
   return 0;
 }

fuzz_test/fuzz_str_utils.cpp

@@ -649,9 +649,9 @@ extern "C" int LLVMFuzzerTestOneInput( const uint8_t *data, size_t size )
 
   print_waypoint( 48 );
 
-  speed_from_remark( datastr );
-  speed_from_remark( str_1 );
-  speed_from_remark( str_2 );
+  try{ speed_from_remark( datastr ); }catch(std::exception &){}
+  try{ speed_from_remark( str_1 ); }catch(std::exception &){}
+  try{ speed_from_remark( str_2 ); }catch(std::exception &){}
 
   print_waypoint( 49 );
 
@@ -661,19 +661,19 @@ extern "C" int LLVMFuzzerTestOneInput( const uint8_t *data, size_t size )
 
   print_waypoint( 50 );
 
-  dx_from_remark( datastr );
-  dx_from_remark( str_1 );
-  dx_from_remark( str_2 );
+  try{ dx_from_remark( datastr ); }catch(std::exception &){}
+  try{ dx_from_remark( str_1 ); }catch(std::exception &){}
+  try{ dx_from_remark( str_2 ); }catch(std::exception &){}
 
   print_waypoint( 51 );
 
-  dy_from_remark( datastr );
-  dy_from_remark( str_1 );
-  dy_from_remark( str_2 );
+  try{ dy_from_remark( datastr ); }catch(std::exception &){}
+  try{ dy_from_remark( str_1 ); }catch(std::exception &){}
+  try{ dy_from_remark( str_2 ); }catch(std::exception &){}
 
   print_waypoint( 52 );
 
-  dose_units_usvPerH( (const char *)data, size );
+  try{ dose_units_usvPerH( (const char *)data, size ); }catch(std::exception &){}
   //dose_units_usvPerH( datastr.c_str(), datastr.size() );
   //dose_units_usvPerH( str_1.c_str(), str_1.size() );
   //dose_units_usvPerH( str_2.c_str(), str_2.size() );

src/Filesystem.cpp

@@ -543,6 +543,13 @@ std::string parent_path( const std::string &path )
   while( strcmp(bname,"..") == 0 )
   {
     char *parname = dirname( &(pathvec[0]) );
+    if( !parname )
+    {
+      // I think we can get here for like invalid utf-8 characters or something, or maybe
+      //  just really long input, or maybe invalid paths; not sure, only hit this during fuzzing
+      break;
+    }
+    
     size_t newlen = strlen(parname);
     pathvec.resize( newlen + 1 );
     pathvec[newlen] = '\0';

src/SpecFile_json.cpp

@@ -306,7 +306,7 @@ bool SpecFile::load_from_json( std::istream &input )
       meas->parse_warnings_.insert(end(meas->parse_warnings_), begin(warnings), end(warnings));
 
     measurements_.push_back(meas);
-    instrument_type_ = "Gamma Handheld";
+    instrument_type_ = "";
     manufacturer_ = "Bridgeport Instruments";
     instrument_model_ = "eMorpho";
     instrument_id_ = serial_number;
@@ -324,7 +324,7 @@ bool SpecFile::load_from_json( std::istream &input )
   }//try / catch
 
   return false;
-}//bool load_from_aram( std::istream &input )
+}//bool load_from_json( std::istream &input )
 }//namespace SpecUtils
 
 

src/SpecFile_xml_other.cpp

@@ -626,7 +626,7 @@ bool SpecFile::load_from_caen_gxml(std::istream& input)
     remarks_ = remarks;
     manufacturer_ = "CAEN";
     instrument_model_ = "Hexagon";
-    instrument_type_ = "Other";
+    instrument_type_ = "";
     detector_type_ = DetectorType::Unknown;
 
     cleanup_after_load();