bail out instead of crash when parsing unicode code (#207)

src/cppcheck.cpp

@@ -32,6 +32,7 @@
 #include <cstring>
 #include <fstream>
 #include <map>
+#include <stdexcept>
 
 //---------------------------------------------------------------------------
 
@@ -297,9 +298,12 @@ unsigned int CppCheck::check()
         if (_settings._errorsOnly == false)
             _errorLogger->reportOut(std::string("Checking ") + fname + std::string("..."));
 
+        try
+        {
             Preprocessor preprocessor;
             std::list<std::string> configurations;
             std::string filedata = "";
+
             if (_fileContents.size() > 0 && _fileContents.find(_filenames[c]) != _fileContents.end())
             {
                 // File content was given as a string
@@ -336,6 +340,12 @@ unsigned int CppCheck::check()
                 checkFile(codeWithoutCfg, _filenames[c].c_str());
                 ++checkCount;
             }
+        }
+        catch (std::runtime_error &e)
+        {
+            // Exception was thrown when checking this file..
+            _errorLogger->reportOut("Bailing out from checking " + fname + ": " + e.what());
+        }
 
         _errorLogger->reportStatus(c + 1, _filenames.size());
     }

src/preprocessor.cpp

@@ -23,6 +23,7 @@
 #include "token.h"
 
 #include <algorithm>
+#include <stdexcept>
 #include <sstream>
 #include <fstream>
 #include <iostream>
@@ -71,7 +72,7 @@ std::string Preprocessor::read(std::istream &istr)
     for (char ch = readChar(istr); istr.good(); ch = readChar(istr))
     {
         if (ch < 0)
-            continue;
+            throw std::runtime_error("The code contains characters that are unhandled");
 
         if (ch == '\n')
             ++lineno;

test/testpreprocessor.cpp

@@ -27,6 +27,8 @@
 #include "../src/tokenize.h"
 #include <map>
 #include <string>
+#include <sstream>
+#include <stdexcept>
 
 extern std::ostringstream errout;
 
@@ -115,6 +117,8 @@ private:
         TEST_CASE(pragma);
         TEST_CASE(endifsemicolon);
         TEST_CASE(missing_doublequote);
+
+        TEST_CASE(unicode1);
     }
 
 
@@ -850,6 +854,13 @@ private:
             ASSERT_EQUALS("[abc.h:2]: (error) No pair for character (\"). Can't process file. File is either invalid or unicode, which is currently not supported.\n", errout.str());
         }
     }
+
+    void unicode1()
+    {
+        const char filedata[] = {'a', (char)200, 0};
+        std::istringstream istr(filedata);
+        ASSERT_THROW(Preprocessor::read(istr), std::runtime_error);
+    }
 };
 
 REGISTER_TEST(TestPreprocessor)

test/testsuite.cpp

@@ -123,6 +123,12 @@ void TestFixture::assertEquals(const char *filename, int linenr, unsigned int ex
     assertEquals(filename, linenr, ostr1.str(), ostr2.str());
 }
 
+void TestFixture::assertThrowFail(const char *filename, int linenr)
+{
+    errmsg << "Assertion failed in " << filename << " at line " << linenr << std::endl
+    << "The expected exception was not thrown" << std::endl;
+}
+
 void TestFixture::printTests()
 {
     const std::list<TestFixture *> &tests = TestRegistry::theInstance().tests();

test/testsuite.h

@@ -41,6 +41,7 @@ protected:
 
     void assertEquals(const char *filename, int linenr, const std::string &expected, const std::string &actual);
     void assertEquals(const char *filename, int linenr, unsigned int expected, unsigned int actual);
+    void assertThrowFail(const char *filename, int linenr);
 
 public:
     virtual void reportOut(const std::string &outmsg);
@@ -58,6 +59,7 @@ public:
 
 #define TEST_CASE( NAME )  if ( runTest(#NAME) ) NAME ();
 #define ASSERT_EQUALS( EXPECTED , ACTUAL )  assertEquals(__FILE__, __LINE__, EXPECTED, ACTUAL)
+#define ASSERT_THROW( CMD, EXCEPTION ) try { CMD ; assertThrowFail(__FILE__, __LINE__); } catch (EXCEPTION &) { } catch (...) { assertThrowFail(__FILE__, __LINE__); }
 #define TODO_ASSERT_EQUALS( EXPECTED , ACTUAL ) if (EXPECTED==ACTUAL) assertEquals(__FILE__, __LINE__, "TODO assertion", "The assertion succeeded")
 #define REGISTER_TEST( CLASSNAME ) namespace { CLASSNAME instance; }