ask CL preprocessor for UTF-8 output

Author: Daniel Kroening

appveyor.yml

@@ -71,7 +71,6 @@ test_script:
     rmdir /s /q ansi-c\Forward_Declaration2
     rmdir /s /q ansi-c\Incomplete_Type1
     rmdir /s /q ansi-c\Union_Padding1
-    rmdir /s /q ansi-c\Universal_characters1
     rmdir /s /q ansi-c\function_return1
     rmdir /s /q ansi-c\gcc_attributes7
     rmdir /s /q ansi-c\gcc_version1

buildspec-windows.yml

@@ -35,7 +35,6 @@ phases:
         Remove-Item ansi-c\Forward_Declaration2 -Force -Recurse
         Remove-Item ansi-c\Incomplete_Type1 -Force -Recurse
         Remove-Item ansi-c\Union_Padding1 -Force -Recurse
-        Remove-Item ansi-c\Universal_characters1 -Force -Recurse
         Remove-Item ansi-c\function_return1 -Force -Recurse
         Remove-Item ansi-c\gcc_attributes7 -Force -Recurse
         Remove-Item ansi-c\gcc_version1 -Force -Recurse

src/ansi-c/c_preprocess.cpp

@@ -335,8 +335,13 @@ bool c_preprocess_visual_studio(
     // understands.
     command_file << char(0xef) << char(0xbb) << char(0xbf);
 
-    command_file << "/nologo" << "\n";
-    command_file << "/E" << "\n";
+    command_file << "/nologo" << '\n';
+    command_file << "/E" << '\n';
+
+    // This option will make CL produce utf-8 output, as
+    // opposed to 8-bit with some code page.
+    command_file << "/source-charset:utf-8" << '\n';
+
     command_file << "/D__CPROVER__" << "\n";
     command_file << "/D__WORDSIZE=" << config.ansi_c.pointer_width << "\n";