Windows Path Normalize Improvements

[mwasplund]

Windows path normalizing currently uses character toupper conversions. This is ugly, does not handle utf8 multicharacter values and will not work well on Linux.

Within the runtime soup should require exact file path matching, however when monitoring operation evaluations other executables may use non unique paths on windows (ie. C:/Windows/ and C:/WINDOWS). This means on windows we have to have a file path comparison logic that ignores case in utf8 strings in the SystemAccessMonitor but during get attributes calls we need to add extra checks to ensure the file paths are exact to help improve interoperability between linux and windows.