purpurproger / dependencyanalyzer Goto Github PK

Issues with populating the "assemblies" variable:

1. (Critical issue) Each assembly is loaded into memory. For my test project that would be 100MB of RAM and 30 seconds of CPU 100%. The problem is application does not use this information. Only the file name is required. As a simple fix You could add file extensions into the "assembliesToLookFor" variable.
   
   In the end the method could be like this:
   
   And the new quick method:
   

2. (Code style issue) AssemblySafeLoad method has side effects. It does load files into memory. I would suggest highlight this in the code by introducing a variable. The problem is the side effects are hidden behind the LINK when reading code quickly. Performance issue like above was hard to find because of the code readability issue.
   

3. (User input issue) Assembly names comparison should be case insensitive. It comes from a kind of user input and could not be trusted. Quick and dirty fix could be like this: