DrewBarlow / CPPCheckDashboard

Planning the backend for a coverage dashboard

Notes:

• sample_data.zip contains a randomly selected chunk of files from the SARD dataset.
• The chunk in question pertains to CWE36, and 78 files are present.

PLANNING:

Goal:

• Generate a .html file that summarizes what sections of code were covered by CPPCheck.
  • i.e., highlight covered sections with green and uncovered sections with red.
  • Comments will be ignored always.
  • Need to take into account multiple preprocessor configs.

Approaches:

1. Configure CPPCheck to generate these files.
   • Pros
     • Faster execution time.
     • Easy to handle variability of multithreaded execution (global shared array + mutex).
   • Cons
     • More difficult to format.
     • Would take significantly longer to implement.
     • Requires HEAVY modification of source code.
2. Create some locally-hosted server to accept results of execution as it happens.
   • Explained differently, use sockets and a threading TCP server to record results and handle them.
   • Pros
     • Server would be easily implementable in Python.
     • Less time to implement.
   • Cons
     • I'm iffy with c-style sockets.
     • Requires more computation than is really needed.
     • Requires HEAVY modification of source code.
3. Use a wrapper program to run CPPCheck and generate the files accordingly.
   • Explained differently, spawn a cppcheck process and capture stdout.
   • Pros
     • Arguably the simplest solution in terms of being able to apply it to other analyzers.
       • Other analyzers would just need to be modified to print similar output to CPPCheck.
     • Relatively quick to implement.
   • Cons
     • A hurdle would be handling the variability of stdout due to multithreading.