It appears that the CodeResources logic is reading the entire file into memory then calling the appropriate digest. This will be very memory intensive for large applications with lots of resources.

One should just take a path to the file and compute and store the cached checksum vectors into the cache. In my implementation, I am computing both the SHA-1 and SHA-256 simultaneously, and I also copy the file into the target directory if needed, so I only read the file once.

Maybe hold off on implementing this. I will try to get clearance from my employer to submit my changes back to the project.