You know what would be interesting? Someone should do the following: 1) Get all of the source code for Exim, and other codebases, codebases for which vulnerab

You know what would be interesting? Someone should do the following: 1) Get all of the source code for Exim, and other codebases, codebases for which vulnerabilities have been discovered -- but before the vulnerability has been removed from source code. 2) Run machine learning on the code, with the understanding that different parts of code may have been written by different authors. 3) Use those ML models to see if other codebases, codebases which are as-of-yet unknown if they contain vulnerablilties, to see if those might have the same coding patterns in certain areas as the affected software above, indicating potential tampering. In other words, via ML, software which has been tampered with -- might be able to be detected preemptively... Disclaimer: I am not a ML expert, and this might be a somewhat futuristic idea, due to various constraints...