Hello,

We do not want AI-generated Slop in our source tree. I get that and I think 
most others on this list will, too. But let us also think about the other end: 
What do we want to use those AI tools for - hey, come on, that "nothing" came a 
bit quickly.

We have taken some decisions already:

 * cppcheck: We have our code screened for what is an obvious contender of 
potential mayhem or just ugly.

And frankly, most of these issues we would not have detected in any near 
future. Some user of LinuxCNC somewhere on this world may have run into some 
weird situation that would be non-reproducible, and likely never be reported, 
just some random fluke, but we go out and fix any such potential contenders. I 
would go as far to say that we fix reports that are false positive if this 
helps the general understanding of the code.

These LLMs are exceptionally good in finding potential issues, too. Yes, they 
also propose patches to fix those issues, but you do not need to apply those 
but can implement this yourself. So:

 * LLMs: Code hardening.

Also, I think we should embrace them for

 * Help with the installation of LinuxCNC.
 * Support for writing error reports when installation fails?

Best,
Steffen




_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to