[sdcc-devel] LLM use

[Philipp K. K. <pk...@sp...>]

Dear SDCC developers,

recently, we've gotten quite some patches from Christopher Bazley for 
various open bugs. Unfortunately, I have not yet found the time to look 
at the patches myself (but hope to find a bit of time to do so early 
next week).

These patches, as clearly stated were created with the help of LLMs. 
Should that affect how we handle those patches? Do we want or need a 
policy for handling contributions created with the use of LLMs? Also, 
while Christopher's patches are clearly marked as having been created 
with LLMs, that might be different for other contributions we receive, 
i.e. we might not know if a given contribution was created with an LLM.

Personally, I have not formed an opinion yet.
Though I kind of like that SDCC so far operated with very little 
organisational overhead (we don't have any formal structure for the 
project, we don't have a code of conduct, the only written expectations 
on devs is what we have in the email greeting new devs (update 
Changelog, check regression tests before committing, check compile farm 
results after), and the coding style / language dialects wiki pages.

Philipp