Hi Roland,
Thanks for your feedback. You raise a very valid point regarding AI-generated code and large-scale refactoring.
To address your concerns while still lightening our process, I propose splitting the rule into two separate cases based on risk, test coverage, and documentation:
1. Review Optional: For minor, manual refactorings—provided they are fully covered by tests AND(/OR?) clear evidence of validation is attached to the PR.
2. Review Mandatory: For major refactorings, AI-generated code, or any changes that lack full test coverage and structural evidence.
Does this compromise sound reasonable to you?
Best regards,
Bruno
Le 17.06.2026 à 08:48, Roland Rosenfeld a écrit :
Hi Bruno!
On Tue, 16 Jun 2026, Bruno Manzoni via Xymon wrote:
To make code review lighter, I propose modifying our governance as follows:
* Refactoring without behavior changes: review optional.
I'm not sure whether this is a good idea. If the refactoring isn't
minimal, there may be issues in the change. Especially if the
refactoring is done by AI, which sometimes doesn't work as exact as it
should do or changes things that don't have to do with the intended
refactoring. This is also the reason for reviewing becoming so time
intensive.
Greetings
Roland