http://bugs.winehq.org/show_bug.cgi?id=60241 --- Comment #15 from Stian Low <wineryyyyy@gmail.com> ---
Actually they did point out 3 issues that are very much real bugs. There are many more real bugs in that report. Even if you are not a fan of AI, you have to admit that it is useful data.
Knowing how the report was produced may also be useful for WineHQ. For example, if it just used Coverty with different settings then that's something WineHQ could reproduce but more integral with existing CD/CI workflows. However if it behaves more like a black box unable to clarify to someone how to reproduce the report and is considered a competitor vs user of Coverty then it approaches conflict with WineHQ's current LLM policy.
If you are participating in bug triaging, you are the public face of the project, and you have a responsibility to be nice and welcoming to users, even if they don't conform to your strict idea of the rules. If you can't do that, you'll be asked to stop participating.
Generally anyone reporting any bug should be met with appreciation. Excessive slop may warrant criticism but when human touch is sensed such as the case of this report then it should be handled in a less rude manner. Other LLM reports with less human touch may need some extra aggressiveness to help them learn better. Bug reporters stating upfront that they used AI should also be appreciated because it helps devs with context which may warrant more skepticism. Also, LLMs that only report bugs and not patches should also be appreciated because they are learning to conform with WineHQ's LLM policy which considers patch risky since internal Microsoft code could have been used as part of training thus violating mission critical clean rule policies. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.