Hi, On Fri, 2026-07-24 at 16:54 +0200, Matthias Geiger wrote: > > (1) > > I still have to manually grep trough the sources of a package I maintain > > to manually create debian/copyright files instead of uploading the > > source package to an LLM asking for a debian/copyright file (and of > > course sampling whether the output makes sense)? > > > Yes, how every package so far has been created. If we can't trust > contributors to write a proper d/copyright file by themselves we might > aswell give up on the NEW queue. Part of Debians' trust is that we > precisely document every source file shipped by the package.
As someone who reviewed uploads to NEW: No, we cannot trust contributors to write a proper d/copyright file, including project members. You will have packages with auto-generated "TODO" in d/copyright (unedited template), packages only listing the license from the package (not checking if some files use a different license), people saying that they don't bother because the FTP team will check, ... We also never precisely documented every source file as many (most?) people don't bother updating d/copyright when updating the source package. A LLM likely doesn't do a worse job. And probably a significantly better job at keeping the file up to date. (I admit I haven't tried, but the bar is not that high.) Nobody (or close to that) enjoys writing d/copyright or similar tasks. If that can be automated, that's great as people can spend more of their time on more enjoyable things. Personally I also really dislike doing things manually that could be automated that's why I automate (slowly) more tasks, certainly the ones I already don't like doing. It's fine if someone wants to do something manually too and it can be relaxing to do a perfectly automate-able tasks from time to time, but nobody wants to forbid that. Ansgar
