On Tue, Apr 24, 2012 at 12:45 PM, Chris Knadle wrote: >> > Try to read between the lines -- it implies "be reluctant to do an NMU >> > unless you're absolutely sure of what you're doing". That's a much >> > higher bar than the spirit that I think is embodied in Zack's email >> > describing NMUs. >> >> This seems like valuable advice for any upload, including NMU's, so I >> don't think this bar needs to be lowered at all, nor do I think Zack >> promotes that. > > Somehow I interpret that differently.
Being careful with all changes in all uploads is important, but that doesn't mean that taking risks is completely disallowed. The person doing that just needs to be ready to fix the problems he or she created. >> > And there's a difference between being correct and being informative. >> > i.e. just because a statement is correct doesn't mean it conveys what DDs >> > need to know. >> >> Is there a concrete problem with DD's misunderstanding this text? > > Is this an answerable question? What would be required to be able to give a > valid answer? > > I can point to emails of people misconstruing what NMUs are for, but that > doesn't mean they've read the Dev-Ref section or even that they're DDs, and > far harder to prove that it's a common "concrete" problem. As a point of reference, as a (new) DD, I had also drawn the conclusion that new upstream versions were likely disallowed in NMUs, but I reached that from a different direction. My understanding came from the fact that the NMU diff needed to be included in the bug report to make it easy for the maintainer to adopt the NMUer's changes. For new upstream versions, the NMU diff is most often huge and unwieldy, and largely unfriendly thing for the maintainer, so I thought that made such an NMU wrong. I am no longer of that mindset, as I now realize that VCS updates are a user-friendly alternative to large diffs. That is what we are now doing for wine: http://lists.alioth.debian.org/pipermail/pkg-wine-party/2012-April/001815.html You may also want to read some of the other ongooings in wine for the past month (and subscribe to the list): http://lists.alioth.debian.org/pipermail/pkg-wine-party/2012-April/thread.html So, the current wording of the developers reference doesn't preclude this work. That isn't to say that it couldn't be further improved (it could certainly use some wording about VCS work in cases where NMU diffs are far too unwieldy), but it seems to be mostly working as is. >> I do think Lucas is right - you are taking a rather large leap of >> interpretation: from very specific ("no cosmetic changes or switching >> packaging style") to rather generic ("nothing other than critical bugs"). >> There's a host of issues in between, they are not excluded in the text but >> they are excluded in what you say the text 'implies'. I would indeed >> suggest, like Lucas, not to try too hard to find 'implications' or >> 'between the lines' text, which isn't actually there. > > As I mentioned in the my most recent reply, the overall tone of the section > overall is why interpret the wording of the section that way. So, like I said, the NMU section can certainly be improved. The best way to do that is to start a new bug report with a diff of the wording that you think should change. Debian values work, rather than discussion, so please put in the work. Best wishes, Mike -- To UNSUBSCRIBE, email to debian-policy-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CANTw=MNwQe54Gaz16S75u_TY5Dq6K1Uw= laafndz5c2o...@mail.gmail.com
