Re: When not to close a bug in a changelog...
> Alternatively, if you include some of the changes from the NMU and
> revert others, you should document *that*, not waste bits talking about
> changes that aren't. IOW, the assumption should be that changes are
> carried forward, unless it specifically says that they aren't.
This assumes that the maintainer has seen the nmu get done, and has then gone
the extra step(s) of fetching the source from the archive.
Of course, if the maintainer has done changes locally to their source, then
this will cause them extra work.
As always, when *EVER* an nmu is done, send communication. Don't just blindly
upload(even to delayed).