Package: dgit Version: 9.7 Hi. IOhannes, I think you won't mind me copying this private mail of yours to the bug system.
IOhannes m zmölnig (Debian/GNU) writes ("Re: dgit with backports"): > On 8/22/19 1:58 PM, IOhannes m zmölnig (Debian/GNU) wrote: > > it refuses with > >> package appears to be new in this suite; if this is intentional, > >> use --new > > > > is this normal? > > so figuring that my package is indeed not yet in "buster-backports" and > thus might be considered 'new in this suite', i went ahead and did a > "dgit -N push-source" and everything went fine so far. > > so i guess the solution wsa that simple. > > however: when interacting with the repository i tend to be overcareful > (at least, if i don't forget to), and "dgit" is a tool in which my > confidence has yet to grow (not because it has shown to be > untrustworthy, but because it is new for me). > > it might be a good idea to mention "--new" in dgit-maint-bpo(7), as when > working with the "normal" suites (unstable->testing->stable) the package > is usually only "new" at special times (e.g. before the first upload; or > because a package has dropped out of the suite). whereas with backports > you get a new and empty suite with every Debian release. What you say makes sense. I think this is a fine suggestion. When uploading to experimental, needing --new is quite common too, because there often isn't a version in experimental. I'm not sure if this needs to be mentioned anywhere. Out of interest, did you read the description of --new in dgit(1) ? I ask because I want to know if improving it would have helped. I just reread it myself and it definitely should be more reassuring. Thanks, Ian.