Hi, On 2024-11-05 12:34, Santiago Vila wrote: > El 5/11/24 a las 6:26, Aurelien Jarno escribió: > > On 2024-11-05 00:33, Santiago Vila wrote: > > > > please find the debdiff attached > > > > > > Hi. Can you post the debdiff? (or just say what you did) > > > > Oops sorry about that. Please find it attached. > > Thank you. I've made a team upload with those changes.
Thanks! > > > NMUs are messy for my taste. I could fix the FTBFS problem and let Anthony > > > (the usual uploader) to decide (for a later upload) if he wants > > > to use a depends, a recommends, or a suggest for the binary dependency > > > on tzdata-legacy (or even nothing). > > > > I can cancel it no problem if you think you have a better fix. > > My aim was to be "conservative". This is the "dilemma" I had in mind: > > - For these kind of team uploads, I try to follow the spirit of a NMU, > i.e. minimal changes (in fact, I joined this team to do QA work). > - In this case, the minimal change to solve the FTBFS problem > would have been to fix the build-depends only. > - However, given than tzdata has been split into > tzdata + tzdata-legacy, the change which really keeps things > "as similar as before" is to add both depends and build-depends. > > So that's what I did in the end (i.e. apply the same changes > in the nmudiff) to be "neutral". > > (And of course, after this upload, Anthony can still relax the > dependency on tzdata-legacy if he considers it appropriate). Ok. If doing so, the autopkgtest has to be adapted because it also relies on tzdata-legacy. Regards Aurelien -- Aurelien Jarno GPG: 4096R/1DDD8C9B aurel...@aurel32.net http://aurel32.net
