Ahh, ok, your warnings did seem overly broad to me. How is that enforced on ASF infrastructure? Is it by an “honor system” that we assume no history rewrites are taking place?
Regards, Alan > On Jan 28, 2015, at 10:12 AM, Mark Struberg <strub...@yahoo.de> wrote: > > Deleting those is fine. But we must make sure that no essential commit gets > lost. And this is only doable by totally forbidding history rewrite on the > repo. > > LieGrue, > strub > > > > > >> On Wednesday, 28 January 2015, 18:27, Alan Cabrera <l...@toolazydogs.com> >> wrote: >>> Deleting committed changes is definitely bad. Deleting branches that >>> ultimately >> participate in the release is also bad. >> >> However, deleting tags for "failed" releases is fine as is deleting a >> branch that never transitively participated in a release should be fine as >> well. >> >> Do I misunderstand something? Is there some infrastructure limitation? >> >> Sent from my iPhone >> >> >>> On Jan 28, 2015, at 6:07 AM, Mark Struberg <strub...@yahoo.de> wrote: >>> >>> Hi folks! >>> >>> I've found another thing which is ugly >>> >>> release-tomee-1.7.2 and >>> tomee-1.7.2 branches which are actually just leftovers it seems? >>> >>> >>> Please be aware that we MUST NOT DELETE anything at ASF repos! We must not >> do history rewrite! >>> >>> So think REALLY hard before creating branches which are just utter garbage. >> In git you just cannot get rid of them that easily. They get mirrored >> downstream >> automatically and we have no whatever control about them anymore! >>> >>> Please read how we handle GIT over at DeltaSpike. >>> >>> We e.g. release with localCheckout=true and pushChanges=false. >>> >>> >>> Just ping me if I need to go into details. >>> >>> LieGrue, >>> strub >>