Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx> wrote: > > 2. skip to next and miss an entry > > I think we can still display duplicates in 2. too since nulls check if > the iteration finished on another bucket? Then 2. means skipped > entries and duplicates. Yes, but its not a change in behaviour, this was never a stable walk. > > I'm not sure whats worse/better. > > Skipping looks simpler, it is less code. But if entries are duplicated > then userspace has a chance to deduplicate? Lets just give up and nuke the entire functionality, i.e. /proc interface removal.
