On Thu, Jan 02, 2025 at 11:31:40PM -0700, Theo de Raadt wrote:
> Theo Buehler <tb@theobuehler.org> wrote:
> 
> > On Thu, Jan 02, 2025 at 11:01:12PM -0700, Theo de Raadt wrote:
> > > Theo Buehler <tb@theobuehler.org> wrote:
> > > 
> > > > On Fri, Jan 03, 2025 at 02:08:01AM +0000, Job Snijders wrote:
> > > > > On Fri, Jan 03, 2025 at 01:33:38AM +0000, Job Snijders wrote:
> > > > > > On Fri, Dec 20, 2024 at 04:31:22PM +0000, Stuart Henderson wrote:
> > > > > > > On 2024/12/19 16:31, Job Snijders wrote:
> > > > > > > > Remove BIRD 1 output, because BIRD 1 no longer is maintained &
> > > > > > > > supported. While there, add BIRD 3 support, which has ASPA.
> > > > > > > 
> > > > > > > 2.16 has ASPA too. For simplicity bird3/output_bird3 probably do
> > > > > > > make sense, but would it be worth a quick note in the manual?
> > > > > > > 
> > > > > > > >  .Pa bird
> > > > > > > >  (for bird2)
> > > > > > > > +.Pa bird3
> > > > > > > > +(for bird3)
> > > > > > > 
> > > > > > > (for bird 2.16 or newer)
> > > > > > 
> > > > > > Ok, let's simplify things and output just one BIRD file while
> > > > > > documenting 2.16 or higher is required.
> > > > > > 
> > > > > > The below works with bird2 and bird3 in ports.
> > > > > 
> > > > > deraadt@ suggested to move faster. I too see benefits to focus on only
> > > > > supporting modern versions of BIRD. -T essentially is a kludge to
> > > > > support older versions, so let's remove it right away. With this, the
> > > > > changes since the last portable release are:
> > > > > 
> > > > > * Remove bird1 output, BIRD v1 is deprecated
> > > > > * Remove -T option which was used to customize the BIRD table names
> > > > > * Add ASPA table to bird output, v2.16+ is required going forward
> > > > > 
> > > > > OK?
> > > > 
> > > > I still think we should do this in two steps. Release what's in tree now
> > > > announce removal, then kill this stuff.
> > > 
> > > Why?   The problem is not our userbase, at all.
> > 
> > What makes this so urgent?
> 
> I think you understand how this works.  If you encourage legacy use,
> you have legacy users, and that isn't limited to VMS.

I still don't see an argument to rush this? What's in tree is a minor
breaking change (bird v1 removal) and a deprecation of -T. Let's release
this as discussed in another thread and announce that as of the following
release we will require bird 2.16. No encouragement of legacy use that I
can see.