►
From YouTube: IETF-TOOLS-20220712-1800
Description
TOOLS meeting session at IETF
2022/07/12 1800
https://datatracker.ietf.org/meeting//proceedings/
B
C
E
B
B
B
D
D
B
B
If
we
do
this
before
meaning
we're
going
to
get
a
lot
of
feedback
early
about,
you
know,
what's
not
working
and
what's
making
people
unhappy,
we
do
it
after
the
meeting
is
going
to
trickle
in
over
a
meeting
cycle
and
we'll
get
a
measurable
portion
of
it
during
ietf
115.
The
next
time
people
wake
up
and
bother
to
touch
anything
ietf.
B
B
B
B
Okay
moving
forward,
I
guess
one
last
thing
I
noted
in
the
notes
I'll
call
it
out
here
we
still
haven't
created
a
replacement
for
daily
dose,
I'm
planning
on
letting
that
break
when
we
make
the
transition
and
backfilling
it
later,
when
we
get
the
resources
to
actually
build
out
a
replacement
for
it,
and
we
still
only
have
a
band-aid
in
place
for
the
rpc
to
have
an
online
copy
of
that.
I
think
that
the
rpc
discovered
that
they
could
use
bap
locally
as
well.
B
B
B
B
We're
asking
a
few
people
who
fell
into
the
signed
up
and
paid,
but
didn't
attend
any
sessions
to
verify
that
they
didn't
attend
any
sessions
to
make
sure
that
we
don't
have
a
problem
with
what
was
happening
at
medico,
but
so
far
we
have
no
indication.
No
no
reason
to
believe
that
there
was
a
problem
there.
B
D
B
B
B
B
B
Test
client
is
up
and
ready
to
go
for
tomorrow's
chair
testing
greg.
If
you
have
any
time
this
afternoon,
it
might
be
nice
to
get
in
ahead
of
ahead
of
that
chair
testing,
you
and
I
can
start
poking
at
things.
One
thing
that
I
have
noted
already
is
that
any
zoo
of
traffic
that
is
not
sent
to
the
jabber
topic
is
not
bridged
to
jabber
at
all.
B
B
B
B
B
We
had
quite
a
bit
of
activity
since
last
meeting
to
check
the
configuration
and
adjust
the
configuration
of
many
mailing
lists.
The
working
group
and
research
mailing
lists,
in
particular
to
ensure
they
were
configured
to
use
the
global
white
list
and
their
accept
these
non-members
setting
their
non-working
group
lists
are
still
being
processed.
Cindy
is
sent
an
update
with
the
particulars
of
that
team.
Several
of
the
people
are
involved,
but
I
didn't
copy
them
in
into
here.
I
think
it's
sufficient
to
know
that
it
is
it's
a
work
in
progress.
B
I
don't
remember
it
was
late
that
late
last
week
or
early
this
week,
it
was
late
last
week
that
there
was
a
flaw
in
the
script
that
updates
the
global
white
list
list
membership.
The
thing
that
at
global
white
list
in
the
except
non-members
actually
reads
the
program
that
calculated
the
new
global
white
list
calculated
the
list
just
fine,
the
step
that
merged
that
into
the
list.
Membership
for
that
global
white
list
list
was
commented
out
by
who
knows
how
long
ago
that's
been
corrected.
B
B
G
Honestly,
yeah,
I
didn't
know
you
were
doing
the
reverse
rewrite,
because
I
would
have
told
you
not
to
do
that
because
it'll
make
because
it'll
it'll
it
will
cause
exactly
the
failures
we're
seeing.
I
mean
I
don't
know
how
hard
this
is
to
implement,
but
if,
if
we
could
only
do
the
reverse
rewrite
on
mail
going
into
the
archives
not
now
going
out
to
people
that
will
probably
decrease
the
failure
rate
somewhat,
because
it
will
break
fewer
dkim
signatures.
G
And
beyond
that,
the
stuff
from
the
I
mean
yeah,
the
problem,
the
problem
with
diagnosing
gmail
failures
is
gmail,
has
a
sophisticated
machine
learning
system
that
adjusts
the
filters
per
user.
You
know
so
it's
entirely
common
for
something
like
it
work.
You
know,
user
one
says
it
fails.
User
two
says
it
works,
fine
for
me
and
they're
both
telling
the
truth
I
mean.
So
I
think
you
know
the
the
rewrite
we
are
doing
is
about
the
best
you
can
do.
I
mean
it's
based
on
one
that
I've
been
using
for
years.
G
That
seems
to
work
pretty
well
for
me.
If
we
can
actually
collect
a
few
more
failure
reports
like
the
charles's
failure
reports,
I
actually
figured
out
what
the
problem
is
there
because
his
mail
is
only
authenticated
with
spf.
So
when
we
relay
it
and
they
and
we
don't
it-
you
know
it
causes
authentication
failure.
So
it's
like
these
are,
I
believe,
there's
some
fairly
straightforward
ways
we
can.
We
can.
We
can
tweak
this,
but
I
can't
ever
promise.
It
will
be
perfect.
B
B
The
biggest
issue
that
we
have
with
changing
the
way
the
reverse
rewriter
behaves
is
very
similar
to
why
we
aren't
doing
forward
rewriting
for
the
aliases
is
that
our
hook
for
it
is
at
a
place.
That's
a
junction
to
the
handoff
to
mailman
just
to
be
happen
to
be
a
convenient
place
where
we
could
put
post
confirm
in
the
path
of
the
message
and
have
it
modify
things
when
we
move
to
mailman
3,
almost
all
of
the
connections
between.
B
B
B
B
Then
there
were
some
final
edits
being
done,
t
that
google
document
and
I'm
hoping
that
this
week
it
will
go
out
for
feedback
and
we
can
start
executing
to
it.
The
integration
with
the
data
tracker
credentials
is
complete,
tested
appears
to
be
functioning
and
I
think
we're
ready
to
start
the
the
transition.
E
B
B
For
the
data
tracker,
a
lot
of
activity
since
the
last
meeting,
you'll
get
the
release
notes
for
eight
four
through
eight
six
one,
a
couple
of
things
to
point
out.
We
have,
as
mentioned
earlier,
explicit
capture,
recession
attendance
and
we
have
given
people
the
ability
to
provide
personal
pronouns.
I
haven't
checked
to
see
how
many
people
have
taken
advantage
of
that
yet,
but
the
plumbing
is
already
in
place
in
with
medico.
B
At
the
same
time,
the
we
also
have
the
processing
for
the
asynchronous
processing
for
submissions
through
new
api,
so
the
first
step
towards
making
even
the
web
interface
submissions
async
and
by
async
I
mean
somebody
hit
submit.
They
don't
have
to
wait
until
all
of
the
tests
are
done
before
the
thing
comes
back
and
acknowledges
that
they've
made
a
submission.
It
will
be
closer
to
what
you're
used
to
with
other
modern
websites
where
it
comes
back
and
says:
okay
right
away,
and
it
gives
you
a
progress
meter
about.
B
You
know
what
steps
it's
going
through
and
what
things
fail
when
the
deployment
of
this
involves
some
components
that
we've
not
had
for
the
data
tracker
before
revenue
and
qm
celery
we're
going
to
be
running
celery
in
a
docker
container,
the
the
celery
instance
that
is
doing
data
tracker
work
and
I'm
working
through
the
deployment
instructions
with
jennifer
right
now
again.
This
is
something
I
hope
that
we
have
deployed
either
later
this
week
or
the
first
part
of
next.
D
B
G
B
B
The
big
thing
we're
going
to
have
in
progress
over
the
next
couple
of
weeks
is
the
time
zone
where
timestamp
work.
You
remember,
we've
been
struggling
with
this
thing
for
many
months.
It's
been
cycle,
starved
we're
able
to
give
it
cycles
again.
Recently,
jennifer's
found
a
way
that
we
believe
we
can
get
through
the
data
migration
in
the
order
of
minutes
instead
of
on
the
order
of
hours,
as
was
the
plan
back
last.
B
So
the
table
has
mixed
data,
the
tables
that
are
related
to
it
have
mixed
data.
Just
what
the
timestamp
means
depends
on
where
in
time
that
time
stamp
happened
to
land-
and
you
have
to
know
about
where
the
time
zones
of
the
meeting
were
now
we're
going
to
be
building
a
large
set
of
migrations
and
changes
to
the
rendering
code
that
converts
everything
to
utc
and
the
time
stamps
that
aren't
currently
in
meeting
time.
B
So
this
is
going
to
be
the
big
chunk
of
our
effort,
the
priorities
that
we
have
queued
up
after
that
I
have
listed
in
that
upcoming
session.
None
of
those
are
are
should
be
particularly
surprising
eric.
It
would
be
good
for
you
to
reconfirm
with
the
isg
since
enough
time
has
gone
by
that
people
may
have
forgotten
that
we
had
agreement
to
start
requiring
that
submissions
be
fully
expanded
and
that
we
need
to
find
a
way
to
give
people.
H
H
H
I'm
also
working
on
vc
print
upgrade
and
that's
list
of
that's
this
bunch
of
pdf
rx
is
generated
with
the
newbies
print
version.
We
sprint
upgrade
and
people
are
reviewing
it.
I
can
add
that
list
to
the
ticket
if
anyone
wants
to
go
and
like
help
us
reviewing
changes
this.
This
change
involves
in
a
pdf
rendering
engine
chain,
so
this
print
is
moving
from
cairo
to
its
own
different
pdf,
rendering
engine
so
there'll
be
some
feed.
The
pdfs
generated
from
both,
which
two
versions
will
be
different.
A
Yeah
yeah,
so
just
briefly,
I
just
submitted
some
changes,
basically
a
refresh
of
the
container
setup
that
we
have
in
the
repo
for
mail
archive
that
you
can
use
in
your
development
environment
that
had
gotten
a
bit
stale
and
then
upcoming.
Actually
the
same
for
both
projects
archive
and
the
red
system
is
a
move
to
engine
x
as
the
front
end,
which
will
allow
us
to
do
the
upgrade
to
python39.
B
D
So
we
get
a
meeting
with
all
the
what
they
call
the
proponents.
So
it's
bernoulli's
and
a
few
others,
including
in
this
case
carsten
and
michael
by
the
way,
but
not
yet
pantene-
is
a
contractor
about
the
next
step
for
young
catalogue
and
the
big
one
is
really
about
the
seed,
and
I
see
a
couple
of
chat
on
this.
We
need
to
get
moving
the
first
six
months
with
pantheon,
where
basically
addressing
minor
or
medium
edit
features,
but
I
think
the
major
one
will
be
basically,
this
schema
id
there.
D
There
are
still
two
improvements
in
progress,
but
I
would
really
like
to
point
is
that
the
server
stability
and
performance
is
much
better
compared
to
one
year
or
two
years
ago.
Thanks
to
greg
and
others,
we
know
got
some
atomos
statistics
and
we
have
about
200
unique
visitors
per
week,
which
is
much
higher
than
I
was
expecting
to
be
honest,
mainly
using
the
the
search
or
when
they
are
searching
for
life
or
module
or
whatever
and
the
details.
D
An
addition
is
to
also
cover
the
api,
because
we
don't
know
how
many
people
are
using
the
api
rather
than
the
web
interface
and
greg.
I'm
not
sure
whether
it's
the
right
place
to
talk
about
it,
but
panty
on
the
contractors
is
requested
the
atf
in
an
email
yesterday.
I
guess
so
the
last
week
whether
they
can
use
yancatalog.org.
D
B
D
E
B
F
B
F
Okay,
so
okay,
so
I
thought
that
was
one
of
the
the
design
features
was
that
we
would
have
a
diff
that
I
would
be
able
to
refer
to
that
was
created
from
well.
You
know
the
last
version
of
the
public
document
and
the
the
version
of
the
document
I
just
generated
on
my
own
computer,
but
haven't
put
anywhere
and
that
I
could
then
share
that
url.
F
So
I
can
do
that
right
now
I
with
rfc
diff,
but
then
the
result
is
private
to
me,
which
makes
it
hard
for
me
to
say
these
are
the
things
that
I'm
the
new
document
has,
and
I
thought
that
that
one
of
the
design
things
in
this
revision
was
that
we
were
going
to
wind
up
with
urls
that
were
stable,
even
though
we
hadn't
that
weren't
automat
weren't.
What's
the
rating.
F
Be
somewhere
yeah
so
so
then
I
ran
into
the
list
of
urls
that
it
would
accept,
which
was
did
not
include
my
site.
So
then
I
had
to
put
the
file
in
on
a
branch
and
github
and
look
at
that
and
anyway.
I
know
how
to
do
that
and
that's
okay.
For
me.
I
think
that
may
be
a
little
bit
difficult
for
others.
So
I
think
you
consider
how
that
was
gonna
work
I
mean.
Maybe
it's
all
gonna
work
great
for
the
full
integrated.
F
B
F
F
There
was
a
design
point
which
was
where
it
was
going
to
be
actually
not
generate
all
the
diffs
on
the
fly,
but
rather
was
going
to
be
able
to
store
the
diff
that
it
produced
for
a
certain
amount
of
time
such
it
will
be
cacheable,
etc.
Blah
blah
blah
but
you're,
saying
that's,
not
wasn't
a
design
goal
in
the
end.
Okay,.
B
F
Well,
but
if
it
disappears
under
the
intermediate
cache-
and
I
give
you
start-
then
it's
gonna-
it's
not
gonna
work
right.
So
so
I
I
don't
know
that
for
the
case
that
I
upload
a
file,
there's
any
value
in
setting
a
cache.
It's
cacheable,
it's
not
it's,
not
cacheable.
It's
got
to
process
it
again
right,
whereas
if
I,
if
I
set
two
urls
and
I
give
you
the
diff,
which
is
something
I've
done
regularly
for
people,
then
you
know
especially
with
you
know,
isg
discusses
this
is
the
changes
I'm
proposing
here.
B
B
D
I
got
a
request
from
one
person
whether
we
can
run
hackathon
per
working
group
and
removed.
Obviously
I'm
not
sure
we,
I
know
we
provide
tools
for
remove
the
cattle.
I
no
clue
what
those
tools
are.
I
got
the
request
this
morning,
my
time,
but
can
we
replicate
them
on
a
per
working
group
interim,
maybe
running
24
hours
around
the
clock
for
one
day.
B
C
D
C
E
B
G
B
B
What
he
reported
for
everybody
else's
benefit
was
an
issue
where
someone
from
japan
submitted
a
draft
where
the
date
was
from
the
server's
perspective
tomorrow,
because
it
was
tomorrow
when
in
japan
and
the
server
complained
I
looked
at
the
code.
The
code
has
a
three-day
allowance
both
ways,
so
something
else
was
going
on
to
cause
that
t
error
to
get
thrown
quite
likely.
B
Yeah
we'll
go
we'll
go
find
it
the
when
we
make
the
changes
to
the
data
charter
become
time
zone
aware.
One
of
the
things
that
we're
planning
to
do
is
at
all
places
in
the
interface
where
we
present
deadlines,
we're
going
to
make
it
pretty
clear
what
the
deadline
is
down
to
the
minute.
We've
been
fuzzy
in
the
past
with
what
that
meant.
B
This
is
something
the
isg
is
probably
going
to
have
to
discuss
and
become
comfortable
with.
We
like
for
the
draft
cutoffs
we
may
set
it
to.
You
know
some,
maybe
even
not
midnight
utc,
but
you
know
set
it
at
some
time,
but
there
is
that
precise
time
where
things
will
be
cut
off
and
the
and
it'll
be
well
well
advertised
and
easy
for
people
to
get
both
in
utc
and
in
their
local
time
zone.