►
From YouTube: Working Group: 2020-11-17
Description
* Ruby Deprecation Dates
A
A
B
Not
right
now,
some
a
lot
of
folks
were
tagged
requesting
reviews.
So
if
you
want
to
weigh
in
asynchronously,
that
would
be
great
thanks.
B
A
A
C
I
think
we
are
still
waiting
on
clarification
as
to
whether
or
not
the
paquetto
stuff
was
covered
under
the
existing
cff
agreement
with
docker
hub
so
like
pending
some
clarification
there
about
whether
or
not
we
need
to
like
resubmit,
an
application
for
paquetto
separately
from
the
cff.
C
A
If
he
doesn't
do
it
well,
he's
not
gonna
do
well
he's
on
holiday.
Probably
I
may
reach
out
and
see
if
I
can
get
an
earlier
answer
than
waiting
for
him
to
return
we'll
see.
A
D
Yeah,
so
this
is
in
an
effort
to
like
move
from
buildback
gamble
to
environment
variables,
like
we've
been
doing
in
the
dot
net
and
the
and
the
go
families
yeah.
So
nothing
much
new
here.
So
please
take
a
look
in
a
way
and
if
you
have
comments.
B
A
Okay:
okay,
good.
A
Okay,
ruby
deprecation
dates.
F
Yes,
there's
a
conversation
that
started
in
slack
about
filling
in
ruby,
deprecation
dates
and,
from
our
end,
I
think
it's
a
little
challenging
for
us,
because
ruby
does
not
publish
deprecation
dates
when
new
versions
come
out.
I
think
the
request
was
to
like
start
adding
them
as
ruby
announces,
deprecation
dates,
and
so
that
would
definitely
pose
a
bit
of
a
challenge
for
our
current
implementation
of
learning
about
new
dependencies.
F
I
was
kind
of
curious
what
your
thoughts
on
like
if
we
determined
this
would
be
too
difficult
to
do
what
what
mitigations
there
would
be
like.
Is
it
just
us
manually,
fixing
it
at
some
point
or
I'm
not
sure,
and
I'm
more
curious
if
there
are
any
other
options
other
than
us,
trying
to
constantly
check
to
see
if
they've
added
a
deprecation
date
at
an
arbitrary
time.
C
One
of
the
ones
that
does
not
have
a
deprecation
date
currently
listed
actually
does
have
an
announced
deprecation
date
from
the
review
like
core
team.
So
at
the
very
least,
I
think
that
one
should
be
like
updated
to
accurately
reflect
the
announce
date
for
the
other
two,
I'm
not
sure.
I
really
think
that
it
matters
if
we
have
an
estimate
of
what
the
future
is
or
we
have
nothing,
mostly
from
the
way
that
we
actually
would
present
this
to
users
is
within
30
days
of
that
date.
C
We
then
start
presenting
a
use,
a
user
with,
like
a
log
output
that
says
this
step.
This
dependency
is
about
to
be
removed
from
the
build
packer
reach
its
end
of
support
date.
You
should,
you
know,
consider
upgrading
to
a
newer
version
so
that
you're,
not
you
know
your
application,
isn't
broken
by
the
fact
that
this
dependent
dependency
is
getting
removed
so
for
those
ones
for
the
the
two
that
are
the
version
lines
that
are
newer.
C
Those
are
at
least
a
year
out
further
than
the
one
that
currently
does
have
a
deprecation
date.
So
for
us,
if
it
has
an
estimate-
or
it
has
nothing
right
now-
is
not
necessarily
all
that
important
because
of
the
fact
that
neither
one
of
those
is
going
to
show
any
sort
of
message
for
at
least
a
year
and
a
half,
and
so
we
have
a
considerable
amount
of
time
to
figure
out
what
the
strategy
should
be
there
for
those.
C
Yes,
so
that
we
can
provide
an
accurate
estimate
to
users
as
to
when
something
is
out
of
support.
A
B
F
Api
that
will
tell
you
the
eol
dates
for
everything
if
they
exist,
and
they
just
show
up
null
until
they've
announced
one.
G
Yeah,
I
believe,
the
last
time
we
talked
about
this
and
mark
brought
it
up
in
the
slack
thread
was.
We
definitely
should
not
estimate
if
they
do
not
give
an
eol
date,
and
this
was
something
that
stephen
also
was
very
opinionated
on,
because
it's
not
our
position
to
give
users
that
contract.
G
If
we
say
if
we're
just
guessing
that
this
is
when
it's
going
to
eol
and
we're
not
really
like
in
the
position
to
provide
users
with
that
guarantee-
and
I
I
would
agree
with
that
position-
I
don't
think
we
should
do
that
either.
So
I
think,
in
the
cases
where
there's
no
deprecation
date
listed,
that
should
just
be
empty.
A
Yeah-
and
I
think
you
should
be
able
to
you-
should
be
just
pulling
against
us.
You
know
once
a
day
once
a
week,
whatever
you
want
to
see
when
these
things
come
in
and
then
updating
the
record
in
the
core
depths
area
and
then
it'll
be
up
to
like
that,
should
not
trigger
a
rebuild
of
the
binary,
that's
associated
with
it.