►
From YouTube: GDK office hour May 2020
Description
Recording from the May'2020 GDK Office Hour
A
Okay,
welcome
to
the
second
office
hour
for
the
GDK
team,
thanks
for
joining
us
and
then
thanks
for
listening
to
the
recording,
if
you
couldn't
make
it
to
our
call,
live
so
on
the
chat
window.
Let
me
just
copy
the
issue
and
you
know
for
some
reason.
This
issue
was
closed
by
somebody,
so
I
assume
that
was
a
mistake
but
I
reopened
it
and
then-
and
it's
some
couple
of
additional
points
and
the
topics
for
some
of
the
main
thing.
A
One
of
the
main
things
that
we
wanted
to
talk
about
and
Lee
I
think
you
joined
our
call
last
month.
One
of
the
things
we
need
to
talk
about
wanted
to
talk
about
was
what
do
like
get
lab
engineers,
for
example,
go
through
in
terms
of
onboarding
for
GDK
and
wanted
to
make
sure
that
there's
I
mean
documentation
was
whatever
is
available
to
the
gait
lab
engineers
is
also
available
to
the
wider
community,
and
I
was
going
through
some
documentation
pages
and
I.
There's
definitely
some
room
for
improvement.
A
You
know
not
surprised,
I
mean
documentation,
always
I
mean
it's
sort
of
working
progress,
but
I
checked
the
docs
page
and
various
like
a
readme
or
MD
files.
So
that
sort
of
the
mean
one
of
the
things
I
want
to
look
at
and
I
mean
ash.
I.
Think
I
posted
this
question
internally
about
two
engineers
about
you
know:
how
did
you
sort
of
get
started
like
using
the
GDK
and
leave
me?
Some
pointers
and
I
can
go
through
those
and
I
know
Lee.
A
B
Don't
think
so,
I'm
I'm
intrigued
to
see
if
we
did
a
live
in
Stowe
and
more
of
a
manual
one.
You
know
whether
that's
more
straightforward,
then
the
the
vagrant
attempt
so
I
I've
been
making
that
the
issue
I
seem
to
see
is
essentially
the
pace
at
which
I
understand.
Gitlab
is
progressing.
So
you
know
the
ruby
version,
changing
the
Postgres
version,
changing
the
the
gem
versions,
changing
etc,
and
the
fact
that
that
the
vagrant
image
doesn't
keep
up.
B
The
shares
I
have
to
increase
the
disk
space
because
15
gig
isn't
enough
when
you're
using
the
guest
disk
to
install
get
there
rather
than
the
host
disk,
etcetera.
So
yeah,
it's
really
last
time,
I
was
suggesting
you
know,
maybe
I'm
the
only
one
using
the
Windows
VirtualBox
vagrant
setup
and
should
I
not
be
using
that
because
you
know,
should
you
guys
stop
attempting
to
maintain
it
because
it
doesn't
currently
seem
to
you
know,
be
possible.
A
C
Yes,
thanks
again
Lina
and
we
really
appreciate
not
only
joining
the
call,
but
also
you
know
your
import
and
the
issues
and
the
conversations
we've
been
having
around
the
challenges
with
reading
GDK
and
a
vagrant
and
vagrant
also
on
windows.
So
again
just
want
to
thank
you
for
your
effort
so
far
yeah.
So
it's
a
great
question
around
just
to
pick
one
part
of
of
what
you
said
around
you
know.
Is
it
something
we
support
and
you
know?
C
Is
it
something
that
we
should
have
should
you
know,
spend
more
more
energy
on
so
yeah
I
mean
we
at
this
stage
we
don't
have
a
really
good
understanding
about
the
you
know
the
numbers
of
exactly
how
people
are
using
the
GDK,
how
they're
installing
what
platforms
that
sort
of
thing.
So,
just
for
a
bit
of
context,
we
are
putting
out
a
survey
very
soon
to
begin
with,
first
just
get
lab
and
then
we'll
open
up
to
a
wider,
wider
audience.
C
That'll
give
some
bigger
numbers
and
some
more
accurate
numbers
around
rather
around
how
people
are
using.
But
obviously
you
know
we.
We
do
have
vagrants
a
vagrant
file
and
end
documentation
around
running
vagrant
in
the
GDK,
so
yeah
I
think
I
think
at
bare
minimum.
We
should
look
to
maintain
that
at
this
stage
at
least
make
sure
it's
it's
accurate
and
up-to-date
and
works
and
as
far
as
you
know,
into
the
future,
as
far
as
you
know,
do
we
look
to
bring
it
on
a
parity,
ongoing
I.
C
Think
that's
another
question
and-
and
you
know
with
regards
to
running
on
Windows,
you
know
in
an
ideal
world,
we'd
love
to
be
able
to
install
the
JDK
as
smoothly
as
possible
on
all
platforms.
That
way
everyone
can
contribute
and
there's
no,
you
know,
there's
no
sort
of
backward
way.
You
need
to
run
something
on
your
platform
of
choice,
just
to
be
able
to
get
in
there
and
do
what
you
need
to
do
so
so
I
guess
what
I'm
saying
I'm
gonna
roundabout
ways.
C
Yes,
we
need
to
put
some
effort
in
to
get
vagrant
working
as
it
is
right
now
and
that
I
think
that
covers
what
you
suggest
it
around.
You
know
making
sure
that
the
Bunty
version
kind
of
more
closely
matches.
You
know,
like
the
the
the
platforms
that
we
target
with
omnibus,
ensuring
that
the
Postgres
version
matches
the
version
that
we're
recommending
and
with
the
GDK.
We
also
try
to
stay
on
the
sort
of
woods
on
the
edge
more
closely
to
the
the
newer
versions
of
things.
C
So,
for
example,
like
Postgres
12
is
out
post
11
as
well,
and
you
know
we
look
to
to
offer
those
as
options
as
far
as
you
know
that
by
default,
we'll
probably
probably
have
Postgres
10
installed.
But
you
know,
because
it's
a
development
kit,
the
it
should
be
possible
without
too
much
difficulty
to
be
able
to
select
eleven
and
twelve
just
to
make
sure
that
what
you're
doing
is
actually
going
to
be.
C
You
know,
sort
of
workable
for
the
future
yeah,
so
I
think
we
definitely
put
some
effort
into
vagrant
and
I'm
very
apologized
profusely
that
I
haven't
spent
much
time
on
the
issue
that
we've
been
conversing
on.
I
haven't
been
yet
having
having
much
luck,
getting
it
to
work
on
my
machine,
but
it's
something
something
else,
but
the
fact
that
I
can't
even
get
anywhere
near
closer
to
where
you're
sort
of
getting
up
to
suggest
that
you
know
we
have
some
work
to
do
so.
C
Definitely
you
know
I'll
commit
to
to
spending
some
more
time
on
vagrant,
but
just
again
for
some
context.
I
started
working
full-time
on
GDK
about
two
weeks
ago
and
prior
to
that
it
was
the
maintainer
x'
sort
of
just
doing
what
they
can
in
between
their
other
other
team
work.
So
so
yeah
I'm
working
on
a
full-time
now
we're
currently
getting
a
lot
of
things
in
order
in
terms
of
CI
updating
documentation,
making
things
a
little
more
tightly
before
we
start
to
launch
into.
C
B
A
I
mean
I
just
jot
it
down
like
a
three
bullet
points
on
the
issue,
I
mean,
if
you
don't,
if
you
don't
mind
like
adding
the
like
existing
issue
that
you
just
mentioned,
it
would
be
great.
Let
me
paste
it
on
the
chat
here
so
go
through.
You
know
what
I
just
typed
in
the
you
should
put
the
office
hour.
C
A
C
Appreciate
that
so
I
think
that
just
the
reason
why
we're
not
initially
kind
of
setting
out
to
everyone
is
I
think
in
the
beginning,
I'd
like
to
get
a
smaller,
just
a
bit
of
feedback
on.
Is
it
actually
asking
the
questions
that
we
should
be
asking
and
then
just
try
and
find
tune
it
a
bit
and
then
roll
it
out
to
everyone
cool.
A
B
No
I
think
I
think
that's
that's
nowaday
again,
as
we
we
mentioned
last
time,
I
think
I'm
open
to
doing
it
a
completely
different
way.
If
it's
I
would
actually
I
run
a
Linux
VM
with
a
GUI,
so
that
I
actually
do
my
development
inside
that
the
yeah.
Well
I
already
do
my
development
inside
the
VM,
because
the
the
host
sharing
of
that
the
files
and
folders
doesn't
work,
which
means
I'm
using
a
text
editor
to
do
all
my
development.
C
Yeah
and
those
issues
around
you
know
things
changing
and
needing
to
be
kind
of
nuked
that
we
see
there
quite
often
even
on
the
native
side
of
things.
So,
for
example,
when
it
you
know
on
a
lot
of
the
gear
lab
engineers
run
Mac,
and
so
when
there's
a
Mac
update,
which
is
great,
you
know
if
your
security
upgrades
law,
libraries
and
then
a
whole
bunch
of
ruby
gems
that
have
C
extensions
which
suddenly
stop
working.
So
we
yeah.
So
it's
something
that's
not
immune
to
just
that.
C
They
even
the
native
install
which
is
there
I,
would
say
right
now,
the
kind
of
preferred,
or
at
least
the
most
common
way
people
are
installed
in
the
JDK,
is
by
installing
locally
other
machines
with
by
homebrew
or
the
you
know.
The
package
managers
of
those
native
platforms,
but
yeah
we've
got
a
few
things.
I'll
add
it
to
the
issue.
There's
a
we're.
C
Looking
at
adding
a
sub
command
called
pristine,
so
in
the
Ruby
world,
there's
a
pristine
command
for
gem,
which
essentially
reinstalls
the
gem,
including
NEC
extensions
that
it
may
have,
and
in
the
pristine
sub
command
for
GDK
will
attempt
to
do
the
same
thing
by
returning
your
GDK
to
a
pristine
state
which
will
repair
any
broken
things.
So
that's
something
that
will
help
when
things
get
really
gnarly.
But
it's.
A
There
you
go.
Hopefully
people
can
see
it
yeah
a
few
windows
I'm
sharing,
so
so
what
I
did
was
I
went
to
doc,
Stuckey,
lab
calm
and
just
search
for
GD
k
and
I
mean
ash
I,
don't
know
if
you
looked
at
like
GD
k
like
documentation
files
on
on
the
docs
page
recently
and
as
far
as
I
can
tell
that
there
doesn't
seem
to
be
a
home
for
GD
k.
A
C
A
A
What
I
thought
was
interesting
when
I
talked
to
some
of
the
engineers
I?
Could
you
know
that
a
joined
gitlab
relatively
recently
was
the
it
seems
like
the
important
page
was
like
the
third
one
setting
up
the
GDK
versus,
like
even
I,
think
like
I,
think
cupping?
Wasn't
that
big
of
a
deal
or
even
using
or
even
like
a
how-to
page,
but
this
is
sort
of
the
page
that
they
talked
about.
Being
this
being
like
pretty
helpful,
that's
just
one
of
the
first
page.
A
They
people
mention
I
mean
Lee
I,
don't
know
if
you
like
it
regularly
like
I.
Look
at
the
page
whenever
you
have
problems
with
the
GDK
to
see
if
there
are
any
updates,
but
you
know
I
guess
the
point
I'm
getting
to
is
from
the
navigation
I
mean
I
saw
like
you,
especially
from
from
like
a
I
mean
I
mean
I'm
I'm,
a
pretty
non-technical
person.
B
The
interesting
actually
I
must
admit:
I,
don't
I,
don't
revisit
it,
probably
often
enough,
so
it's
possible
that
it
does
get
changed
and
updated
and
would
help
to
prevent
me
hitting
some
of
the
obstacles
that
I
do
here,
especially
for
example,
when
there
was
the
recent
change
to
the
way
that
you
modify
settings
so
with
the
implementation.
I
think
it's
of
the
GDK
dot
yeah
Mille
when
that
happened.
B
I
guess
maybe
that
is
over
not
an
alarm
bell,
but
a
sort
of
suggestion
that
I
might
have
will
should
all
contributors
developers
have
somehow
be
on
a
newsletter
or
get
some
kind
of
notifications
when
something
is
updated
or
something
major
changes.
I
know
this
is
completely
off
topic.
Apologies,
but
when
masters
broken,
for
example,
the
the
I
think
it
was
Stan.
Who
I
believe
is
that
that's.
C
B
We're
not
responded
to
it
a
couple
of
questions
in
the
contributors
channel,
which
was
gray
and
sort
of
confirmed
our
suspicions
that
master
was
broken,
but
you
know:
did
your
internal
engineers
get
a
notification
so
that
they
don't
start
going
crazy
down
rabbit
holes,
tryna,
diagnose
problems
with
their
their
pipelines
with
their
builds,
etc?
When
actually
our
masters
broken?
So
you
know
we
just
need
to
be
patient,
wait
for
someone
to
fix
Master,
rebase
and
then
continue.
C
C
Any
number
of
those
can
be,
you
know
potentially
non-functional
and
cause
the
overall
like
get
lab
I'm
instance
to
not
function
as
correctly
as
correct
as
it
should.
But
going
back
to
your
point
around,
you
know:
how
do
we
communicate
these
things?
When,
because
we,
you
know,
we
have
status,
get
lab
comm,
for
you
know
for
people
wishing
to
understand
the
production
side
of
things
like.
When
is
the
production
get
lab,
calm
suffering
some
sort
of
degradation
or
hiccup,
but
how
do
we
communicate
when
the
development
side
of
things
is?
Currently?
C
You
know
it's
not
possible
for
a
developer
to
do
their
job.
So
it's
a
it's
an
excellent
question.
I
have
raised
an
issue
for
it
a
little
while
back
we
I
suggested
at
the
time
to
potentially
have
you
know,
pre
and
post
sort
of
update
messages
when
we
do
a
GDK
update,
but
that's
not
going
to
cover
just
paste
it
in
the
chat
it
just
for
long,
but
that
doesn't
cover
kind
of.
But
that's
more,
you
know,
heads-up
we've
made
a
minor
change.
C
You
may
need
to
do
something
here
or
there,
but
for
your
your
sort
of
point
around
you
know,
there's
a
transient
issue.
We
you
know
we're
actively
working
on
it.
How
do
we
communicate
that
fact?
So
it's
a
great
question
newsletter,
Google,
Group,
Gitter,
I,
think
they're.
All
you
know
if
we
can
have
one
way
to
create
the
message
and
have
it
sent
to
all
those
platforms,
I'd
be
more
than
happy
to
to
maintain
or
manage
that
so
I
think
it's.
A
If
you
can
just
like
just
post
it
even
on
like
a
gator
channel
I
think
that
would
be
helpful
because
I
think
Li
and
like
other
active
contributors
are
definitely
there
and
that's
why
it's.
You
probably
seen
me
post
questions
on
on
our
GDK
slack
channel
anyway.
When
is
whenever
I
see
him
so
I
think
that
would
be
pretty
effective.
I
mean
you
can
do
it
straight.
I
mean
just
you
know
straight
on
getter,
or
you
can
just
utilize
that
channel
that
I
created
on
slack.
So
you
can
just
post
it.
A
There
they'll
be
broadcast
over
over
together,
I
mean,
unlike
adjudicate
things,
I
think
I
mean
we
have
a
pretty
small
team.
That
might
be
a
pretty
easy
way
of
doing
it
as
a
starter
and
and
see
if
that
works,
and
then
I
mean
we
can
kind
of
calibrate
like
what
needs
to
be
broadcasted
and
what
doesn't
but
I
mean
LeAnn
I
mean.
Let
me
know
if
you
guys
disagree
or
I
mean
or
you
you,
you
think,
that's
a
reasonable
place
to
start.
B
Stopping
that
sounds
like
a
good
starting
point.
I
I
kind
of
wondered
whether
some
of
the
integrated
GDK
commands
like
the
GDK
are
GDK
update,
might
even
be
able
to
spit
out
something
but
again
I.
Guess
it's
it's
somebody
having
to
maintain
that
and
and
that-
and
that
is
probably
the
challenge,
as
ash
mentioned
there,
that
the
kind
of
dependency
chain
that
GDK
depends
on
all
of
those
products
work
horse
and
get
lab
itself
and
various
other
components.
It's
is
what's
changed
with
those
I
the
reason.
B
C
There
would
have
been,
or
there
there
was
I,
don't
know
the
specific
link,
but
in
the
in
the
release,
so
like
the
gitlab,
CA
and
a
releases
would
have
been
noted
as
a
change
in
the
like
a
deprecation,
so
that
does
highlight
a
gap
we
have
in,
and
this
is
one
of
many
gaps
we
have
between
the
version
of
gitlab
that
you're
developing
on
this
is
the
version
of
get
lab
comm
or
the
version
of
get
lab.
You
may
have
on
premises
that
you
know
they
are.
They
are
quite
separate
projects,
so
we
do
see.
C
You
know
that
a
lot
of
issues
are
a
lot
of
mode
requests
being
made
in
get
lab
code
base
and
then
kind
of
needing
to
be
recreated
in
the
GDK
code
base
to
you
know
fix
documentation,
so
it
is
a
bigger
problem.
Then
I
guess
just
say
that
the
Postgres
9.5
it's
even
another,
little
minor
thing,
like
the
name
of
the
services
don't
match
in
the
GDK
version,
to
say
what's
an
omnibus,
which
is
the
kind
of
you
know,
production
most
common
production
method
for
deployment.
C
C
You
know
the
actual
grid
layer
project.
You
know
things
like
that.
We
need
to
start
to
think
about
how
we
can
more
closely
bridge
those
gaps,
but
yeah
they're,
definitely
gap,
something.
Sorry,
I
don't
have
a
more
concrete
answer,
but
definitely
yeah,
we've
I
think
anything.
We
can
do
to
spread
any
issues
further
than
just
that.
You
kind
of
slack
channels
that
we
have
a
good
lab
would
be
great.
B
Now
I'm
sure
it'll
be
a
team
effort.
I
guess
the
other
thing.
Maybe
I
should
be
a
bit
more
disciplined
raising
issues.
You
know
the
the
Postgres
issue
that
I
had
and
again
I
think.
Maybe
this
is
what
all
the
developers
there
and
maybe
why
we're
in
part
of
this
situation,
because
I
tried
to
do
a
ser
GDK
up,
they
I've
got
an
issue
with
one
of
them.
My
great
database
migration
scripts.
It
was
Trinitron
googled.
That
issue
found
a
Stack
Overflow
post
that
suggested.
B
That
was
because
functionality
that
doesn't
exist
in
Postgres
nine
point,
five
and
and
then
sort
of
I
think
I
tried
to
spin
out
my
new
VM
and
that
still
had
exactly
the
same
problem.
So
I
then
thought!
Well.
Maybe
it's
because
I've
been
to
16
is
just
so
old.
It
doesn't
all
that
might
be
to
do
the
golang
version
not
being
on
there
or
whatever
it
might
have
been
and
then
I
just
kind
of
start,
plugging
away
until
I
get
it
working.
B
And
whilst
that
have
been
spamming
that
post
asha
bahut
trying
to
get
GD
k,
you
know
working
and
and
kind
of
exactly
the
steps
that
required,
which
hopefully
will
be
helpful
for
someone
or
we
can
use
to
try
and
dig
out
all
of
those
individual
issues.
I
wonder
whether
it
would
be
right.
I
know
it
would
be
a
lot
more
beneficial
if
every
time
I
hit
a
problem,
I
raised
an
issue
and
even
though
I
probably
solve
it
I
then
post
the
solution
as
well.
B
B
Do
we
need
to
whatever
it
might
be,
but
having
much
more
of
a
kind
of
catalog
so
again,
slack's
great
for
you
guys
get
is
great
for
us,
but
only
while
the
informations
kind
of
at
the
top
I
think
is
it
especially
with
the
threads
that
appear.
Now
it's
going
to
be
very
hard
to
search
yeah,
so
I,
don't
know
what
your
thoughts
are
on.
That.
C
Yeah
well,
I
think
you
know
having
the
ability
to
I
mean
I
would
like
to
sit
in
the
the
get
a
challenge,
even
throw
the
terminology,
Channel
I,
suppose
and
be
more
active.
It's
just
it's
just
a
bit
full-on
at
the
moment,
trying
to
maintain
I
guess
making
making
progress
on
improving
the
GDK
and
also
supporting
the
GDK
channel
that
we
have
internally,
so
yeah
I
think
to
begin
with
having
a
place
the
ability
to,
even
if
it's
manual,
just
you
know,
post
the
updates
or
use
the
tunnel.
A
Think
I
I
think
I
pinged
you
on
the
like
a
channel,
but
it's
no,
no
biggie.
If
you
go
to
the
go
to
that,
let
me
if
you
go
to
slack
right
now,
look
for
git
or
contributor
room,
then
you
should
be
able
to
see
messages
that
are
coming
across
and
then
the
message
you
post
Airwolf
will
get
broadcast
together
as
well.
A
Yeah
I
mean
that's
a
very
good
question.
I
mean
this
is
something
that
we
I
mean
I.
Think
in
especially
the
community
relations
team
that
we
talked
about.
We
even
thought
about
like
having
like
a
like
instances
like
one
internal
forget,
lab
and
just
a
public
instant,
that's
open
to
like
everybody,
or
that
goes
long
as
you
like
sort
of
register
and
then,
but
we
haven't
I,
think
it's
been
a
while,
since
we
had
that
discussion.
So
this
this
sort
of
you
know
creating
our
bridge
is
sort
of
a
interim
solution
for
it.
A
B
I
know
when
we
looked
at
it,
although
I
it's
the
free
version,
there
were
some
limitations
that
made
us
kind
of
consider
whether
we
wanted
to
go
for
the
pay-for
version,
but
with
the
remote
working
we
found
that
yeah.
It
works
using
zoom
for
calls,
or
and
or
Skype
and
various
other
tools
that
we've
got
but
they're.
B
What
I
think
they
bill
is
passive
voice
chat.
Feature
of
discord
is
brilliant
for
kind
of
recreating
our
our
office
environment
as
we
call
it.
So
it's
very
easy
for
me
to
kind
of
just
shout
and
if
anyone's
around
they
can
respond
or
or
not,
rather
than
me,
having
to
very
deliberately
ring
them,
but
yeah
that
that
tall
seems
to
be
working
quite
well
for
us
and
it
mimics
lack
in
a
lot
of
ways
from
what
I
understand.
I
don't
know
if
I've
you
guys
familiar
with
that
yeah.
A
B
A
Okay,
sorry
I
mean
the
other
tool
that
I
just
thought
of
we
have
is
like
a
formula.
Calm
and
I
mean
that's
sort
of
an
under
underutilized
tool,
but
this
allows
I
mean
I,
don't
know
if
you
ever
participated
in
our
forums
or
been
to
our
forum.
Caleb
calm
I
mean
that
this
is
where
it's
it's
easily
more
easily
searchable,
and
then
you
can
have
more
like
sort
of
discussions
right
rather
than
trying
to
deal
with
threads
in
like
slack
or
or
getter,
and
we.
A
B
Thing
but
again
it
seems
to
me
I
guess
maybe
this
laid
out
in
a
slightly
nicer
fashion,
but
it's
it's
again.
Sometimes
a
negative,
almost
stead
of
having
too
many
tools
or
I
assumed
I
think
I
was
getting
the
feeling
that
the
forum
might
have
been
deprecated.
It
was,
it
was
actually
legacy,
and
now
that
you
know
you
use
issues
when
when
you've
got
a
question
or
you
want
to
start
a
threaded
discussion,
you
create
an
issue
so
again.
B
Maybe
that
is
something
that
doesn't
exist
or
or
does
existing
could
be
improved
or
something
that
I've
not
seen.
That
sort
of
really
helps
to
highlight
and
say,
hey
guys.
This
is
get
a
contributor
chat
channels
great.
If
you
want
to
try
and
get
an
immediate
response,
you
know,
but
often
there
may
not
be
someone
around
issues
can
be
used
for
this.
The
forum
can
be
used
for
this
slack's
used
internally
for
this
etc.
A
Yeah,
so
yeah
I
mean
I,
do
like
to
throw
it
out
there.
We
actually,
this
is
like
a
discourse
based
forum,
is
something
that
we're
trying
to
do.
I
mean
we
actually
have
people
like
starting
to
work
on
it
again,
so
that
it
doesn't
so
it
gets
revised
some
it
gets
revived
and
expect
us
to
sort
of
sort
of
pick
up.
So
it's
another
option,
but
I
mean
as
started
like
I
mean.
Maybe
I
mean
we
just
you
know
we
start
with
trying
to
see.
A
If
you
know
starting
to
use
get
er
I
mean
it's
sort
of
a
simple
tool
that
we
can
use
with
the
integration
with
slack
and
see
how
that
works,
and
if
it
doesn't,
then
we
can
certainly
explore
other
options.
I
mean
I,
wouldn't
necessarily
rule
other
tools
out
I
mean
I'm,
not
that
familiar
with
this
court
I
mean
I,
should
I'm,
not
sure
if
you
are
but.
A
B
It
was
just
the
fortunate
in
a
way
for
us
taking
enterprises,
that
is
that
we
we
weren't
using
I,
suppose
we've
got
skype
and
teams
and
probably
various
other
tools,
but
Microsoft
Skype
is
notoriously
flaky
teams.
I
can't
remember,
we
haven't
really
started
to
use
a
bit.
They've
been
trying
to
push
people
on
to
it
and
I
think
we've
got
a
couple
of
gamers
on
our
team
and
they
use
discord
for
from
a
kind
of
gaming
point
of
view,
and
they
were
the
ones
that
recommended
that.
A
B
Is
there
any?
Is
there
any
highlight
ash
I
think
you
mentioned
that
the
deprecation
of
Postgres,
nine
and
a
half
would
have
been
included
in
release
notes
for
the
gitlab
core
product
to
wonder.
If
there's
any
way
of
flagging
certain
release
notes
in
the
change
logs
or,
however,
you
would
do
it,
because
what
you
really
want
is
a
culmination
something
that
what's
the
word
I'm
looking
for
that
didst.
B
Of
no
no
definitely
that's
that's
definitely
condensed,
but
there's
something
I'm
looking
for
that
almost
says.
Okay,
let's
take
you
can
flag,
let's
say
a
changelog
there's
a
major
change
or
a
dependency
change,
or
something
like
that,
and
then
there
would
be
something
that
collects
all
of
those
changes
from.
You
know
workhorse
from
yet
lab
from
GDK
and
so
yeah
somewhere,
where
you
can
see
unified
view
of
yes,
it's
just
another
idea.
B
It
doesn't
immediately
strike
me
as
as
brilliant,
but
I
think
it's
again
because
I'm
thinking,
oh
I'm,
using
GDK
I'm
looking
at
gbk
I'm
interested
in
GDK.
But,
as
you
mentioned
earlier,
all
the
dependencies
are
all
the
core
products
in.
If
I'm,
not
keeping
on
top
of
what's
happening
with
all
of
those
core
products,
then
I'm
potentially
missing
out
on,
but
then
again
there's
a
question.
B
C
There
are
all
great
points
yeah.
This
has
a
lot
they're,
just
something
that
just
picked
just
popped
into
my
mind.
So
this,
the
general
I
guess,
structure
of
how
GDK
works
is
that
gitlab
the
projected
lab
master.
It
drives
the
dependency
structure.
So
when
you
do
a
GDK,
install
or
update
it
pulls
down
master
of
gitlab
and
then
in
inside
the
master
repository
there
are
a
bunch
of
to
do.
There
are
files
that
tell
GDK
and
as
well
as
omnibus,
when
those
packages
are
created,
which
versions
of
giddily
shall
work
horse
to
pull
down.
C
So
that's
the
kind
of
driving
point
and
that's
that's
the
way.
The
get
lab
needs
to
work
get
lab
engineers
in
that
they
we
all
work
out
of
master
and
that's
sort
of
hard-coded
into
the
way
the
current
GDK
works,
but
and
that's
kind
of
the
source
of
where
a
lot
of
these
incompatibilities-
and
you
know,
breakages
around
dependencies,
come
in
because
master
moves
so
quickly.
But
for
the
usage
pattern
for
all
the
workflow,
I
should
say
for
other
other
people
other
such
as
yourself.
You
may
not
need
to
work
out
of
master.
C
It
may
be
sufficient
for
you
to
be
developing
against
the
you
know
the
most
current
stable
tag,
and
so
therefore
that
would
move
far
less
quickly
and
introduce
far
fewer
dependency
breakages
and
would
allow
I
guess
more
closely,
aligning
your
like
the
development
against
sort
of
production
releases.
Still
with
the
you
know,
there's
work
to
be
done
around
and
communicating
that
you
know
deprecation
is
coming
in
and
say
you
know,
thirteen
that
needs
to
be
reflected
somewhere.
B
B
And
you
know
all
the
the
reviewers
maintain
us
whatever
we
want
to
call
them
the
guys
that
help
me
out
with
my
mrs,
be
regularly
kind
of
say:
oh,
can
you
can
you
rebase
from
master,
so
I
think
that
it
probably
is
quite
important
that
you
know
to
make
sure
that
I'm
not
changing
something?
That's
already
been
changed
or
yeah
they're
doing
anything
along
those
lines.
That's
I
think
it.
B
It
could
be
a
very
good
option
if
something's
broken
and
I
need
to
get
up
and
running
to
say:
oh,
is
there
a
command
I
can
use
to
do
a
GDK
update
and
tell
it
to
use
a
certain
branch
or
tag
that
I
think
it's
a
brilliant
idea,
just
to
kind
of
say
that
I
really
need
to
keep
working
and
and
can
I
do
it
that
way,
but
for
the
kind
of
standard
development
cycle
processor
think
I
probably
do
need
to
be
a
master.
So.
A
C
Great
question
so
the
just
to
go
back
to
the
question
of
postally,
we're
not
saying
using
master.
Sorry,
if
I
wasn't
clear
that
what
I
meant
to
say
was
a
gitlab
master.
So
definitely
at
this
stage
we
don't
have
a
concept
in
the
GDK
project.
We
don't
have
a
concept
of
tags
and
and
releases
of
GDK
that
may
come
in
the
future.
But
definitely
master
is
the
only
place
right
now
that
you
can
work
out
of
in
the
in
the
get
lab
development
kit
project.
C
What
I
was
meaning
is
the
the
get
lab
master
branch
of
the
main
project
is
kind
of
what
drives
all
the
dependencies
and
updates.
So
when
you
do
leading
to
your
question
array
around
the
reason
why
you
would
do
a
GDK
update
is
for
one
most
of
the
time,
it's
because
you're
working
on
a
feature
and
you
need
to
keep
up
with
all
the
active
changes
that
are
going
on
in
get
lab
master.
It's
not
so
much
the
GDK
master,
although
there
are
changes
that
come
in
when
you
run
a
GDK
update.
C
The
first
thing
it
does
is
pull
down
the
GDK
master
and
then
from
there
that
allows
you
know,
fixes
and
improvements
around
various
things
and
then
the
second
thing
it
does
is
pull
down,
gitlab
master
and
from
there
that's
where
it
determines.
You
know
newer
versions
of
gems
and
things
like
that.
So
if,
if
that
flow
and
I
hope
that
answers
your
question
array
now,
if
that
flow,
that
that's
currently
the
kind
of
standard
flow
and
that's
the
flow
most
people,
so
most
engineers
at
get
lab
kind
of
follow.
C
But
if
the
flow
for
say
yourself,
Lee
or
others
isn't
necessarily
you
know
required
to
use,
get
lab
master,
then
that
flow
could
be
substantially
kind
of
hardened
by
by
using
the
current,
stable
or
at
least
saying
you
know,
I
wanted
to
develop
against
the
bleeding
edge,
which
is
master
or
the
current
stable
or
the
previous.
You
know
stable
release
that
kind
of
thing
that
would
protect
from
some
of
the
the
kind
of
transient
type
of
dependency
breakages
and
that
type
of
stuff
that
occurs
from
working
directly
out
of
master.
That's
a
possibility.
B
C
Yeah,
it's
definitely
preferable
to
stay
as
close
to
master
as
possible,
because
you
have
all
the
mentioned
all
the
reasons
you've
just
mentioned,
however,
currently
sort
of
comes
at
the
cost
of
of
potentially
writing
out.
You
know
the
waves
of
you
know
breakages
that
we
also
as
a
company
experience
when,
when
masters
broken,
I
think
collectively
it
just
requires.
C
You
know
this
sort
of
ongoing
effort
to
make
sure
our
pipelines
are,
you
know,
have
got
the
best
coverage
possible
and
that
we,
you
know,
continue
to
review,
but
it
will
always
happen,
but
I
think
the
problems
we're
talking
about
here
and
not
so
much.
You
know
the
transient
type
of
problems
with
master.
B
B
Obviously
the
numbers
for
everything
else
would
be
really
interesting
to,
but
I
tend
to
believe
that
I
may
be
one
of
a
few,
which
is
why
I
sort
of
suggest
well.
Is
it
worth
dropping
support
and
stopping
trying
to
maintain
this,
because
you
know
just
for
a
select
two
three,
four
five
developers:
it's
requiring
one
full
member
of
a
team
to
actually
maintain
the
the
vagrant
VirtualBox
image
and
that
just
doesn't
kind
of
make
sense.
B
But
if
there
are
20
very
a
hundred
people
using
the
windows,
vagrant
VirtualBox
setup,
then
surely
we
must
all
be
having
similar
issues
and
fixing
them
ourselves.
We
need
to
take
the
time
to
share
those
fixes
to
raise
merge
request
in
GDK
I
think
maybe
that
is
kind
of
another
problem
that
we're
kind
of
in
our
little
bubble,
and
we,
like
oh
I've,
seen
this
before
I
know
how
to
fix
it.
I'll
fix
it
for
me,
not
okay.
Well,
this
is
a
problem.
C
A
So
I
mean
just
one
idea
that
I
think
Lee
you
brought
up
a
few
minutes
ago.
I
mean
I.
Think
you
use
a
term
like
a
change
log
like
in
the
readme
file.
Should
there
be
a
link
to
or
table
of
like
here's
like
the
latest
components
that
are
that
are
part
of
the
GDK,
for
example,
whatever
Postgres
version
and
whatever,
like
you
know
here
at
the
latest
version
that
I
support
it
like?
A
B
B
So
yeah
it's
even
that's
interesting
because
it's
sort
of
like
well
surely
alarm
bell
should
ring
somewhere.
When
somebody
changes
anything
on
this
page,
they
should
have
either
previously
and
again
this.
This
isn't
a
blame
game,
as
you
said,
ashes
everyone's
paid
to
do
their
own
different
things,
but
if
somebody's
updating
this
document
here
and
saying
well,
no
js'
13
is
required.
Now,
then,
presumably
somebody
then
needs
to
go
and
look
at
GDK
and
actually
or
vagrant.
B
C
C
What
is
the
version
of
Postgres
that
we
we
want
to
actively
by
default,
install
and
kind
of
work
with,
and
then
what
are
the
versions
that
we,
you
know
have
facility
to
to
be
able
to
select
between,
and
we
don't
have
a
good
way
to
kind
of
have
that
cross-platform.
So
right
now
it's
baked
in
the
Ruby
code
and
that's
to
you
know
to
make
that
change
effective
in
the
vagrant
file.
C
For
example,
you
know
we'd
need
to
make
a
separate
change,
so
there
is
something
again
if
I
created
an
issue
but
I'm
abstract
about
the
services
and
the
software
that
the
GDK
requires
into
something
that's
centralized
and
then
I.
Think
I've
mentioned
something
like
the
GDK
core,
like
having
a
core
sort
of
a
core
part
that
can
be
utilized
by
both
vagrant
and
native,
and
then
you
know,
and
just
for
heads
up
in
the
future.
We
are
going
to
start
offering
some
of
those
data.
Oriented
services
like
Postgres
Redis,
mean
I,
owe
it
cetera.
C
We
definitely
don't
have
a
way
to
kind
of
connect,
all
the
dots
between
vagrant
and
other
ways
to
do
things
so
that
will
improve
but
yeah
it's
a
bit
of
a
YUM
and
at
the
moment
the
focus
is
on
the
GDK
native
side
of
things
with,
but
yeah
I
will
commit
to
improving
the
vagrant
stuff.
So
at
least
you
know
matches
right
now
with
with
what
we're
saying
as
far
as
versions
and
things
so.
A
Yeah
we're
few
minutes
over,
but
I
mean
on
the
dark
side
of
things,
I
think
what
I'll
do,
if
you
don't
mind,
I,
might
just
open
an
mr,
like
even
on
the
main,
maybe
read
me
page
suggest
some
improvements
that
that
I
mean
that
we
sort
of
discussed
here
in
the
last
50
minutes
or
so
and
then
I
mean
Lee
I.
Think
you
raise
a
couple
of
good
questions
about.
You
know
how
old,
like
updates
on
the
important
things,
be
communicated.
A
I
think
that's
definitely
worth
looking
into
yeah
and
then
yeah
I
didn't
even
think
about
this.
One
master
is
broken,
I
mean
this.
This
obviously
doesn't
get
communicated
to.
Like
I
mean
there
was
a
Twitter
handle
for
status,
but
maybe
we
need
to
figure
out
a
way
to
sort
of
communicate
that
somehow,
because
it
obviously
because
I
mean
I've
noticed
that
the
over
the
past
couple
there's
is
something
flaky
about
master
and
then
your
pipeline
fails
with
you
know,
and
it
wasn't
due
to
any
changes
that
people
have
been
making
from
the
wider
community
right.
A
B
C
B
C
Shall
get
you
the
link
yeah,
so
the
master
broken
I'm
not
from
here
with
the
master
needs
investigation,
but
typically
when
master
is
detected
as
being
broken,
that
obviously
impacts
everyone.
So
an
issue
was
created
with
that
tag
and
the
tag
then
I
think
that
is
there's
something
there's
some
sort
of
automation
around.
That
being,
then
any
issues
that
are
created
with
that
tag
then
get
posted
back
into
slack
I.
Think
it's
the
development
channel,
but
I'll
just
find
the
okay.
B
Okay,
because
similar
to
what
I
was
saying
about
the
the
readme
for
the
kind
of
dependencies
I'm
not
aware
of
any
real
kind
of
notifications,
or
she
call
it
your
subscriptions
that
you
can
do
within
the
get
lab
product
itself
to
sort
of
say.
Oh,
let
me
know
if
an
issue
gets
raised
with
with
this
tag,
or
let
me
know
if
this
file
gets
changed
or.