►
From YouTube: Test/Release Working Group Meeting 6/4/2018
Description
Agenda and meeting notes:
https://goo.gl/6G4dwx
A
A
So
once
again,
there
is
no
plan
for
point
8
1,
unless
we
find
some
critical,
critical
bugs
in
point
eight
zero
that
we
need
to
fix.
Hey
I
created
that
tracker,
a
github
is
record
just
to
collect
information
about
all
these
potential
issues
or
potential
quitting
critical,
bugs
and
based
on
this
information.
We
evaluate
the
need
for
creating
dispatch
release
and
order.
Communication
will
happen
in
the
checker.
B
B
B
And
so
we're
using
so
first
the
tight
configuration,
so
you
can
define
basically
queries
to
beat
up
on
how
you
wanna,
you
know,
merge
what
kind
of
labels
you
want
to
expect
on
your
PR
so
that
in
can
be
merged
and
what
kind
of
PR
you
should
be
missing
and
and
then
you
can
define
the
kind
of
method
for
merging.
So
in
all
the
ways
your
project
will
use
flash.
So
each
time
we
add
a
new
report
and
we
have
to
make
basically
add
the
squash
method
for
all
those
things.
B
So
we
I
created
basically
that
plug-in
that
lets,
use
that
and
so
that
I
guess
I
want
to
talk
about
that
option
too
much,
and
so
the
second
part
of
the
top
of
the
tight
integration
is
that
basically,
each
time
we
are
going
to
add,
you
required
check,
meaning
something
new
from
circle
CI.
We
we
want
to
add
them
actually
in
the
pro
configuration
and
pro
will
push
it
to
the
top
and
the
reason
we
want
to
do.
B
That
is
because
we
want
to
be
able
to
audit
the
change,
because
we
want
to
know
why
the
test
was
added,
you
know
and
why
it
was
removal
as
well
and
and
currently
the
little
API
doesn't
allow
us
to
edit
to
audit
that.
So
that's
the
reason.
So
don't
do
social
truth
is
actually
the
pro
config
and
as
soon
as
you
change
it,
and
you
commit
your
PR,
it
will
be
pushed
to
the
prowl
and
type
will
read
it.
You
cannot
just
push
to
a
repo
to
change
it.
B
You
have
to
create
a
PR,
but
that's
only
and-
and
here
one
of
the
question
people
might
ask-
is
why,
where
are
all
the
products
listed?
So
because
this
is
the
pro
configuration?
It
actually
inferred
the
required
check
from
the
product,
and
you
can
see
some
of
the
drugs
have
the
optional
optional,
optional,
yeah
optional,
equal
true,
and
that
means
that
that
job
is
not
required
to
merge.
So
if
and
also
you
can.
B
There
is
a
Skip
branch
option
in
every
presumed
configuration,
so
if
there
is
a
job
that
doesn't
actually
apply
to
the
branch
that
you're
working
on
you
can
just
do
a
skip
branch
and
and
the
branch
protection
will
not
consider
it
for
for
merge,
but
that's
basically
it
so
right
now,
when
the
change
is
pushed
to
the
branch
protection
configuration,
it
will
update
it
up
once
a
day.
It's
not
gonna,
get
a
big
hit
around
on
npr's
I.
Think
in
the
future.
B
We'll
we'll
have
something
like
this
that
can
actually
look
for
web,
who
can
and
directly
actually
push
on
Meetup
right
now.
It's
not
like
that.
It's
a
cron
job
that
basically
runs
once
a
day,
so
people
can
still
change
the
thing
on
and
like
admins,
only
can
change
on
github
and
make
the
DPR
push
but
fully.
This
is
a
no-no.
B
B
B
And
you
see
all
the
things
aren't
actually
not
being
merged
because
they
have
a
failing
test.
So
if
I
click
on
any
random
one
I
should
see
that
required
tests
are
not
being
passing
so
yeah
this
time.
This
test
is
that
is
required
and
is
not
passing.
So
that's
why
this
one
is
passing,
but
it's
probably
not
being
merged
for.
B
A
B
So
if
people
change
on
github
but
don't
change,
prop
protests
have
doesn't
have
any
ID
in
what
tests
are
required,
so
it
will
still
wait
for
the
tests
that
are
failing
to
be
passing,
even
though
it's
inconsistent
with
github.
So
that's
why
I
like
the
way
to
do
that
is
actually
to
push
it
directly
in
Pro
and
probable
reach
right
away.
B
Even
though
it's
not
updating
Geetha,
it
will
still
use
that
data
right
away
and
and
then
like
once
a
day
out
push
it
to
beat
up
and
then
the
UI
it's
going
to
be
consistent,
but
hopefully
we
should
be
changing
required
check.
Like
you
know,
once
a
week
or
something
or
even
once
a
month
it
shouldn't
be
changing.
We
should
be
fixing
the
test
as
opposed
to
disable
them.
I.
Think.
B
A
A
A
A
D
E
D
A
B
A
B
So
there
was
basically
some
issues
with
creating
cluster
in
some
of
the
disappearance
and
so
that
kind
of
trigger
some
in
choosing
the
boscoe's
and
Mason
code
that
we
deal
with,
but
basically
now,
instead
of
creating
all
the
cluster
in
one
specific
region,
we're
just
one
webbing
across
all
region.
So
we
should
see
em
much
less
issues
and
also
increase
the
pool
of
the
number
of
projects
we
have
100
projects
now
and
I
think
in
very
busy
Dee
Dee's
day
we
use
like
at
most
50.
So
we
should
have
plenty
of
room
for.
B
So
the
problem
was
because
we
ran
out
of
we
sauce
in
one
region
yeah,
but
you
can't
tell
from
the
project
country
you
don't
know
it's
only
when
you
after
you
create
that
feels
like.
Oh,
there
is
not
enough
resource
and
I
didn't
find
any
API
to
actually
query
that
so
I'm,
just
looking
at
all
the
regions
and
and
creating
one
in
a
different
publicly
which
time,
okay.
B
A
A
B
A
A
A
This
in
the
submitted
codebase
and
on
the
pre,
submit
tab.
You
can
see
the
test
results
for
other
peels,
so
in
case
the
test
is
failing
for
your
PR.
You
can.
You
can
take
a
look
and
see
if,
if
the
test
is
also
failing
for
appeals-
and
that
could
give
you
some
ideas
about
where
this
infrastructure
problem
or
is
just
your
PL
right,
so
this
should
be
a
very
useful
for
deep
for
troubleshooting,
debugging
or
development.
F
It
easily
accessible
from
like
your
invented
test
fails
like
the
github
UI.
If
it
is,
that's
a
big
I
think
super
helpful.
Well,
what
what
do
you
mean
no
like
if
I
go
to
my
github
page
and
then
I
see
some
test
failure
right
and
from
there
I
can
go
to
the
test
itself,
but
like
can
I
actually
trampoline
to
this
one
or
do
I
need
to
keep
the
link
to
this
somewhere
in
my
bookmarks.
B
F
F
Okay,
so,
ideally
I
might
you
know
I
submit
my
peer
and
then
my
one
of
the
test
fails
right
and
the
boss
says:
okay,
you
I
noticed
that
one
of
your
test
fails.
Here's
a
link
there.
You
can
go
and
see.
If
this
is
a
flake
or
not,
and
that's
it
right,
it
doesn't
have
to
be
formatted
or
like
it
doesn't
have
to
give
you
the
landing
page
like
the
context-specific
link
for
your
test.
It
can
be
just
the
main
home
page
for
this
system,
but
even
that
is
I
think
would
be
super
useful,
yeah.
A
G
E
G
E
Steps
are
Kumiko,
it's
it's
kind
of
on
the
back
burner
for
at
the
moment,
but
she's
going
to
take
a
better
look
at
requirements.
Basically,
we
have
two
apps
right
in
our
current
test
framework.
We
have
I
think
the
simple
intent
tests
use
for
do
kind
of
as
their
little
client
to
do
to
kind
of
force
to
simulate
traffic
between
services.
The
pilot
tests
use
its
own
app,
so
we
basically
want
to
have
one
app
for
the
test
framework.
So
Kimiko
is
going
to
look
at
for
Tod.
E
G
G
B
G
Looking
food
is
a
separate
need
for
that
right.
The
need
is
to
make
sure
we
don't
break
the
demos
that
outside
so
I
mean
should
not
be
used
really
for
functional
testing.
It's
more
like
to
make
sure
that
the
tasks
on
the
site,
the
website
still
the
very
wrong
Okin
accidentally,
oh
yeah.
We
should
we
should
move
more
feature,
oriented
testing,
synthetic
apps
and
leave
booking
for
just
as
the.
G
If
what
you
is
the
library
right,
not
just
like
you
can
run
it
as
a
service,
but
you
can
run
it
as
the
library
to
so
I
would
think
them
and
if
you've
got
the
way
the
web
socket
is
actually
invested
in
this,
it's
really
not
kind
of
hockey
she's.
He
and
then
just
a
pose
one
line.
What's
okay
and
tries
to
read
it
back
so
I
think
that
can
stay
as
a
site.