►
From YouTube: video1788798686
Description
October 5, 2022 Jakarta EE Platform TCK call #24. Minutes can be viewed via https://docs.google.com/document/d/1V1dDLJkd14EDRMPeuI0VzPtU4Lbli8FFBd1pLDLlOrY/edit#
A
A
Hey
welcome
everybody
so
yeah,
so
yeah
we
created
a
a
branch
for
10.0
in
which
we
had
missed
a
few.
You
know
we
had
missed
a
few
changes
somehow
so
we
corrected
that
and
now
the
the
you
know,
the
the
10
branches
tagged
for
final
and
also
yeah
contains
all
all
of
the
changes
that
were
you
know
that
were
made
in
in
the
ee10
platform
tck.
So
yeah,
thanks
to
Guru
for
pointing
out
that
I.
A
You
know
that
we
had
missed,
missed
a
few
and
so
not
much
more
to
say
about
that.
We
just
have
a
branch,
so
we
you
know,
can
handle
tck
challenges
in
the
future,
which
is
great
and
got
some
feedback
on
the
pull
request
to
up
update
from
ee
10
to
11
with
just
you
know,
versioning
and
then
pck,
doc,
user,
doc,
just
yeah
not
much
to
just
to
discuss
there,
but
just
wanted
to
mention,
and
in
case
there
was
feedback
here.
A
I,
don't
see
anyone
in
the
in
the
in
the
agenda
in
the
list
participants
list
that
gave
feedback
so
far,
so
maybe
you
know
maybe
we
just
move
on
but
feel
free
to
speak
up.
If
you
do
have
feedback
on
that
or
comment
on
it,
but
I
think
the
media
is
really
on
the
you
know
the
tck
refactoring
and,
let's
see
so,
we
have
issue
five
four.
A
A
Okay,
so
you
know
one
one
general
idea,
you
know
which
we
talked
about
just
to
kind
of
you
know
mention
it
again
is
to
you
know
to
do
continue
the
refactoring
in
the
master
branch
we
talked
about
about
that
before
I,
think
on
the
mailing
list
and
in
previous
calls,
but
you
know
the
idea
is
to
put
us
into
the
situation
where
we
have
you
know.
Will
we
have
you
know
new?
A
B
A
Know
will
be
and
have
the
refactored
tests
you
know,
live
in
the
the
platform
tck
a
bit.
You
know
a
a
bit
longer
or
long
enough
until
all
of
the
tests
are
ready
to
move
to
a
particular
spec
and
kind
of
you
know
kind
of
a
way
to
avoid
needing
to
bring
the
old.
You
know
some
of
the
old
ant
based
tests
into
us
into
the
spec
along
for
the
ride
when
we
move
the
refactor
refactor
test
over.
A
A
I'll
say
yeah
taller
yeah
in
platform.
Pck
is
I,
guess
is
a
good
is
one
way
to
you
know
one
way
to
State
it
where
the
tests
that
I
moved
over
you
know
aren't
sorry
the
tests
that
aren't
refactored,
maybe
those
either
go
somewhere
else
as
ant
tests
and
I
had
a
few
suggestions
for
what
that
could
be
or
maybe
they're
removed.
A
Because
they're,
you
know
not
really
I
mean
they
are
part
of
the
requirements
for
that,
for
whichever
spec,
but
it's
it's
just
it's
something
we
could
decide.
We
could
decide
to
do
and
we'd
have
to
talk
about
that
more.
You
know
to
get
to
a
point
where
we
have
nothing
but
the
new.
You
know
the
new
style
tck
test
for
that
technology
instead
of
having
all
the
old
Legacy
ones.
So
either
you
know
the
tests
that
aren't
refacted.
A
Maybe
they
move
to
it
like,
let's
just
say,
I'll
just
add
some
options
here.
So
you
know
see:
remove
any
not
refacted
tests
from
the
platform
tck
or
move
those
tests
to
a
Legacy,
just
not
yeah
that
this
wouldn't
be
the
name,
but
a
legacy
amps
based
QE.
Let's
call
it
a
QA
test,
repo
thing
where
they
could
live
forever
kind
of
thing
and
I'm,
not
I'm,
not
sure.
A
Personally,
you
know
you
know
what
yeah,
which
is
best
because,
like
you
know
it
it's
better
to
not
remove
any
of
those
old
to
delete
to
delete
those
old
tests
because
they
asked
pile
the
specs
still
and
that
yeah
that's
been
covered,
but
there's
a
heavy
price
to
keep
them
around,
as
well
as
a
requirement
to
be
run
as
opposed
to
you
know.
Maybe
they
could
be
acute
like
a
treated
as
a
QE
test.
Suite
I,
don't
know
just
to
throw
that
idea
out.
There.
B
A
A
So
we
so
so
the
only
thing
that
went
wrong
with
you
know
doing
that
for
ee10
is
we
went
from
running
the
you
know
those
tests
that
were
moved
over
from
running
those
as
a
platform
as
platform
tck
tests
to
run
on
them
as
Standalone
tck
test?
Because
it's
that's
what
you
know
it
was
kind
of
going
into
a
standalones
spec.
A
A
You
know
set
of
problems
for
implementations
to
deal
with,
so
it
kind
of
you
know
it
got
a
lot
of
yeah
late
problems
that
were
run
into
with
that
and
if
we
just
do
a
lot
of
that
it
just
it
slows
the
implementations
down
that
I.
That
might
not
be
ready
to
you
know
to
test,
but
I
mean
we
could
smooth
that
out,
but
we
can
improve
that
so
that
it's
easier
just
to
use
the
platform
dck
configuration
and
we
have
ideas
for
that.
Well,
yeah.
B
B
B
B
A
A
A
B
A
While
others
did
a
lot
of
the
of
the
work
on
the
refactoring
on
those
different
specs
like
Alwyn
spent
a
lot
of
time.
You
know
you
know,
you
know
worked
on
that.
I
didn't
I
didn't
personally
spend
as
much
time,
but
you
know
this
time
around
I
wanted
to.
You
know:
Jump
Right,
In
and
and
spend
more
time
on
that
I.
B
B
A
B
B
B
A
Yeah
there
was
I
I,
I
vaguely,
remember
the
magic
Andre
I
think
worked
with
Scott
Scott
on
that
and
I
I
vaguely
remember
some
magic
stuff
that
they
did
to
to
deal
with.
You
know,
generally,
it's
like
a
maven
Plug-In
or
something
that
they
that
they
that
they
did
but
I
remember
what
it
was
exactly
but
I
do
know.
You
know.
A
Scott
Scott
has
worked
on
the
batch
tck
for
many
years
and
it
lived
in
CTS
and
now
the
platform
tck
and
you
know,
with
the
changes
instead
of
doing
the
work
in
the
platform
tck
and
then
he
would
always
like
migrate.
The
like
the
I
think
that
he
you
know
he
would
run
you
know
he
would
always
take
the
work
out
of
there.
A
B
Forget
yeah,
my
big
recollection:
is
they
layered
the
so
they
basically
had
ever
the
test
as
basic
junit
test,
and
then
they
had
another
module.
That
was
using
one
of
the
arcillian
runners,
but
they
would
basically
create
an
archive
that
included
all
the
dependencies
and
ship
that
over
and
then
run
it
basically
embedded
in
a
in
a
container,
and
so
they
had
an
example
of
doing
it
with
Liberty.
B
A
A
B
Yeah,
that
was
that's
a
long-standing
issue
that
I
raised
and
probably
at
the
start
of
E10,
that
we
certainly
never
got
to
it's.
How
do
we
break
these?
The
circular
dependencies
we
have
at
the
tck
level
between
a
spec
that
you
know
technically
does
not
have
an
API
dependency
on
it
Downstream,
but
it
tries
to
describe,
or
at
least
test
behaviors
and
containers
that
are
Downstream
api-wise
yep.
A
B
A
A
You
know
on
time
every
you
know
which
wouldn't
be
awful,
because
that
means
there
is
only
going
to
be
one
wave
and
then
it's
all
at
once,
but
it
also
means
we
can't
test
anything
until
we
have
everything
done,
which
is
really
bad,
so
we
kind
of
need
to
go
yeah,
I
I.
My
feeling
is
that
we
need
to
go
in
the
direction
of
having
the
integration
tck
like
the
platform,
dck
or
yeah.
Something
like
that.
That
kind
of
you
know
separates
that
out,
so
so
that
specs
can
complete.
A
A
B
Yeah,
but
you
know
by
definition,
we're
talking
about
things
that
are
part
of
a
platform
or
profile
specification,
see
you
know
or
or
they
are
cross-cutting
cross-cutting
concerns
that
yeah
a
specification
team
has
tried
to
find.
So
you
know,
there's
got
to
be
at
least
some
cross-section
or
some
set
of
the
two
specification
teams
that
have
interest,
or
else
you
know
we're
we're
not
aligned
correctly.
I
mean
servlet
is
the
worst
because
they
pretty
much
by
definition,
are
cross-cutting
thing.
I'm,
not
I'm.
Sorry,
not
sorry
about
security.
B
B
B
B
A
A
B
A
A
A
Pushing
Jay,
you
know,
testing
JS,
let's
say
J
unit,
so
yeah
there's
a
lot.
A
lot
of
you
know
like
they
were
I
I
was
asked
to
to
spend
time
with
user
groups.
They
kind
of
like
help
train
some
of
the
user
groups
out
there
on
helping
with
tcks
and
yeah
I.
Just
see
that
as
something
that
comes
after
we
do
the
refactoring
I
don't
know
if
that's
right
or
wrong,
but
yeah.
A
For
my
money,
the
the
part
of
the
platform
TCU
is
confusing
is
not
really
the
ant
part.
It's
the
Java
test
harness
like
that's
the
part
that
requires
the
massive
jte
file
and
all
that
stuff
like
that.
That
is
the
complicated
portion
of
making
sure,
you're
configuring
and
getting
stuff
running
and
maybe
I,
know
amp
well
enough
that
you
know
that
stuff
doesn't
really
bother
me,
but
yeah.
B
A
A
Yep,
so
platform
tck
could
be
easier
with
ants
without
test
Java
test
harness
and
yeah
I.
Don't
know,
I
haven't
really.
You
know
thought
about
that,
but
yeah
it.
It
I
think
that
there's
a
lot
of
complexity
with
with
things
that
we
don't
necessarily
you
know
need
so
there's
a
lot
of
nice
features,
and
you
know
things.
A
B
A
Yeah,
so
if
if
you've
removed
I
mean
you
wouldn't
know,
really
know,
what's
used
and
not
used
by
the
different
users,
really
I
guess
it's
hot
yeah,
but
in
terms
of
the
plug
points,
but
there's
things
like
yeah
definitely
there's
features
that
could
be
removed,
I
mean
it
could
be
simplified,
but
anyways
I.
Guess
it's
besides
the
point,
but
but
yeah.
A
A
Was
it
in
the
runner
or
in
the
examples
I
figured
now,
but
there
is
one
example
in
there
anyway,
so
you
that
you
did
Scott
forget
now,
but
I
know.
We
have
an
example
that
we
can
go
by
to
to
I.
Think
I
think
it
was
like
to
run.
It
I
think
you
to
run
a
Json
test
in
a
container
or
something
like
that.
If
we
wanted
to
it
was
like,
but.
A
A
Is
hard
to
to
hard
to
think
about,
but
I'll
just
throw
some
ideas
out
and
I'm
going
to
throw
it
under
standardization
because
it's
like
you
know
if
we
kind
of
pull
together,
you
know
figure
it
out.
This
stand
standard
knives
and
all
of
us
I
think
we
kind
of
have
to
do
it
on
some
timeline
yeah
before
yeah
yeah
before
certain.
You
know
like
at
some
point
like
ballots
have
to
you
know.
You
know,
ballots
have
to
happen.
It's
at
some
point.
A
Some
point
so,
if,
if
the
work
is,
you
know,
originates
in
the
platform,
tck
or
or
maybe
it
moves,
you
know
moves
elsewhere
and
not
tied
to
that
idea.
But
wherever
the
work
happens,
if
we
come
up
with,
if
we
St,
if
we
come
up
with
a
framework,
that's
that's
one
thing,
and
then
you
know
we,
you
know
we
kind
of
have
to
prove
it
with
a
few.
A
A
A
But
what,
after
we
start
like
you
know
the
ballots,
we
kind
of
you
know
maybe
yeah.
Maybe
we
have
that
flexibility
to
to
do
that
still
somehow
that
haven't
really
figured
that
out
yet
it'd
be
nice.
If
we
could
do
that,
and
they
were,
you
know,
I
think
there's
some
ideas
for
that,
but
but
it's
it's
either
one
way
or
the
other,
though
it's
either
yeah
yeah.
A
A
You
know
period
of
time
where
we
can
do
validation,
and
then
we
have
a
period
of
time
that
we
can't
do
much
validation
or
we
can't
validate
as
much
or
everything
until
we
get
near
the
ends
like
we
did
in
like
in
may.
We
finally
started
doing
platform.
Tck
runs
with
glassfish
and
there
was
like
a
big.
You
know,
I
think
it
was
like
like
in
September.
A
We
could
still
run
against
ee9,
one
implementations
it
may
be,
or
maybe
it
was
August
I
forget
now,
but
then
yeah,
then
we
we
made
changes
from
for
many
months
and
merge
changes
and
made
changes,
and
we
had
no
idea
dear.
If,
if
we
had
no
way
of
testing
those
changes,
so
I
just
kind
of
wanted
to
point
out
that
we
kind
of
have
that.
You
know
that
to
deal
with
and.
B
A
B
B
B
B
B
You
yeah
so
the
main
thing
that
I'll
that
I'll
ask
Alwyn
and
Guru
is
just
you
know
what
their
thoughts
on
you
know.
Is
there
any
way
that
we
can
get
like
an
initial
script,
or
you
know
at
least
a
pattern
to
okay?
How
do
I
take
this
set
of
you
know
container,
or
this
set
these
set
of
tests
that
are
involved
in
both
the
technology
in
a
given
container
and
move
those
over
to
a
junit
based
J.
B
B
A
A
A
I,
don't
yeah
I,
don't
think
it
was.
You
know
it
it's
it's
not
accepted
as
as
the
plan,
but
you
know
it
was
mentioned
as
a
possible
option,
yeah
that
we
could
do
to
if
we,
if
we
wanted
to
have
something
that
could
run
on
I
I
think
they
did.
Was
you
know,
so
you
could
run
on
jdk,
21
or
any
of
the
you
know
any
of
the
jdks
that
you
want
to
run
on
without
the
security
manager
was
kind
of
the
idea
with
ee10.
Essentially
so.
A
B
A
A
A
Don't
know
if
it's
a
big
a
big
deal,
but
it
came
up
before
491
that
yeah
that
you
know
eager
versus
lazy
class
loading
should
be
you
know,
should
be
handled
as
well.
Since
it's
you
know
since
it's
something
that
any
any
jvm
could
could
use
so
so
Scott.
Do
you
think
a
10.1
like
your
proposed
Terror
would
would
that
kind
of
be
a
good
idea
as
a
step
in
the
refactoring
to
kind
of
do
the
the
security
manager
thing
isolated
or
or
will
it
delay
further
refactoring.
A
I
I
think
there's
a
small
group
of
people
who
will
do
the
who
will
do
the
work
you
know
would
be
taken
away
from.
You
know
like
from
doing
the
general
work
like,
but
it
it
might
be,
I
guess
if
it's,
if
it's
just
a
security
manager
aspect
I
you
know,
then
maybe
that's
not
the
case.
Maybe
that's
just
you
know.
A
A
A
A
A
B
B
Could
we
move
to
Java
SE
21
for
11.,
without
some
intervening
update
to
move
first
to
17.,
that's
kind
of
where
I
was
seeing
that
fall.
But
you
know
at
this
point
in
the
store
early,
but
it
is
a
lot
of
work
for
maybe
not
a
lot
of
gain.
So
I
mean
I.
Think
that
the
conversation
that's
still
going
to
have
to
get
fleshed
out.