►
From YouTube: Cells : Demo & Office hours 2023-08-24
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
B
C
Okay,
I
think
I'll
start
yeah,
so
my
agenda
is
to
demo
group
creation
of
cell
2
today,
so
we
have
reached
the
mobile
State,
even
though
all
the
fixes
are
not
on
Master
yet,
but
we
have
solutions
to
all
the
problems
that
we
face
for
the
bloggers
that
we
face
for
the
demo,
so
I
have
fixed
it
in
my
local
and
I
have
a
branch
for
that.
C
So
I'm
running
that
branch
in
my
local
today
and
showing
you
the
demo
so
before
that
I
think
I
will
go
through
the
current
state
of
where
we
are
so.
There
is
just
one
more
table
to
be
moved
to
sell
local
schema,
which
is
the
members
table.
The
review
is
currently
ongoing
for
that
and
with
that
we
will
finish
all
the
tables
that
needs
to
be
moved
and,
after
that,
it's
just
about
fixing
the
problems
that
come
up.
B
C
The
stage
where
we
try
to
create
a
group,
so
my
estimate
is
that
there
are
six
more
issues
like
that,
which
happens
when
we
try
to
create
a
group
on
Cell
two,
so
I've
listed
them
down.
Three
of
them
are
loose
foreign
key
issues,
which
means
that
since
we
have
moved
tables
across
databases,
foreign
Keys
need
to
be
converted
to
lose
foreign
keys.
So
these
three
issues
are
straightforward
because
we
have
a
scripture,
generator
migration
and
once
you
undergo
the
migration,
those
problems
are
fixed.
C
So
the
first
is
the
part
where
so,
when
you,
after
you
create
a
group,
you
are
redirected
to
the
group
to
show
the
group,
and
and
for
that
you
need
the
roots
table
and
currently
We
join
the
roots
table
and
the
namespace
is
stable
to
find
the
namespace
with
that
root.
But
Roots
is
a
cluster
by
table,
but
name
specials
is
a
local
table.
So
you
cannot
join
these
two.
C
So
surprisingly,
there
is
already
an
issue
for
that,
because
the
performance
of
the
Cross
join
itself
is
not
really
that
great,
so
I
am
just
taking
up
the
same
issue
either
database
team
will
take
it
or
we
should
take
it,
but
anyway
the
the
point
is
that
we
should
avoid
this
cross
join.
Otherwise
we
cannot
navigate
to
a
group
on
Cell
too,
so
this
is
the
first
part,
but
to
solve
it
on
local.
Currently,
the
the
approach
that
I
have
taken
I
will
share
my
screen.
A
C
Lines
so
that
the
joints
do
not
come
into
picture
and
since
you
already
have
it
will
just
perform
a
query
as
described
in
the
issue
itself,
so
that
fixes
that
problem
then
the
second
cross
join
issue
is
yeah
over
here.
When
you
try
to
join
across
users
and
members
users
is
castoroid,
members
is
a
local,
so
you
cannot
join
them
and
I
saw
that
the
error
happens
in
first
owner,
so
I
have
fixed
first
toner
in
my
local,
where
so
this
was
the
problem
area.
C
If
you
look
at
owners,
has
many
owners
is
joining
across
members
and
users,
so
we
cannot
use
it
anymore.
So
I
am
just
trying
to
figure
out
first
owners
from
like
getting
all
the
owners
and
then
get
the
first
member
record
and
the
first
user
of
that
member
record
will
be
the
first
donor.
So
it's
a
very
small
fix
and
that
solves
that
issue
and
the
third
one
is
a
small
bug
fix,
for
which
we
also
have
a
solution.
C
So
what
happens
is
when
you
every
time
you
log
in
it,
was
trying
to
create
a
new
username
space
which
is
not
really
needed?
You
you
need
a
username
face
only
when
you
create
a
user,
so
the
difference
here
you
can
see
so
we
should
create
a
new
namespace
for
a
user
only
when
the
user
record
itself
is
new,
so
that
fixes
that
issue.
So
we
have
solutions
for
all,
but
the
the
pain
Point
here
is
checking
the
performance
of
the
new
approach
and
getting
it
past
review
so
yeah.
C
So
I
will
proceed
to
demo
now
I
have
my
GDK
running
with
like
two
different
gdks
for
two
different
cells
yeah.
So
this
is
the
first
cell,
as
we
usually
do.
C
C
Yeah
I
can
navigate
to
the
group
on
cell
one
so
that
yeah
user
will
create
a
group
in
cell
one
yeah
that
that
works
as
expected
and
in
a
new
incognito
window
I'm
going
to
delete
all
support
3001,
which
is
where
the
second
GDK
is
running.
So
this
says
that
already
login
user
can
try
to
sell
to
the
already
logged
in
part:
that's
not
working,
yet
it
might
be
a
stretch
goal
for
this
Milestone.
But
currently
we
are
just
logging
in
again
on
the
second
vdk.
C
C
Yeah,
so
this
also
works
now
and
now
we
can
navigate
to
yeah
navigate
to
this
page,
where
you
can
see
all
the
groups
on
Cell
2
so
cell
to
group
is
here
and
I
go
back
to
cell
one.
C
A
A
I
haven't
written
it
in
the
document
yet
about
like
the
why
the
user
is
not
logged
in
in
on
Cell
two,
when
the
user,
logs
in
and
cell
one
and
what
I,
when
I,
try
to
do
that
locally.
I!
Try
to
configure
the
the
cell
to
reduce
the
sessions
with
this
to
point
to
the
readers
of
to
the
sessions
to
this
of
cell
one,
but
that
did
not
solve
the
problem.
So
I'm
wondering
like
what
what
would
be
the
like,
because
I
remember.
We
had
this
discussion,
I
think
with
Tong
and
the
song.
A
You
said
like
it's
not
about
cookies.
It's
about
the
sessions
with
this,
but
that
didn't
solve
the
problem
for
me.
So
I'm
wondering
like
if
there's
something
else,
if
that's
something
we
want
to
solve,
maybe
not
for
the
demo,
not
for
16.4,
but
I'm
still
curious
like.
Why
didn't
this
didn't
help
solve
the
problem?.
B
In
this,
this
I
guess
relates
to
number
two,
which
is
that
you
know
I've
added
that
that
issue
as
a
blocker
but
as
mono
says
it
probably
is
a
stretch,
but
that
issue
is
scheduled
for
16.4
as
well,
so
I
guess
you're
already
working
on
it.
Uma
yeah.
A
B
B
B
B
Yeah
so
come
here,
we're
trying
to
figure
out.
Why
is
sharing
the
redis
sessions
between
GDK?
Doesn't
work
I,
don't
know
whether
you
successfully
shared
ready
sessions
with
gck
I
did
and
that
worked
without
any.
You
don't
need
to
modify
application
code.
D
You
need
to
have
the
same
cookie
encryption.
Key
GDK
May
generate
a
new
one,
GDK.
A
B
Yeah
because
my
my
script,
when
I
set
up
Audi
I
only
updated
DB
key
base,
I
think
yeah
I
need
to
check
that
out.
D
B
Yeah:
okay,
okay,
my
GDK
secondary
script
is
only
making
the
DBK
base
equal,
but
not
the
SQL
key
base.
So
yeah.
D
A
A
D
A
We
fine
for
the
demo,
then,
is
that
cell
two
shares
the
same
sessions
with
this
with
cell
one.
A
A
D
D
D
D
So
yes,
like
ideal
individual,
basically
you
know
you
would
not
have
to
re-log
in
interestingly,
you
can
run
Mana
git
Labs
under
different
parts
and
as
long
as
the
same
as
the
hostname
is
the
same,
the
cookies
for
that
hostname
will
be
searched
across
different
ports.
A
A
D
A
Okay,
well
because
I
I
intentionally
like
tried
this
so
I
can
also
take
over
this
issue.
If,
because
I
already
started
the
question
so
I
can
assign
the
issue.
D
D
Sorry,
gck
I
run
exactly
from
the
same
declare
price
for
the
I.
Don't
have
separate
it
up
race,
folder
at
all,
so
every
new
cell,
that
you
run
you
kind
of
runs
from
all
the
same
Source
from
the
from
the
same
files.
So
if
you
modify
like
fi,
it
gets
modified
on
ulcers
automatically.
So,
basically
for
me,
in
my
case,
everything
is
held
pretty
much
just
what
is
different
like
there
is
some
yaml
configs
with
ERB
content
that
changes
based
on
the
nth
variable.
B
Got
to
know
yeah
we
can.
We
can
do
that
with
GDK,
because
in
order
to
I
think
run
it
run
it
and
it's
and
it's
the
process
of
a
different
folder
I
mean
yeah,
I.
Think
I
think
we
can
do
that,
but
that's
quite
hacky,
so
yeah,
let's
see,
let's
see
what
the
minimum
we
need
to
share
is
no
it's
good
to
find
out
anyway.
What
what
we
need
to
share
this
way,
VIA,
gtk
so
I,
think
so
perfect,
Secrets.
B
Cool
sorry,
manosh
I
think
we
we
jumped
ahead.
Is
there
anything
else
from
you.
C
No
so
yeah
so
yeah,
so
like
everything.
So
we
have
the
solutions
but
getting
into
an
MR
and
checking
performance
of
the
new
approach
and
getting
it
through
review
that
is
like
now,
but
otherwise
we
are
ready
for
a
demo.
D
C
C
Like
it's
a
certainty
conversation,
so
we
might
need
to
fix
that
as
well,
because
we
are
expecting
an
error
for
a
join,
but
we
did
not
get
the
error.
Rather,
it
was
joining
with
the
roots
table
in
the
set
local
database
itself,
which
is
problematic
because
it
will
not
surface
PR
as
we
expected
to.
D
C
D
C
C
Yeah,
just
one
one
more
one
more
question:
if
you
have
time
yeah,
so
there
are
a
lot
loose
foreign
Keys,
also
to
be
changed
and
I
was
wondering
what
usually
like.
Do
we
look
at
all
the
cases
where
we
use
like
table
dot,
funky
name,
and
do
we
all
like
replace
all
those
instances?
How
is
it
usually
done,
because
the
the
associated
record
might
be
gone,
but
a
table
dot
foreign?
The
value
will
remain
in
the
table
for
quite
a
while
at
least
a
minute.
So
I
was
wondering
like.
C
B
D
D
A
D
The
first
one
is
non-issue
really
because
of
the
relation
we're
gonna
remove
namespace,
so
actually,
in
the
worst
case,
entry
in
the
rules
we're
gonna
be
still,
but
we
we
are
doing
close
one
or
like
we're
doing
the
second
stage
select
of
the
namespace
entry
user
notifications
typing.
So
we
remove
users
and
we
are
left
the
notification
settings
right.
D
A
On
the
main
database,
it's
it's
fast
like
most
of
the
time.
It's
like
a
I
mean
by
fast,
like
you
like.
It,
takes.
C
A
Every
minute
the
job
runs
and
one
for
Main
and
one
for
CI
and
usually
the
main
database
is
empty.
But
now
the
CI
database
is
the
the
one
that's
struggling
a
little
bit.
It's.
A
Yeah
I
mean
if
the
if
the
yeah
I
mean
if
the
source,
if
they
are
Source
record,
that
was
deleted
on
Main,
then
we
don't
have
problems.
D
It
could
be
probably
even
like
a
manual
test
if
your
accuracy
could
probably
do
the
same
for
the
rules,
but
from
my
experience
probably
come
to
the
roots
like
we
just
look
close
to
him
very
often
or
like
the
second
stage,
patch
of
the
record
I
think
we
don't
use
rules
as
a
source
of
tune
just
for
getting
IDs.
We
always
go
with
intellectual
namespace
obituary.