►
From YouTube: Web IDE Terminal Discussion
Description
A discussion about the current state of the Web IDE + Terminal, where we're going, and what we need to do to get there.
A
So
this
is
a
video
for
the
latest
feature
for
the
web
terminal,
which
is
talking
about
file
syncing,
so
before
people
would
spin
up
the
web
terminal
would
start
would
start
a
pre-configured
environment,
but
as
I
make
changes
in
the
web
IDE
those
changes
were
never
synced
to
the
terminal.
So
now
in
12-point,
oh,
we
just
released
a
feature.
We're
making
changes
here
and
the
web
ID
are
reflected
in
terminal.
A
So
this
is
all
this
running
locally
I
just
created
this
demo
for
for
UX
to
help
review
it,
and
it's
important
that
I
note
that,
because
it's
not
intuitive
to
get
this
running
to
start,
as
you
can
see
actually
here
to
the
left,
is
a
configuration
file.
That's
needed
for
if
you
want
the
web
I
web
IDE
terminal
available
to
you
at
all.
So
you
have
to
specify
it
is
nice.
You
can
specify
what
image
is
gonna
be
running
for
that?
A
A
But
yes,
so
there's
a
little
bit
of
configuration
to
get
this
web
ready
terminal
running,
but
once
it's
running
it's
pretty
nice.
For
some
reason
it
doesn't
default
to
my
project
directory.
But
that's
fine.
You
can
run
just
commands
in
the
terminal
and
we
can
see
the
our
project
directory
nicely
and,
as
you
see,
I
can
make
changes
and
those
changes
are.
A
Reflected
into
the
repositories,
all
of
that
is
very
sweet,
but
going
back
to
our
document,
there's
there's
a
lot
of
assumptions
around
the
usability
of
this.
If
one
getting
the
kubernetes
runners
all
set
up
getting
this
specific
web
IDE
configuration
set
up,
I
have
to
be
a
maintainer
to
even
run.
This
there's
also
some
some
issues
that
come
up
since
I'm
coupled
to
CI.
This
is
a
positive
demo,
so
I
haven't
shown
anything
negative
here.
A
There's
the
web
idea
terms
the
web
body
dots
which
have
the
terminal
bit
to
it,
and
you
see
a
lot
of
the
notes
highlighting
that
there's
assumptions
here
to
your
environment,
so
I'd
like
to
maybe
itemize
that
list
of
weaknesses
that
we
see
and
itemize
if
there's
strengths.
So
we
see
about
this
or
we
can
spend
time
talking
about
where
we'd
like
to
be
I,
really
want
to
just
start
this
holistic
conversation
and
from
that
now
I'm
gonna,
step
back
and
I'm
gonna
hit
the
new
button.
B
Okay
I'd
like
to
talk
about
those
restrictions.
I
mean
we
are
okay,
it
doesn't
seem
so,
but
we
are
moving
quite
fast
and
we
are
working
with
things
that
we
don't
really
know
how
they
are
going
to
evolve.
I
mean
those
restriction
regarding
only
maintainer
countries
or
the
terminal
or,
for
example,
we
can
only
have
one
way
by
the
pipeline
at
a
time,
I
mean
if
you
run
a
terminal
and
you
open
a
new
tab
with
the
same
project
and
you
try
to
well.
You
won't
be
able
to
run
the
a
new
terminal.
Why?
B
How
our
CI
environment
will
behave
with
with
so
many
CI
jobs,
I
mean
in
my
test.
I
could,
for
example,
reuse
the
running
pipeline,
the
running
way
buddy,
for
example,
you
are
running
a
web
ad
terminal
and
if
you
open
a
new
tab
with
the
same
project,
the
same
pipeline
that
the
same
terminal
will
be
open.
B
Okay
in
order
to
reuse
this
terminal,
so
we
still
have
to
decide.
How
do
we
body
terminal
and
all
this
stuff
is
going
to
work,
so
those
restrictions
were
the
most
restrictive
ones
we
could
set
at
that
time
in
order
to
see
how
is
feature
box,
so
it's
intended
to
be
to
be
for
for
developer,
mostly
because
it's
what
makes
sense,
but
this
is
still
a
new
born
a
new
feature.
So
we
are.
B
D
A
D
B
To
open
it
to
all
the
users,
because
we
want
to
limit
the
risk
a
little
bit
about
the
startup
time.
I
talked
to
Paul
comment
on
idea
regarding
how
we
could
boost
this.
This
start
time.
I
was
thinking
that
the
about
the
same
mechanism,
virtual
rockstars,
with
snap
salt
and
resume
and
post
thing,
but
I,
talked
to
Steve
at
Sephardi
in
the
CI
team,
and
if
the
image
is
already
cars,
the
mean
time
goes
into
starting
the
day
container
starting
the
bill.
B
B
Another
thing
we
are
facing
right
now
is
that
in
this
scenario,
we
are
going
to
start
more
than
one
container
I
mean
also
Paul.
So
how
for
the
weaponry
in
that
video,
we
needed
the
main
image,
which
is
the
big
container
image
and
an
additional
image
folded,
but
for
the
file
syncing
process,
okay
in
CI.
B
Well,
let's
say
in
kubernetes:
this
matching
starts
in
parallel,
so
it
may
happen
that
the
file
sync
image
starts
before
the
build
image.
Okay,
so
for
the
file
sync
mirroring,
we
need
the
cloned
repository,
which
is
cloned
in
the
build
image.
So
when
the
file
sync
image
starts,
it
tries
to
get
the
repository
clone
in
the
build
image
and
because
the
build
image
is
not
ready.
Yet
that
process
spray.
B
Okay,
so
and
the
same
goes
for
the
for
the
for
the
terminal,
it
may
happen
that
we
try
to
connect
to
the
terminal,
but
the
pill
image
is
not
ready
yet.
So
that
is
something
I'm
going
to
create
an
issue
about,
because
I
discussed
it
with
also
with
the
steam.
We
are
going
to
create
a
new
endpoint
in
individual
Runner,
something
like
health
check
service
or
something
like
that
in
order
to
know
when
all
the
containers
are
in
a
ready
state.
Okay.
B
The
get
lab
we
body
German
file,
we
can
provide
templates
the
same
way
we
do
with
Scylla
adapt
ci.
So
we
can,
for
example,
have
a
select
box
saying
what
do
you
want
to
configure?
Do
you
want
live
preview?
Do
you
want
file?
Sync
mirroring?
What
do
you
want
and,
depending
on
what
you
select,
we
can
feel
automatically
that
German
fire
some
useful
some
users
with
these
first
step
with
the
body
channel.
B
Okay,
first
of
all,
for
all
all
that
doesn't
know
how
the
web
are.
A
channel
works
is
a
restricted
version
of
the
kidnapped,
see
a
jungle.
Okay,
it
doesn't
support
all
the
syntax
that
they
comply
channel.
Does
okay,
it's
a
restricted
person
just
for
this.
In
the
same
way,
we
are
adding
something
taxed
to
the
way
body
german.
Do
that
they
killed
up.
Ci
channel
doesn't
support
okay.
We
are
starting
to
divert
a
little
bit
but
I
think
part
yeah.
B
Okay,
we
are
fine
for
now,
but
we
may
have
problems
in
the
future
if
we
are
too
far
from
each
other
in
this.
In
this
configuration
file
and
yeah
I
mean
I'm
also
worried
about
this
start
a
time
because
I
mean
when
you
use
code
samples
or
these
kind
of
tools-
and
you
click
the
terminal
tab
automatically
pop
up.
So
it
feels.
A
Different
and
that's-
and
that's
my
and
that's
what
every
week
so
I
think
we
listed
it,
itemized
I
think
a
lot
of
the
weaknesses
that
you
and
I
have
identified
here
fun,
but
that's
the
main
reason
I
want
us
to
potentially
talk
about
this
from
a
very
large
media
vision
perspective.
So
I'm,
looking
forward
to
jumping
to
the
next
section
is:
where
would
we
like
to
be
cuz
yeah?
You
look
at
these
other
these
other
products
that
provide
a
web
environment
with
a
terminal
and
editor,
and
it's
seamless.
A
There
is
no
user
configuration
and
I
realize
that
we
did.
We
get
some
flexibility
with
approval
with
providing
this,
but
yeah
there's
no
startup
time.
This
is
all
there
and
I.
Think
part
of
us
I'm
worried
that
we
were
looking
for
implementing
functional
requirements
when
we
didn't
really
have
the
quality
attributes
that
we
wanted
like
formalized
like
it
needs
to
start
out
the
next,
a
second
so
whatever.
A
C
You
know
we're
looking
at
the
web
IDE
now
right
and
we've
got
all
of
these
features
that
we
want
to
go
do
and
then
we
want
to
start
bringing
in
live
coding
right,
like
that's
kind
of
one
of
the
next
big
milestone
things
that
we
want
to
do.
Is
we
want
to
make
a
place
where
to
developers
connect
to
the
same
web
IDE
at
the
same
time
and
can
and
can
do
those
things,
and
so
you
know
making
sure
the
foundational
pieces
of
the
web
ID
are.
There
are
really
important.
C
C
One
of
the
other
things
that
we've
talked
about
you
know
the
James
has
brought
up
before
is
sometimes
you
have
to
go
back
and
patch
older
versions
of
your
app
right
and
well
now
we're
on
oh
well,
my
local
environments,
five
Ruby,
Ruby
versions
ahead.
The
gem
files
are
all
different,
like
yeah
sure
I
could
figure
out
and
spend
a
day
resolving
this
to
get
an
old.
C
Are
we
in
a
place
now,
where
we're
kind
of
not
in
the
right
spot,
to
go
and
do
some
of
the
long
term
things
that
we
want
to
go?
Do
you
know
and
to
be
fair,
that
we
hadn't
necessarily
thought
of
the
use
cases
or
hadn't
well
mapped
out
so
I?
Don't
it's
not
it's
not
a
fault
of
anyone.
It's
just.
It
is
what
it
is
based
on.
C
The
state
to
get
get
to
the
shippable
product,
so
you
know
those
are
the
things
that
I'm
thinking
about
when
I'm
thinking
about
where
we
want
to
go
and
I
think
this
was
a
good
for
me,
a
good
level
set
when
when
call
him
yep
and
he's
like
hey.
This
is
what
you
said.
It
doesn't
quite
work
like
that.
It
was
like.
Oh
well,
I'd,
be
good
to
know,
oh,
and
so
so
those
are
the
things
I'm
thinking
about
the
other
thing
that
is
in
the
doc
that
we're
we're
thinking
about
is.
C
Are
there
other
editors
like?
Could
we
rip
and
replace
this
and
not
build
all
of
this
functionality
from
scratch?
Could
we
go
find
something
that's
more
complete
and
do
that?
And
so
you
know,
that's
the
other.
The
other
big
question
and
I
think
it's
important
to
understand
where
we
want
to
be
when
we
think
about.
If
we
want
to
go,
invest
on
the
rip
and
replace,
because
is
one
of
these
other
things
actually
going
to
enable
us
to
do
that?
F
So
Kai
I,
just
like
that
one
other
thing
in
there,
so
I,
don't
know
how
mentioned
this
happened
on
the
live
coding
part,
but
what
the
technology
we're
thinking
or
at
least
I
could
see
being
very
useful.
There
is
the
way
barred
WebRTC,
which
doesn't
necessarily
requires
an
intermediate
step
on
the
server.
C
And
those
are
the
things
that,
like
yeah
I,
don't
have
a
clue.
How
we'll
do
that?
That's?
Why
you
guys
are
all
much
smarter
than
I
am,
but,
but
those
are
the
things
that,
like
I,
think
we
need
to
be
taking
it
at
this
point,
I
feel
like
we're
in
a
place
where
we
need
to
be
taking
these
things
into
account
as
we
move
forward.
B
Yeah
I
have
a
point
here,
because
I've
been
thinking
about
these
additional
features
we
are
going
to
be
in
the
future
and
I
mean,
for
example,
something
really
easy
that
we
give
a
lot
of
body
to
everybody
would
be
linking.
For
example,
it
doesn't
make
sense
that
you
have
to
open
the
terminal
for
half
for
having
the
limiting
functionality
there.
So
I
think
we
are
attaching
too
many
features
to
the
fact
of
having
the
terminal,
so
I
I
think
we
should
think
about
the
terminal.
B
B
So
if
we
change
this
fact
a
little
bit,
if
you
know
when
you
study
body,
it
requires
a
couple
of
seconds
to
load
a
whole
environment.
So
if,
at
the
same
time,
we
trigger
some
enmity,
some
helper
images
or
some
additional
services,
for
example,
we
can
boost
their
the
terminal
or
the
main
build
image.
At
the
same
time
we
can
we
can.
We
can
trigger
the
preview
image.
We
can
serve
that
start
time
that
peel
that
container
start
time
with
the
way
body
loading
time.
B
Okay,
what
is
the
problem
here
that
we
are
going
to
start
a
lot
of
peels
with
every
new
way
body?
So,
if
you
are
going
to
open
your
weapon
is
in
different
tabs,
you
are
going
to
start
different
and
different
to
are
going
to
start
different
containers.
Well,
maybe
we
can
feel
it
out
how
to
serve
these
containers.
If
you
are
the
same
user
in
the
same
project,
I
mean
we
can
figure
something
out,
but
I
think
this.
A
Yeah,
you
know
saying
either
way
like
yeah.
This
still
just
the
cost
is
makes
me
feel
a
little
little
shaky,
but
I
think
it's
a
really
good
idea
and
I
love
that
yeah
you're
right.
Why
are
we
coupling
now?
You
have
a
full
integrated
environment
once
open.
This
Web
terminal
may
be
really
neat
if
gonna
be
enhance
the
usability
of
the
whole
web
IDE.
A
If
we
actually
provided
the
environment-
and
it
was
a
configurable
environment
for
admins
on
start,
where
yeah
I,
don't
I,
don't
need
to
type
anything
a
terminal,
but
yeah,
please
lend
my
stuff
for
me
would
be
great
I.
Think
it's
a
really
cool
idea,
but
I'm
I,
don't
know.
If
I
mean
this
kind
of
work,
we
can
either
do
on
the
server
side
or
on
the
client
side.
If
we
do
it
on
the
server
side,
I
wish
I
knew
I
wish.
A
We
did
how
much
well
how
how
much
dollars
are
we
spending
for
someone
to
open
up
the
web,
ID
and
web
terminal
for
30
minutes
and,
if
that's,
not
economical,
we
gotta
figure
out
a
way
it
gives
on
the
client
side
I
feel
like.
But
if
we
don't
have
any
answer
that
question
I'm
concerned
with
us
going
too
far
and
I'm
realizing
we've
made
a
non
economical
decision
and
and
maybe
engineers
we
should
be
able
to
do
the
math
there
I.
C
Guess
it
would
depend
on
so
like
if,
from
a
user
perspective
and
from
like
a
financial
model
right
like
if,
if
we're
burning,
if
the
web
IDE
appropriately
burns,
CI
minutes
so
uses
the
CI
minutes
that
everyone
gets
for
free,
then
in
theory
were
covered.
I
actually,
don't
know
if
those
are
instrumented
correctly
too.
You
see
like,
even
though
we're
using
CI
functionality.
I,
don't
know
if
minutes
in
the
web
IDE
especially
like.
A
I
think
it's
a
good,
listen
as
a
good
perspective.
Seeing
like
is
it
that
quick
in
lab
calm
or
good
lab
us,
but
if
it's
also
not
economical
for
the
users
and
because
this
is
an
ultimate
only
feature
and
right
now
the
file
syncing
only
works.
As
someone
does
does
the
file
syncing
only
works
its
own
sets
up
their
own
kubernetes
runner.
If
it's
not
economical
for
the
user,
then
it's
then
then,
we've
also
kind
of
but
I'm
not
gonna,
go
so
far
and
say
that
we
have
to
build
something.
That's
valuable.
A
C
C
You
can
have
ten
hours
a
month,
a
free,
complete
web
IDE
time
and
then
beyond
ten
hours
you
gotta
start
buying
and
whatever
in
ten
hour
blocks
or
something
for
for
very
little
money
right,
I
think
I
think
there's
a
a
business
model
right
on
the
calm
side.
That's
there
I
think
for
self
hosted,
I'm
slightly
less
I.
Think
there's
an
initial
concern
about
self
hosted
right
because
she's
so
our
self-manage,
because
some
people
don't
necessarily
attach
kubernetes
clusters
they're,
not
using
runners.
G
A
You
think
do
you
think
it
makes
sense
to
maybe
I'm
just
imagining.
If
we
could
do
anything
resume
needs
to
split
up.
You
have
CI
minutes
and
then
you
have
web
ID
environments
minutes.
My
people
should
be
able
to
use
client-side
web
ID
like
for
free.
But
if
we're
gonna
do
web
ID
environments,
you
think
it
makes
sense
to
split
up
I.
Guess
a
users
accident
aquarium
through
their
CI
minutes
and
now
you
know,
I
can't
merge,
merge
requests
because
my
bills
aren't
goin
serially
I,
guess
you'd
have
being
a
huge
frustration
for
users,
yeah.
C
I
think
from
like
an
instrument
and
measurement
perspective,
right,
I
think
users
would
prefer
it
like
they
had
2000
CI
minutes,
and
then
we
say,
like
you,
get
X
number
of
complete
web
IDE
minutes
right,
I!
Think
on
the
business
side,
that's
I'm,
I'm
less
concerned
about
solving
it
now,
because
it
is
most
of
these
things
sitting
ultimate
only
anyways
and
so
therefore
they're
paid
for
so
not
necessarily
a
big
deal
about
how
we
deal
was
with
it
right
now.
C
I
think
CI
minutes
exist
for
everyone
on
free
plans,
but
we
capped
it
somewhere
because
it's
right
there
has
to
be
some
kind
of
resource
usage
mitigation
there
just
to
deal
with
some
of
that,
and
so
I
think.
That's
to
me.
That's
where,
like
we
would
get
in
a
in
a
long
term
state,
but
it's
I
think
the
feature
set
right
now
is
still
kind
of
we're
still
working
through
the
kinks
of
the
feature
set,
which
is
good
and
bad.
C
A
A
How
can
we
get
there
to
make
sure
that
we're
talking
about
and
I'm
I
wanted
to
just
put
the
question
out
here:
do
we
think
that
do
we
do
we
feel
like
we're
on
the
right
path,
or
do
we
feel
like
there's
some
adjustments
that
need
to
be
made
and
maybe
even
communicate
higher
up
yeah
I?
Don't
I,
don't.
C
Know
what
what
everyone
thinks
I
think
that's
my
question
for,
for
you
in
fron
is
like.
If
you
go
and
I
think-
and
maybe
you
need
the
time
to
digest
it
I
think
that's
fine
I
think
the
live
coding
strategy
was
recently
updated
and
the
web
IDE
strategy
was
recently
updated
just
to
kind
of
reflect
more
of
what
we're
thinking
about
further
out.
C
The
other
the
other
pieces
here
are
we
on
the
right
track,
or
would
you
if
you
could
start
over
today,
knowing
everything
that
you
know,
would
we
go
do
something
different?
Would
you
pick
a
different
underlying
architecture?
You
know,
if
you
could
do
anything
you
wanted,
would
we
go
do
it
differently
and
I
think
that's
the
question
we
should
try
and
get
an
answer
to
like.
Would
we
do
it
differently
and
would
that
enable
these
other
things
that
we
want
to
go?
C
A
A
Look
at
our
console
in
terminal
stuff,
I,
look
at
competing
products
that
granted
they're
just
hundred
percent
focused
on
that
product
right
I
mean
we've
hit
some
functional
checkboxes,
where
we
can
on
a
marketing
page,
say:
look
what
we've
got
but
there's
some
there's
some
significant
quality
check
boxes
that
don't
go
and
notice
that
are
better
missing
there
and
unbaked
and
if
we
think
those
quality,
actually
it
so
important.
I
do
think
yeah
we're
gonna
have
to
rethink
our
strategy.
A
A
G
C
No
I
so
I
think
yeah
I
mean
that's
the
last
item.
Here's
right
to
see
ya,
I'd,
be
proof
of
concept
and
so
I
think
one
of
the
things
and
Mark
is
asking
for
this
mark
on
sex,
so
I
think
one
of
the
things
we'll
do
is
we'll
spend
some
time
in
twelve
to
some.
People
need
to
spend
some
time
figuring
out.
If
this
helps
us
get
where
we
want
to
go
faster
and
is
better
or
if
we
could
build
all
of
these
things,
and
it
would
be
the
same.
C
A
If
that
might
be
the
kind
of
architecture,
that's
that
would
be
useful
for
us
kind
of
what
ron
was
talking
about.
Why
do
we
have
to
start
because,
right
now
and
even
for
life
preview,
what
we're
planning
on
you'll
have
to
start
the
terminal,
because
that
kind
of
kicks
off
the
environment
for
things
that
happened
it'd
be
really
neat.
A
If
we
allowed,
we
always
provided
the
the
full
client-side
version
of
the
editor,
but
if
an
admin
enabled
it
there's
the
server-side
version,
the
editor,
where
yeah
we
spin
up
a
docker
container
environment
for
every
editor
and
that's
environments
kicked
off
and
that
would
decouples
from
CI
and
we
could
figure
out.
Maybe
how
or
maybe
we're
still
borrowing
some
function
ID
from
there,
but
I
think
that
we
could
borrow
maybe
some
of
their
ideas,
if
not
their
solution.
A
It's
be
interesting
and
I
like
a
lot
of
fronts.
Talking
about
of
yeah.
It's
weird
right
now,
everything's
triggered
by
this
we're
starting
the
web
terminal
from
a
user's
perspective.
I,
don't
think
that
would
make
probably
a
lot
of
sense
and
I,
really
like
maybe
more
cell
or
for
UX,
to
look
at
this
list
of
weaknesses
and
maybe
look
at
what
is
our
what's
our
order.
Our
actual
requirements
here
would
be
really
helpful.
B
I
want
to
point
out
something:
I
mean
I,
like
the
Creator
approach.
The
client
approach
to
this,
but
the
main
the
main
thing
for
this
CIE
workflow
is
to
have
a
sandbox
of
the
repository.
If
we
use
a
client
setup
Roche,
how
do
we
perform
all
these
operations?
All
these
new
functionalities?
Are
we
going
to
build
a
map
only
in
the
client
side?
Are
we
used
to
give
that
API
and
where
are
going?
Those
API
calls
point
to
to
the
real
repository.
C
But
here's
a
crazy
architecture
thought
for
you.
What
if
you
require
docker
to
be
available
on
the
person's
computer
and
we
docker
containers
on
their
compute
instead
of
on
an
attached,
kubernetes
cluster
and
the
web
idea
was
a
portal.
So,
like
look
there's
a
bunch
of
ways
that
we
could
solve
for
like
if
we
need
docker
compute,
if
we
need
containerized
compute
somewhere,
you
know
there
are
ways
to
do
this
and
data
centers
there's
a
way
to
do
it
and
these
clusters
there's
ways
to
do
this.
C
I
think
those
are
the
things
that
we
need
to
think
about
in
terms
of
like
what
are
the
best
ways
forward
to
get
us
the
most
stuff
and
I
think
we
just
need
to
be
more
creative
to
make
sure
that
we're
accounting
for
longer
term,
where
we
want
to
be
and
I
would
say,
like
take
a
look
at
the
strategy.
Docs
I
think
this
was
helpful
to
get
a
good
understanding.
C
So
I
think
people
understand
there's
some
some
issues
around
where
the
web
ID
and
the
terminal
stuff
sits
today
and
where
how
these
compute
things
are
happening,
I
think
and
I
think
you
know
we
should
spend
some
time
in
12.
I
think
we're
gonna
do
a
little
bit
of
technical
research
versus
actual
delivery
and
I.
Think
that's
fine
I
think
we
need
to
make
sure
that
we're
all
on
the
same
page,
moving
forward
about
where
we
want
to
go.
A
A
If
there's
an
architecture,
we
could
offload
this
on
the
client-side
I
think
that
they'd
prefer
that,
because
they'd
have
start
a
run
up
time
and
that'd
be
easier
connection
for
them,
and
we
prefer
that,
because
it's
cheaper
for
us
and
it
also
let
us
probably
lower
the
barrier
of
entry,
which
kind
of
kicked
off
this
whole
discussion.
So
little
thanks,
everybody
and
I
hope
this
is
helpful
and
I
know
we
didn't
get
to
create.