►
From YouTube: Meshery Development Meeting (Mar 3rd, 2021)
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).
A
A
As
a
there
were
a
couple
of
other,
so
we've
got
two
topics
I
suspect
that'll.
Take
us
not
all
of
our
time,
and
there
are
more
things
to
talk
about.
One
of
those
things
that
I
was
considering
we
might
chat
through
is
well
has
to
do
with
some
work
that
I
think
jaguarity.
A
A
B
A
A
C
We
should
regulate
and
make
it
make
it
more
presentable
so
like
when
we
created
all
they
are
pretty
good,
looking
apis
right.
C
So
the
first
step
which
we
have
done
over
here
is:
I
have
jot
down
every
apis,
which
we
have
in
misery
most
of
the
eight
years,
and
we
have.
We
have
segregated
them
by
categories.
As
you
can
see,
there
are
a
few
things
which
are
still
to
be
done,
like
the
descriptions
and
the
parameters
and
how
they
are
passed
and
stuff
like
this.
This
will
probably
be
used
when
we
are
trying
to
document
it
by
go
swagger
and
for
now
what
I
wanted
to
discuss
was.
C
There
are
a
few
changes
which
we
need
to
make
when
we
look
at
the
overview
of
how
api
points
are
like
one
of
the
things
could
be
like
whether
we
want
to
use
camel
case
or
kebab
case,
because
in
few
of
the
cases
we
are
using
camera
cases
and
few
of
the
cases
we
are
thinking
about
to
make
it
more
cons,
it
makes
sense
to
make
it
more
consistent.
So
there
are
a
few
things
like
that
and
and
general.
There
are
few
endpoints
which
are
serving
the
same
things
like
api.
C
Results
are
basically
used
for
fetching
data,
but
one
is
used
for
getting
us
singular
data
and
one
is
used
for
getting
multiples,
so
that
doesn't
really
make
sense
so
like
finding
this
kind
of
things
in
api
endpoints
and
like
trying
to
improve
them.
So
we
are
open
to
suggestions
on
what
we,
how
we
want
to
change
it,
how
we
want
to
what
we
want
to
change
like,
if
any
of
you
guys
have
work
in
any
of
the
endpoints.
C
It
would
be
greatly
helpful
if
you
could
like
fill
out
the
descriptions
and
how
you
are,
what
kind
of
methods
you
are
using
in
the
parameters
which
how
you
are
passing
them.
If
you
are
passing
them.
That
would
be
greatly
helpful.
If
you
don't,
then
slowly
and
steadily,
this
sheet
will
be
filled
up
eventually,
no
worries
about
that
and
yeah.
We
are
also
planning
to,
as
you
can
see
he
doing
right
over
here.
We
are
planning
to
change
few
of
the
end
points
in
a
better
way,
so
that
they
reflect
the
particular
category.
E
Are
we
trying
to
be
as
rest
as
restful
as
possible.
E
Meaning,
for
instance,
that
we're
sort
of
trying
to
identify
the
resources
that
we're
the
resources
that
we
want
to
handle
and
be
consistent
with
that,
for
instance,
when
I
look,
but
I'm
not
I'm
I'm
just
guessing
right
now.
If
I
look
on
line
34,
I.
E
E
Because
what
I'm
trying
to
sort
of
say,
if,
if
we're
thinking
rest,
then
the
end
point
would
be
or
the
the
resource
path
will
be
just
bored
and
posting
would
import
or
create
a
port
which
would
correspond
to
an
import.
That's
a
bit.
What
I'm
sort
of
trying
to
ask
and
another
one,
for
instance,
is
the
query
range.
C
C
Yeah
and
I'm
seeing
what
we
are
saying
we
we
should
probably
do
that
to
like
putting
it
all
together
in
one
place
where,
if
there's
a
gate,
we
could
get
there,
there's
also
its
own
free
post.
Then
we
would
post
the
results.
Something
like
that
like
this
is
a
complete
jungle
right
over
here,
because
few
of
them
are
restaurant
per
se,
but
few
of
them
are
not
at
all
tasteful
like
they
have
different
end
points
for
for
posting
and
for
deleting
and
getting.
F
E
C
F
C
A
Yeah
this
michael's
spot
on
about
what
kind
of
where
he's
headed,
which
is
well.
I
think
I've
felt
slightly
like
my
ocd
has
felt
so
slightly
out
of
whack
over
the
last
year,
and
maybe
it's
because
of
these
endpoints,
and
it's
just
like
oh
they're,
just
great
on
you
they're,
just
they're,
just
they're,
just
a
collection
of
300
something
contributors
coming
to
work
in
various
areas
at
various
times.
I
don't
know,
is
it
not
that
300
people
have
touched
the
endpoints
but
but
yeah
to
michael's
point
the
it.
A
What
will
really
improve
code
quality
improve
our
conversations
as
we
go
to
refer
to
something
and
call
it
something
our
our
language
will
become
easier
is
is
when
we
start
to
focus
on
a
noun
and
and
then
to
michael's
point
like
restful
things
or
http
verbs
that
you
can
perform
on
that
noun
and
as
we
do
that
that
that
will
encourage
us
to
have
a
cons,
consistent,
endpoints
and
points
that
you
would
expect
you
could
apply
a
filter
to
or
that
you
could
expect
that
if
you
do
a
post
to
this
will
happen.
A
If
you
do
a
put
to
this
will
happen
and
you
you
come
to
expect
like
fair,
a
fair
amount
of
consistency
across
the
methods
and
across
like
the
query
strings
that
are
available
like
just
the
something
for
us
to
be
using
something
like
gorilla
or
mux,
or
any
of
these
frameworks
like
other
than
routing.
I
don't
know
what
advantage
we're
getting
from
them.
A
If
we're
not
like
I'm
ignorant
of
whether
or
not
they
provide
any
additional
standardization
around
things
like
filters
or
provide
a
framework
and
an
encouragement
to
have
consistent
behavior,
they
must,
and
we
must
just
not
be
taking
advantage
of
that
or
having
enforced,
that
in
the
past,
because
yeah
something
like
while
it
it
can
be
bothersome
and
a
bit
mind-bending
to
give
a
little
bit
of
forethought
to
something
like
grafana
prometheus,
and
if
you
keep
going
with
what
other
telemetric
signals
there
are.
A
There
are
any
number
of
others
like
anytime,
there's
a
bunch
of
there's.
This
just
makes
me
wanna.
I
just
wanna
throw
up
all
over
the
spreadsheet.
It's
just
the
reason
being
is
like
you
should
really
never
ever
ever,
even
when
we're
calling
out
kubernetes
down
below
like
kubernetes,
is
it
the
centerpiece?
Is
it
the
rock
the
center
of
gravity
in
the
middle
of
much
of
what
we're
doing
yeah?
A
Is
it?
Does
that
mean
that
things
don't
run
in
vms
that
we
don't
want
to
consider
for
things
that
are
outside
of
kubernetes?
No,
that
we
definitely
want
to
consider
for
things
that
are
outside
of
kubernetes,
like
80
percent
of
the
world's
workloads
are
outside
of
kubernetes
or
90,
so
we
definitely
want
to
you
know:
people
want
to
on
come
to
a
service
mesh
come
to
cloud
native
because
they
want
to
modernize
their
existing
workloads
and
because
they
want
to
create
new
things
anyway.
A
My
point
is
even
though,
even
for
as
center
central
as
something
like
kubernetes
is
to
have
it.
If
you
don't,
if
you
don't
mind,
if
you
scroll
down
to
like
lines
50-ish
or
something
where
mesh
sync
is,
I
think
this
is
what
this
is
this.
This
type
of
an
exercise
requires
one
to
like
between
two
to
three
and
more
can
come
to
bear,
but
it
requires
a
minimum
one
individual
to
go
to
to
wear
an
architect
hat
to
sit
here
and
say.
A
A
Why
is
that
not
a
category
center
and
I'm
not
asking
that
at
dhruv
or
toward
drove
in
any
in
any
way-
and
maybe
it's
fine
like
it
should
be
separate
and
that's
because
we
really
we
favor
short
endpoint
references
over
hierarchical,
categorized
things,
that's
a
statement
that
needs
to
be
made
and
the
guiding
principles
and
like
it's
not
an
objective.
It's
a
go.
It's
a
guiding
principle
to
as
to
the
design
of
something
like
this
so
yeah.
This
thing
is
like
I've
avoided
having
this.
A
For
my
part,
I've
avoided
having
this
conversation
for
a
long
time,
because
it's
like
look
because
it
requires
a
deep
breath,
sit
down,
think
about
it
for
a
bit
and
then
get
well.
This
is
gonna
sound
like
a
bad
kebab
joke,
but
and
then
get
fairly.
Tighten
your
sphincter
a
little
bit
get
get.
Can
I
get
well,
I'm
not
gonna,
say
the
word,
but
be
meticulous
and
scrutinize.
A
How
try
to
make
it
perfect
because
else
else
it
kind
of
doesn't
matter
and
it's
haphazard.
So
for
my
part
to
michael's
point,
I'd
love
to
see
something
like
no
more
are
we.
I
don't
know
that
we
should
ever
reference
prometheus
or
grafana,
it's
easier
when
you
do
when
you
call
out
those
things
explicitly,
but
how
many
tracing
programs
are
there?
I
don't
know:
there's
jaeger,
open
tracing,
zip,
not
open,
tracing,
there's
jaeger,
zipkin,
there's
a
third
one.
I
don't
know
anyway.
A
Are
we
only
ever
going
to
use
prometheus
for
metrics,
or
are
we
going
to
allow
datadog
to
plug
in
and
get
metrics
as
well?
What
about
grafana?
Is
that
the
only
way?
So
why
are
we
calling
these
things
out
explicitly
in
first
class?
Maybe
we
should,
but
that
seems
like
they
would
be
categorized
under
a
slash,
telemetry
or
we're
fine.
Let's
just
call
it
slash
metrics
if
just
grafana
and
prometheus
are
going
to
be
there
and
then
we'll
have
logs
and
tracing
separate.
G
A
Anyway,
michael
this
sort
of
this
is
this
is
one
of
those
areas
that
you
are
that
speak
to
some
of
your
your
many
strengths
dhruv.
This
is
great.
Thank
you
for
doing
this.
This
is
the
first
time
I've
ever
seen.
A
The
collection
put
together,
someone
to
have
organized
them
and
said,
like
I
had
no
idea
what
methods
were
supported
against
what
endpoints
and
to
michael's
point
like
it's
egregiously
wrong
that
we
would
have
something
like
slash,
prometheus,
slash,
query
range,
and
I
was
there
when
some
of
the
first
ones
were
put
in
where
it
was
like
slash,
slash,
perf,
slash,
result
and
then
slash
perf,
slash
results
like
I.
That
was
the
dumbest
thing
when
we
put
it
in
and
nobody
none
of
you
were
here
when
that
happened.
A
I
don't
remember
the
justification,
but
we've
been
doing
you
all
have
been
doing
a
ton
of
tech
debt
cleanup
michael
just
published
a
post
on
that
was
basically
like
here's
how
he
and
a
collection
of
the
rest
of
you
went
off
and
totally
rewrote
the
adapters,
and
I
sleep
better
at
night
because
we're
using
and
we're
using
a
library
for
it
anyway.
Point
is
same
kind
of
thing
needs
to
happen
here
and
it's
a
it's
a
beautiful
thing,
it'll
feel
really
good
and
it
doesn't
it.
I'm
expressing
all
this
exasperation
about.
A
You
know
the
the
way
that
it's
designed
and
it's
actually
such
a
sweet
opportunity,
it's
kind
of
a
neat
way
to
get
in
and
wear
an
architect
hat
kind
of
design.
A
few
things
engage
with
some
others
in
discourse
about
like
there's,
multiple
right
answers
here,
yeah,
let's
we
are
going
to
get
to
logs,
like
we've.
Almost
vanith
had
demoed
to
us
a
ui
a
little
while
ago.
A
That
was
showing
like
the
ability
in
mesherie
to
have
a
tail
a
tail
of
logs
like
a
live
tail
of
logs
coming
from
the
meshes
and
the
services.
A
So
we're
probably
going
to
deal
with
logs,
like
that's
that
makes
sense,
we'll
probably
like
we
should
absolutely
be
dealing
with
traces,
so
we
can
categorize
things
under
this.
You
know
slash,
telemetry,
category
or
or
not.
You
know
like,
like
it's
correct,
either
way,
there's
a
couple
of
things
that
I
think
that
should
be
given
as
some
some
rules
in
part,
two
I'll.
Take
it
one
step
further
from
what
michael
was
saying-
and
that
is
you
know,
rules
like.
A
E
A
Quality
yeah
yeah,
there's
a
you
know,
quality
is
consistency,
even
if
your
consistency
is
kind
of
kind
of
crap.
Still
it's
still
quality
crap,
because
it's
the
same
crap
everywhere.
D
A
A
A
But
is
that
most
of
the
software
you
look
at
you
don't
have
like
a
v1
api
api
v1.
That
was
there
from
the
start
and
has
only
ever
been
there
for
the
last
10
years.
It's
like
like
very
rarely
if
I've
seen
you
know,
sometimes
you'll
see
them
move
to
a
v2.
E
No,
I
think
it
would
be
because
there
are
like
three
or
four
ways
to
version
an
api,
and
all
of
them
are
wrong
or
difficult
like
having
it
in
the
in
the
in
the
path
which
is
easy
to
understand,
or
you
can
have
it
in
the
in
the
header
or
you
can
have
it
in
the
in
the
in
the
what's
it
called
the
content
type.
What
is
it
called
schema
if
you
do
it
sort
of
proper,
rest
and
stuff?
E
But
I
personally,
I
think,
having
it
in
the
in
the
path
is
so
easy
to
understand.
E
E
E
C
On,
like
the
action
item
over
here,
which
I'm
getting
from
your
conversation
is,
maybe
first
define
a
few
few
things
which
you
would
consider
while
constructing
the
apis.
You
could
like
look
at
few
of
the
good
apis
implementation
and
like
define
that
now
we
will
use
burps
like
we
won't
use
crude
rules
itself,
and
once
we
decide,
then
maybe
you
could
go
through
each
of
these
sections
one
by
one
and
decide,
particularly
on
that
how
we
want
to
reach
it
like
we.
C
D
E
E
Because
in
pattern
file
we
know
what
the
resource
I
will
am.
I
I
think
we
know
what
the
resource
is.
The
resource
is
a
pattern
file,
that's
you
know,
and
then
how
do
we
sort
of
implement
the
functionality
we'd
like
to
that
is
listed
in
sort
of
like
represented
by
the
the
parts?
On
the
right
hand,
side?
E
I
can
think,
how
can
we
is
this
restful
or
is
it
not
and
if
we
want
to
make
it
really
restful?
How
should
it
look
like
I
mean
an
import
would
be
a
post
post
and
you
know
I'm
just
making
up
right
now
an
import
would
be
in
post
and
an
export
would
be
a
get,
and
then
where
do
you
define
the
the
format
like
whether
you
want
to
have
a
site
to
escape
js
as
format.
F
E
In
yeah
and
then,
then
you
would
use
the
what
is
it
called
the
media
type
right
or
the
the
content
type
and
the
accept
type?
That's
maybe
what
I'm
trying
to
if
you
want
to
be
really
restful
or
you
would.
E
E
C
Yeah
yeah
that
sounds
good
like
oh,
I
can.
I
don't
have
much
idea
way
to
look
for
good,
good
inspiration
for
ap
endpoints.
Like
do
you
guys
know
anywhere
where
there
are
certain
there
are
certain
set
of
rules,
people
use
for
their
finding
apis
or.
E
E
We
were
not
quite
successful,
I
think,
but
this
is
not
public,
but
I'm
sure
I
find
a
public
contact
we
can
use
or
if
someone
else
has
one,
but
I
can
have.
I
think,
if
that's
okay,
I
can
get
back
to
you.
E
E
But
I
I
can
have
refresh
my
memory
and
look
at
the
resources
I
used
when
we
did
that
a
couple
of
years
back,
you
can
share
it
with
anyone
on
there
on
slack
or.
D
F
A
It's
yeah
definitely
a
design
document.
E
A
A
We
can
continue
to
perp.
This
is
gonna
sound
bad,
but
we
can
continue
to
perpetuate
the
problem.
A
It's
and
it's-
and
it's
mostly
fine,
like
you,
know,
there's
issues
with
it,
but
but
when
we
do
go
to
address
it,
yep
hey
we'll,
want
to
have
you
know,
given
a
lot
of
thought
and
have
a
well-established
system
with
you
know
design
system
with
some
kind
of
some
rules
if
you
will
or
some
formatting
rules
or
what
have
you
and
so
yeah
the
better
to
take
some
time
with
it
and
restructure
it's
one
of
those
things.
A
In
last
week's
meeting
minutes
we
have
sort
of
laid
out
high
level
milestones
for
what
the
next
set
of
releases
might
be
and
I'll
need
to
go.
Look
in
the
meeting
minutes,
but
this
is
one
of
those
things.
It's
like
a
pre
1.0
that
we
would
want
to.
A
You
know
that
we'd
want
to
get
to.
We
want
to
stabilize
on
the
refactoring
of
these
by
then
some
of
the
way
that
I'll
do
so,
michael's
probably
got
you
know
a
better
response
about
like
or
more
a
better
response
about
it,
the
design
methodologies
and
how
you,
you
know
how
you
approach,
one
of
the
quick
quips
that
I'll
that
I
would
I
would
offer
is
well.
It
is
for
me
it's
it's.
A
Not
it's
not
a
lot
different
than
how
you
might
think
about
designing
classes
like
in
an
object-oriented
piece
of
software,
that
you
create
a
class
about
or
the
way
that
you
design,
sql
tables
to
be
in,
like
third,
normal
form
or
something
that
you
would
with
those
classes
or
those
objects
or
those
tables
and
the
rest
of
the
endpoints
these
endpoints.
A
What
are
what
are?
What
are
they
comprised
of
resources,
so
a
resource,
a
class,
an
object
like
those
are
all
about
synonymous
with
one
another
and,
in
my
mind,
they're
really
all
about
to
use
a
grammatical
word
to
describe
them,
they're,
all
about
a
noun,
a
resource,
a
noun.
A
So
if
you
have
a
table,
that's
a
that's
a
resource!
That's
an
endpoint!
There's
a
table,
that's
a
class
right!
There's
it
has,
it
has
has
legs.
Maybe
legs
are
an
attribute
of
it
or
maybe
legs
are
its
own
end
point
as
well.
You
want
to
create
new
legs
or
delete
it
like.
Basically,
anything
you
want
to
do.
A
Crud
over
those
types
of
crud
operations
are
very
well
aligned
to
restful,
http,
verbs
and
so
yeah
to
have
things
like
so
in
under
grafana,
like
there's
grafana
as
a
resource
that
we
are
measuring
is
connected
to.
So
it
does
make
sense
that
there
is
a
configuration
like
there
may
be.
An
api
part
of
that
configuration
is
what's
its
endpoint.
A
What's
this
api
key
and
how
many
of
these
instances
am
I
connected
to
okay,
that
kind
of
makes
sense
that
maybe
there's
such
a
thing
called
config,
or
maybe
it's
just
called
grafana
like
and
graph
the
slash
api
cophana.
That
is
the
config.
That
is
the
thing
the
entity
when
you
want
to
do
a
a
get
on
it,
you
either
get
back
the
config
or
you
get
a
ping,
and
maybe
you
get
both.
It
would
probably
be
you'd
get
a
config,
and
when
you
do
a
post
you
would
create
or
update
that
config.
A
When
you
do
a
put
you
would
that
actually,
I'm
confusing
the
the
verbs
here
I'd
actually
say
it
differently
when
we
do
and
when
you
do
a
put,
maybe
you're
actually
doing
a
ping
at
that
point
like
like
the
fact
that
we
have
verbs
called
out
kind
of
scan
and
ping
and
and
then
this
range
and
stuff
is
or
this
query
is
like
you
know,
I
don't
you,
don't
have
to
have
a
like
the
the
fact.
Typically,
the
fact
that
you're
calling
an
endpoint
often
is
a
query
unto
its
own.
A
It's
more
about
find
and
replace
you
know
in
in
the
code
and
and
yeah
when
things
start
to
get
described
and
swagger
part
of
where,
where
michael
almost
touched
on,
or
was
it
kind
of
did,
was
in
the
mystery
architecture,
deck
there's
a
that's,
probably
the
the
that's,
the
best
collection
of
what
would
be
our
logical
object
model,
which
is
to
say
there
is
such
a
noun
or
there
is
such
an
object
within
measuring
called
grafana
there's
another
one
called
like
what
are
these
resources?
What
are
these
nouns?
What
are
these
endpoints?
A
What
are
they
like?
They're
all
whether
you're
modeling
it
visually
so
that
little
brains
like
mine
can
understand
that,
like
it's
okay
for
meshri
to
connect
to
one
or
more
grafanas,
I
love
to
see
that
visually
laid
out.
Because
just
you
know
it
helps
like
okay,
that
one's
talking
to
that
one.
And
it's
okay,
that
it's
talking
to
more
than
one
of
those
great
and
then
that
directly
maps.
To
some.
You
know
an
exercise
like
this,
which
directly
maps
to
something
like
a
you
know:
swagger
spec,
which
then
hopefully
maps
to
importing
into
guerrilla.
A
You
know
gorilla
or
whatever,
we're
using
and
but
those
same
artifacts
are
closely
tied
together,
which
actually
then
closely
like
identically
maps
to
sql
design
language,
where
you
take
that
object
model,
and
now
it's
actual
sql
statements
that
create
the
the
tables.
And
so
maybe
this
is
just
the
way
that
my
mind
works
but,
like
I
was
saying,
although
and
then
that
directly
maps
to
end
points
behind
which
you
have
objects
and
classes
that,
like
these
things,
all
interconnected
across
the.
A
Spectrum,
nice
I'll
be
quiet,
it
sounds
like
michael
and
drew
you
guys
were
on
a
roll
there
with.
You
know
some
next
steps
and
things.
E
A
Michael
was
just
being
polite
and
saying:
hey,
would
you
would
you
like
to
hold
hands
for?
Would
you
like
to
partner
up
and
sort
of
yeah,
definitely
yeah.
C
Oh
yeah,
so
the
first
thing
I
should
first
action
item
should
be
to
have
a
good
define
our
methodology,
which
we
would
be
using
and
then
implement
in
one
of
the
easier
sections
like
patent
file
and
then
move
on.
C
D
E
A
A
Is
alonso
still
on
oh
yeah,
so
I'm
sensitive
to
the
fact
that
alonso,
you
know
alonso
and
a
couple
of
others,
but
alonso
specifically,
like
he's
put
in
a
lot
of
work
here
and
he's
done
something:
that's
not
easy.
Basically,
you
know
what
the
work
that
he's
done
is
like
touches
all
of
the
documentation
pages.
It
touches
it's
a
change
to
the
entire
docs
site
and
it's
a
massive.
You
know
it's
an
uplifting
and
positions
us.
It
changes
the
platform.
A
It
positions
us
for
multiple
languages
that
the
larger
the
pr
the
longer
that
it
takes
to
land
is
sort
of
a
natural
thing,
one
one
twist
that
I'd
like
to
it
doesn't
have
to
be
thrown
in
and
totally
considered
for,
or
rather
it
doesn't
have
to
be
accounted
for
in
this
pr.
A
But
but
please
do
give
this
some
thought
a
little
bit
of
consideration.
Chewbunk
now
that
you've
been
involuntarily
volunteered
here,
chewbacca
and
alonso
is,
as
you
set
out
or
as
you
kind
of
conclude,
on
a
system
of
multiple
languages.
There's
another
system
that
we've
been
missing
and
anuru
jain
has
given
thought
to
this.
Some
shriti
chandra
has
given
a
thought
to
this.
A
It's
versioning
of
the
docs
today.
It's
just
assumed
that
you're
running
the
latest
measuring,
which
is
a
fine
assumption
like
if
you're
not
running
the
latest
measure
you
should
be.
We
don't
want
to
spend
time
talking
about
older
releases
at
some
point,
though,
hopefully
it
becomes
important
to
people
who've
been
using
meshri
as
a
tool
that
they
rely
on
to
help
them
run
this
infrastructure
they're
going
to
be
concerned
about
what
the
differences
are,
especially
the
differences
in
behavior
are
between
one
mesh
version
and
the
next
so
long
story
short.
A
A
So
my
the
thinking
is-
or
my
my
perspective-
is
that,
when
multiple
version
that
it's
it's
actually
good
to
do
multiple
languages,
because
that's
a
subset
of
a
concern
of
multiple
versions,
because
would
you
version
the
the
the
different
language
pages?
A
Okay,
as
we
go
to
do
this
next
release,
we'll
snapshot
the
the
current
build
of
the
docs
and
just
and
just
save
them
off
into
a
different
folder,
so
that
when
people
come
here
and
they
want
to
go
to
a
certain
version
of
the
docs,
it's
just
they're,
either
implicitly
on
latest.
Whenever
they
you
know,
come
and
then
the
switcher.
Much
like
the
language
switcher
that
you
guys
were
looking
at,
there
would
be
a
version
switcher.
That
is
a
top
level
concern.
H
I
I
don't
know
if,
if
is
the
right
time
for
but
but
I
have
to
do
to
make
a
question
about
this
documentation
site.
That
is
the
permalinks
that
I
I
always
thought
of
the
permalinks,
also
as
a
unique
kind
of
identifier.
H
H
A
Yeah,
I
think
the
jekyll
yeah
it's
a
little
bit
of
a
conundrum,
because
it's
true
that.
A
One
way
of
solving
the
issue
is
that
we
just
don't
use
permalinks
like
the
the
and,
if
you
don't
use
permalinks,
that
means
that
the
the
structure,
the
folder
structure,
in
which
you
lay
out
the
the
files
on
disk
inside
the
inside
the
docs
folder
itself,
that
if
you
don't
use
permalinks,
that
folder
structure,
it
becomes
a
direct
representation
of
like
the
higher
on
the
information
architecture
that
you
have,
the
hierarchy
that
you
have
in
which,
like
you,
have
to
have
a
concept
folder
under
which
you
would
have
a
an
architecture
folder
under
which
you
would
have,
and
so
we
just
have
to
be
strict
about
adhering
to
that
kind
of
a
layout.
A
If
you
don't,
which-
which
I
mean
we
we
halfway
get
there,
but
but
everything
doesn't
always
fit
cleanly
and
neatly
as
to
how
you'd
like
to
represent
it
here.
So
we
do
use
permalinks
to
help
us
directly.
Get
to
you
know.
You
know
get
from
one
place
to
the
next.
That
the
problem
is.
The
challenge
is
that
when
you
look
at
a
given
page
and
and
here's
a
permalink,
it's
slash
concept
architecture.
A
A
What's
what's
it
called
not
ruby,
but
the
jekyll
macros.
A
Like
liquid
templating
up
in
the
front
matter,
I've
tried
to
do
it
a
bunch
of
different
times
for
different
things.
It
might
be
possible
yeah
it
it's
like
well,
okay!
Well,
if
that's
either
that
is
possible
or
we
have.
We
have
to
write
our
own
liquid
tag
or
find
every
page
has
implicit
javascript
that
looks
at
it
says:
hey.
What
is
this
person
looking
at?
What
language
is
this
person
looking
at
and
by
a
person
looking
at
that
might
just
be:
what's
the
current
query
string?
A
Does
it
equal,
yes
or
whatever
the
language
identifier
is
fine,
great,
make
sure
that
we
in
javascript,
like
redirect
the
window?
Some
solution
like
this,
where
you're
saying
redirect
the
window
to
es
slash
whatever
the
permalink
is.
A
I
guess
what
I'm,
what
I'm
getting
to
is
like
yeah,
it
actually
kind
of
is
an
issue,
because
the
the
site
itself
will
kind
of
override
the
language
selector.
A
Because
the
permalinks
assume
a
single
language,
the
default
language,
in
this
case
being
english,
and
so
it
sort
of
assumes
that
the
user
is
navigating
to
an
english
english
page.
This
is
there's
about
four
different
ways
to
like
address
this
issue.
One
don't
use
permalinks
two
have
a
javascript
redirect
after
this
like,
I
was
just
saying
three
get
liquid
templating
going
on
up
in
here,
four
just
publish
another
copy
of
the
site
like
like,
like
just
like.
We
were
talking
about
with
the
version
numbers
where
we
have
like.
A
We
don't
need
to
try.
We
don't
have
to
have
a
copy
of
the
docs
for
every
single
little
point,
release
that
we
make.
Rather,
we
need
about
five
copies
of
the
docs
right
now,
one
for
the
v
zero
one,
zero
v,
zero,
two
zero
v,
zero
for
the
minor
releases
like
the
more
significant
ones.
So
we
would
about
we
about.
We
would
just
soon
we
would
have
a
fifth
copy
of
the
dogs.
A
Well,
as
part
of
that
I
mean,
storage
is
cheap
and
this
is
not
a
big
site.
A
So
you
could
also
have
like,
like
when
someone
navigates
to
docs
dot
messages,
dot,
io,
slash,
yes
or
slash
whatever
that
you
know
slash
language,
that
the
the
site,
the
jekyll
the
site,
when
it's
statically
built
that
it
considers
it's
that
all
of
the
pages
there
that
they're
all
they
all
consider
that
the
root
is
stops
at
the
language
barrier
such
that
like
we
don't
we
actually
not
changing
the
permalinks,
it's
just
that
the
the
site
and
the
folders
as
we
publish
them
the
the
base
url.
A
A
They're
always
published
in
cons
in
god,
I'm
missing
the
word.
H
Oh
yeah,
okay,
but
more
of
the
question.
I
don't
know
if,
if
I
don't
get
the
answer
right
but
more
of
the
question,
it
goes
to
that.
If
is
everyone
okay
for
getting
the
permalinks
translated?
Also,
oh,
because
the
there's,
the
there's
the
the
this
language
abbreviation
like
es
or
e
e
in
or
turkish
will
will
be
tr's,
etc,
so
that
that
the
permalink
is
stays
the
same.
H
And
then
this
a
review
share
of
the
language
is
goes
in
the
base
url,
like
you
were
saying
so
chandra
was
talking
about
translating
also
that
permalinks,
but
then,
when
we
will
be
navigating
on
on
the
dark
side.
H
A
Yeah,
I
was
hitting
the
issue
from
slightly
so
right
now,
when
you're,
when
we're
publishing
a
spanish
version,
when
you
click
on
a
page
that
has
a
permalink,
does
it
and
you
navigate
from
page
to
page?
Does
it
stay
within
spanish.
H
H
So
it
has
the
the
header
also
translated
and
the
right
sidebar
too
and
yeah
concepts
also.
A
I'll
I'll
toss
out
a
quick
vote,
I'm
right
now
toss
out
my
quick
suggestion,
which
others
can
vote
on
and
you
guys
don't
have
to
take,
but
my
suggestion
would
be
yeah
just
leave
the
permalink
in
english,
even
though
like
yeah,
it's
not,
as
you
know,
ideal.
For
other
I
mean
other
people
just
navigate
the
site
here.
They're,
like
oh
I'd
like
to
read
more
about
adapters
and
even
though
it's
not
adapted
or
adapted
it
is
in
the
url.
It
doesn't
matter
that
like.
H
Yeah
also,
I
I
second
your
your
vote
as
an
individual
vote
too,
but
but
we
we,
I
don't
know
chihuan
can
shinji
call
also
give
us
a
thought
about
it
and
maybe
take
a
decision
on
it.
Sure.
A
You
know
one
way
that
this
kind
of
thing
gets
solved
is
like
is
to
go
back
to
those
individuals
and
say
yes,
it
is
so
long
as
you'll
provide
a
solution.
Well,
we'll
get
that
you
know,
let's
get
that
done
if,
if
they
aren't
gonna,
if
they
don't
provide
a
solution,
then,
like.
F
Yeah,
I
think
we
should
go
with
the
non-translated
version
of
permalink
now
that
we
have
discussed
the
problem
and
it
sort
of
makes
sense
for
us
to
keep
it
the
english
language
itself,
because
the
function
of
permalink
is
just
to
like
preview,
the
structure
of
the
website,
instead
of
focusing
on
the
keywords
of
it
right.
So
now
that
I
see
the
demo
and
the
problem
and
I
feel
like
we
should
keep
the
permalink
intact.
So
that's
like
my
word
sort
of
thing
yeah.
We
should
keep
it
in
english.
A
But
I
can
see
where
you
guys
were
doing
was
suggesting
the
right
thing,
just
on
the
surface
of
it
as
you're
reviewing
a
markdown
file
and
you're
seeing
and
it's
supposed
to
be
in
one
language
and
you're,
seeing
it
some
stuff
still
left
in
english.
It's
like
hey,
you
know
shouldn't
that
be
translated
as
well
like
on
the
surface
of
it.
It's
like
yeah.
It
should.
A
A
You
had
other
feedback
chewbunk
looking
for
the
green
light
from
a
review
from
a
peer
reviewer
like
you
on
alonso's
good
work,
are
we
ready
for
this
or.
F
Yeah,
I
think
one
of
the
things
also
that
I
wanted
to
highlight
is
that
the
issue
that
we
were
facing
on
the
talk
site
is
that
whenever
we
refresh
the
page,
it
goes
back
to
the
english
translated
version
right.
So
one
of
the
things
that
lee
highlighted
was
that
maybe,
instead
of
having
the
root
of
the
page
directly
directed
being
to
the
normal
translated
version,
how
about
we
do
it
the
way
that
for
english
website,
it
should
be
somewhere
like
en
slash
en
for
a
spanish
website.
F
It
should
be
slash
yes,
so
in
that
way,
even
if
we
refresh
the
page
and
if
there
is
any
sort
of
leak
or
anything
like
the
problem
that
we
are
facing
the
404
or
like
anything
like
that,
so
it
will
be
redirected
to
the
same
translated
language.
I
think
if
we
try
to
go
with
this
approach,
I
think
the
404
issue
can
be
solved
in
that
way.
F
A
A
So
I'm
just
like
I'm
going
over
here,
a
b
c
d,
e,
f
and,
and
that
page
doesn't
exist.
Okay.
So
if
I
oh
so.
G
H
F
Sure
sure
and
the
other
thing
I
think,
that's
not
specific
to
alonso's
pr
but
the
the
snap
bar
the
site
nav.
But
what
it
does
is
that
if
I
navigate
down
to
a
page,
let's
say
to
a
link
and
if
I
click
on
it,
the
navigation
bar
resets
back
to
the
topmost
position
so
like.
If
you
go
to
the
most
bottom
link
on
the
navbar
right
right,
left
hand,
side,
navbar
and
you
click
on.
F
F
A
Okay,
yeah,
that's
a
good
item
to
create
a
separate
issue
on.
A
Nice,
I
think
what
I
just
heard
is
a
long
is
that
we
should
merge
the
pr
that
alonso
okay.
A
F
A
Else,
quick,
quick
question:
abhishek
any
any
update
that
you'd
like
to
give
on
it's
great
work.
Alonso
abhishek
update
on
some
of
the
final
tweaks
on
mesh
sync.
I
Yeah
sure
so
there's
no
much
of
an
update,
but
the
work
is
currently
shifted
towards
the
front.
End
part
where
we
are
where
we
are
like
tying
off
the
back
end
queries
to
the
front,
end
elements
and
yeah.
Basically
it
could
be
like
basically
you
can.
It
will
be
available
for
people
to
test
or
in
the
edge
release
by
today
tonight
or
tomorrow,
early.
G
A
There's
just
one,
I
think
I
think
some
of
you
already
know
it's
worth
calling
this
out.
There's
some
of
you
have
worked
on
various
blog
posts
for
this
upcoming
release
announcement
and
there's
a
pretty
good
one
that
asuko
and
michael
have
been
working
toward,
which
is,
I
think,
it's
it's
sort
of
the
main
post.
The
main
release
announcement
that
it
kind
of
it
kind
of
looks
like
this,
where
it's
sort
of
like
it
has
the
highlight
the
top-level
highlights
of
the
you
know:
significant
new
features
and
changes,
and
but
it
also
it
links.
A
It
has
a
lot
of
links
that
point
out
to
either
links
to
docs
or
links
to
other
blog
posts
that
go
in
depth
on
what
is
mesh
sync
or
like
what
is
the
operator?
What
is
the
so?
Not
all
of
these
I
mean
in
an
ideal
world,
we'd
have
a
one
blog
post
per
topic
or,
or
you
know,
does
anyone
know,
because
I
actually
don't
know:
does
anyone
know
where
we're
a
blog
post
behind
or
like
where,
where
we
would
ideally
have
one,
but
we
don't,
but
no
one's
working.
A
A
This
is
great
like
we
should
have
blog
posts
on
there's
a
number
there's
a
couple
of
different
measures.
Ctl
commands
that,
like
first
of
all,
we
should
have
a
message
ctl.
We
have
a
mastery,
ctl
command
reference
great
and
that
should
be
auto
generated
at
some
point,
but
we
should
also
have
a
cheat
sheet
the
cheat
sheet.
It
gives
people
more
of
a
behavioral
like
just
like
common
uses
and
some
behavioral
tips
and
we're
we
don't.
You
know,
have
those.
A
So
that's
actually
what
I
was
going
to
bring
up
with
jagriti
and
aditi
and
they're.
Not
here
we'll
try
I'll
bring
it
up
on
slack,
but
that's
an
open
call
for.
A
A
Okay,
anyway,
thank
you
all
very
nice
to
talk
to
everybody
michael
and
drew
very
encouraged
about
your
conversation,
guys.