►
From YouTube: gRPC Community Meeting 3.15.18
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
Okay
yeah,
so,
as
we
were
saying,
you
know
it's
a
little
bit
of
a
smaller
career
today,
things
like
the
first
day,
I
think
I've,
actually
adjusted
to
daylight.
Savings
I
actually
got
a
decent
night's
sleep
last
night,
whereas
that
has
not
been
happening
and
it's
silly
for
us,
because
it's
brain
Dolby,
so
we
haven't
even
seen
the
extra
fun
completely
unfair.
A
A
Trying
to
find
the
tab
that
has
been
working
notes
in
it.
There
we
go
so
I
did
post
a
link
again
in
the
chat
and
I
see
that
them
both
have
added
some
questions
in,
which
is
great.
If
you
have
anything
that
you
want
to
share
with
the
community,
please
you
know
either
add
the
Lincoln
or
as
a
Lincoln
or
you're,
welcome
to
just
kind
of
leave
on
mute
and
say:
hey
would
like
to
talk
about
XYZ.
A
One
thing
is
that
we
typically
do
for
the
sort
of
new
meetings
as
we
encourage
everybody,
who's
new
to
the
hall
or
new
to
working
with
your
FPC.
If
you
want
to
say
hi
and
tell
us
a
little
bit
about
you
and
what
you're
working
on
and
if
we
have
anybody
this
new
this
week,
but
if
there
was
one
thing.
B
A
D
C
Honor
ball,
oh
hi,
so
there's
this
Naresh
and
one
of
the
students
interested
in
applying
for
Chuseok
through
G
RPC
this
year
and
so
I've
been
interested
in
this
project
where
we
had
to
build
CRP,
C
Python
using
basil
and
that's
a
project
I've
been
working
and
I
mean
I've
been
attempting
and
Nathaniel
has
been
very
helpful
with
reviewing
well
my
we
have
so
far,
which
really
helped
me
with
you
know
getting
familiar
with
the
foodways.
So
thank
you
so
much
so
I
do
have
a
couple
of
questions
today.
C
A
couple
of
questions
today,
so
I
hope
I,
don't
take
too
much
time,
so
so
I
guess
I'll
discharge.
So
one
of
my
I
mean
literally
the
biggest
question
that
I
have
is
from
my
understanding
of
basil.
So
far,
is
that
so
basil?
So
currently
it
has
a
lot
of
build
tools
for
five
and
something
like
PI
underscore
binary,
which
will
so
since
we
are
dealing
with
Python
here,
a
binary
does
not
make
sense
as
such,
so
it
is
going
to.
C
You
know,
rearrange
our
code
in
two
basil
bin
and
you
know
just
keep
it
there
and
it.
So
if
we
were
to
just
use
that
directly,
then
that
would
mean
that
we
have
to
change
a
lot
of
our
code.
So,
for
example,
things
like
tests
wouldn't
work
because
we
are
using
import
G
RPC
there
and
like
you,
would
need
absolute
directory.
So
that
is
a
problem.
C
So
what
I
instead
did
was
that
I
had
a
look
at
tensorflow,
how
tense
of
law
is
doing
things
and
what
tensorflow
is
doing
David
actually
running
so
right
now,
what
we
do
is
set
up,
thought
piping.
They
were
doing
that
using
basil,
but
I
am
not
fully
convinced
that
it
delivers
the
advantages
that
basil
would
typically
give.
So
because
their
answer
to
this
is
there
something
that
you
have
distress.
You
have
in
mind
for
what
can
be
the
ideal
solution
to
this
problem.
D
D
The
extent
to
which
we
should
change
and
the
extent
to
which
battle
is
going
to
change
and
where
we're
going
to
meet
in
the
middle
and
how
we're
going
to
accomplish
goals
like
running
our
tests
in
a
sensible
way
that
that
will
understand
and
possibly
building
our
distribution
artifact
in
a
sensible
way
that
bad
law
understand.
So
so
we're
not
making
a
lot
of
statements
saying
like
well,
since
the
current
state
of
basil
is
the
way
it
is.
D
D
The
message
that
I
linked
here
says
the
project
is
open-ended.
We
we
would
not
put
specific
goals
on
it.
Well,
we
would
put
specific
goals
on
it.
The
specific
goal
that
we
would
put
on
it
would
be
things
like
progress
toward
running
some
of
our
continuous
integration
in
the
Bassel
integrated
way,
building
at
least
some
of
our
distribution
artifacts
with
babel
sort
of
sort
of
moving
us
toward
our
eventual
goal,
and
maybe
getting
that
is
a
very
open-ended
project.
Is
a
project
idea.
D
C
Yes,
sort
of
that
does
so
I
have
been
following
so
I
found
this
one
special
I
had
looked
at
the
response
that
it
is
linked
and
also
I
found
the
special
interest
to
for
basil
to
spy
on
that
was
so
they
have
been
discussing
a
lot
of
things
related
to
this,
so
I
probably
so.
This
is
probably
a
follow-up
question,
so
how
would
it
work
I
mean?
C
D
Out
there
and
say
yeah:
okay,
frost
repository
contributions
will
be
great.
This,
as
I
sort
of
said
I.
Think
part
of
this
story
is,
can
be
we're
bending
toward
basil
and
bath
ovens
a
bit
toward
us.
That
sort
of
seems
to
think
there's
going
to
be
cross
project
in
a
ration
to
at
least
some
moderate
degree
is
not
a
strong
degree.
D
Okay,
so
the
picture
in
the
current
state
of
affairs
is
saitama's,
an
open
source
project,
bezels
an
open
source
project.
Erp
C
is
an
open
source
project,
we're
all
aware
of
each
other.
Some
of
the
same
people
are
involved.
I
can
tell
you
that
inside
Google
there
rules
for
building
scythe
on
with
the
basal-like
system
that
we
used
internally
and
everybody
agrees
that
there
should
be
such
rules
outside
I.
Don't
remember
the
current
state
of
affairs
of
them.
D
C
Okay,
so
I
saw,
but
so
I
mean
I
just
want
to
make
sure
that
there's
no
redundancy,
because
if
this
was
to
be
I
mean
if
this
is
in
the
roadmap
some
place
to
basil,
then
that's
probably
something
that
I
mean
so
I
think.
Maybe
that's
not
something
I
should
be
covering
in
my
purposes
but
yeah
I
guess
you
have
answered
the
question
so
yeah.
Thank
you.
So
much.
B
Because,
frankly,
the
attitude
right
now
is
a
little
like
unfortunate,
and
you
know
I
just
like
to
say
you
know,
like
we're
working
a
lot
of
open
source
projects
as
well
and
I
understand
like
it.
There
can
be
a
lot
of
volume
and
it
can
be
frustrating
to
deal
with
with
people
filing
bugs
and
stuff
like
that.
B
But
at
the
same
time,
like
one
thing,
we
tried
not
to
do
especially
a
doctor
and
especially
on
container
D
projects
and
other
CNCs
projects
is
not
to
like
not
to
like
in
stuff
close
issues
or
just
kind
of
closed
issues
summarily
after
you
know.
No
resolution
is
sound,
so
I
hope
that
I
know
at
this
point.
You
guys
are
a
CNCs
project
and.
E
F
B
G
B
F
So
it's
impossible
to
avoid
breaking
backward
compatibility
like
across
the
board.
Right
like
it's
literally
impossible.
We
had.
You
know
we
have
existing
api's
that
have
problems
with
them
in
implementing
new
features
which
are
required
by
G
RPC
because
of
some
unfortunate
design
decisions
that
were
made
in
the
past,
and
so
essentially,
there
are
occasionally
places
where
some
subtle
behavior
has
to
change
and
it
will
unfortunately
break
some
users
backward
compatibility
guarantee.
We
we
published
these
things
as
broadly
as
we
can.
We
published
a
G
RFC
for
this
particular
feature
in
July
of
last
year.
F
This
was
posted
to
the
mailing
list.
It
was
published
in
the
repo.
All
of
the
work
you
know
was
undergoing,
you
know
publicly
and
you
know
took
place
and
reviews
were
available.
There
were
plenty
of
opportunities
for
people
to
point
these
problems
out
and
for
most
users.
You
know
this.
This
was
not
a
major
impact,
so
I
think
you're.
A
particular
use
case
was
unfortunate
because
it
was
you
know,
like
the
syntax
of
your
target,
happen
to
exactly
match
the
URI
parsing
scheme
that
you
know
the
G
RPC
spec
initially
said
it
supported.
B
B
But
actually,
when
you
look
at
the
implementation
like,
you
can
see
the
implement,
like
one
implementation
uses,
four
slashes
I
think
to
go,
and
then
the
C
implementation
uses
three
slashes
and
then
I
was
looking
at
the
PR
for
against
the
G
RPC
go
library
to
add,
actual
UNIX
socket
support
and
it's
using
two
slashes
in
the
example
so
like
it's,
it's
it
doesn't
seem
like.
Even
if
you
look
at
those
specifications
and
proposals,
you
wouldn't
see
this
breaking
change
coming
down
the
pike.
B
The
other
issue
is
like
the
standard
way
and
go
of
adding
like
new
functional
you're.
Changing
functionality
is,
is
to
add
a
new
method
or
like
a
new
function
to
the
package,
and
you
know
it
seems
like
the
nature
of
these
functions
has
completely
changed
and
it's
not
even
it
uses
the
term
dial,
but
it
doesn't
really
dial
an
actual
connection
or
anything.
It's
really
like
a
connection.
F
So
you
know
I
mean
so:
I
I've
been
on
the
project
for
a
year
now,
personally,
I
don't
have
before
I
did
have
a
blog
post
about
a
month
or
two
ago
about
some
of
the
changes
that
we
have
made.
If
you
haven't
read
that,
please
do,
and
you
know,
I
mean
I
hope
that
maybe
changes
your
opinion
of
how
we're
running
the
project,
because
I
think
we're
doing
a
lot
better
job
now
than
we
were
back.
B
A
It
sounds
like
maybe-
and
this
is
April
I
know
I'm
just
directly
off
the
camera.
It
sounds
like
it
probably
would
be
a
good
idea.
If
maybe
we
had
a
you
know-
I
say
the
term
offline,
but
another
conversation
where
we
can
kind
of
begin
with
straight
offline
online,
where
we
can
like
begin
to
use
them
of
the
specific
issues
that
yellow
thing,
because
it
sounds
like
there's.
F
B
B
F
It's
not
that
I
didn't
take.
You
seriously
is
that
there
was
like
an
unfortunate
breakage
that
was
caused
by
you
know,
implementing
the
new
resolver
imbalance
or
API
that
you
know
affected
a
small
percentage
of
our
users
and
I
did
you
know,
provided
that
that
was
expected
to
be.
You
know
forward
compatible
for
you,
but
just
not.
F
Like
nothing
else,
I
can
really
do
there.
Besides,
like
okay,
we
have
to
rollback
that
entire
design
and
re-implement,
like
so
I,
don't
I,
don't
know
exactly
what
you
were
hoping
for
there,
but
I
mean
I,
provided
a
workaround
and
I
quote
the
issue
and
that
workaround
should
work
for
you.
If
that
workaround
does
not
work
for
you
please,
you
know.
F
Let
me
know
we
can
work
on
that,
and
really
that
is
a
way
that
should
work
going
into
the
future
is
not
just
like
a
work
around
the
patches,
a
problem
that
we'll
be
fixing
right
now.
You
did
point
out
a
difference
in
the
way
that
we
are
handling
the
URI
scheme,
and
you
know
the
way,
the
way
that
we
had
implemented.
F
It
was
actually
according
to
documentation
that
is
in
our
gr,
PC
specs
doctor
yet,
but
when
I
went
in
talked
to
some
of
the
team
members
I
just
let
me
finish
real
quick
when
I
went
and
talked
to
some
of
the
other
team
members
more
senior
like
they
indicated.
Oh,
you
know
that
specs
really
just
there
to
highlight
example,
use
cases
and
the
actual
spec.
That
should
be
implemented.
Vri
scheme,
as
you
know,
in
the
RFC,
forget
the
number
396
at
which
point
okay.
This
never
was
clear
to
us
before
and
so
I've
held.
E
B
So,
like
I
mean
first
of
all
like
this,
isn't
you
know,
while
it
looks
like
it's
only
like
a
small
percentage
of
users,
a
lot
of
people
use
container
D
through
doctor,
so
like
the
impact
is
much
wider
than
then
it
looks
like
the
other
thing
is.
We've
also
had
other
breakages
that
that
impact
docker
users
as
well.
So,
like
you
know,
while
it
you
know
well,
it
looks
like
there's
it's
just
a
small
number
of
users
complaining
about
it
on
your
your
issue,
tracker.
B
F
F
B
F
B
A
Another
look
at
this
yeah
just
so
we'll
take
another
look
at
it
and
see.
You
know
if
there's
a
communication
breakdown
or
something
kind
of
doing
analysis
there
and
see
what
we
can
do
to
fix
this
going
forward
and
then
it
sounds
like
we
should
probably
have
a
you
know,
a
small
huddle
with
your
team
and
some
of
the
folks
on
this
team
here
to
kind
of
talk
through
some
of
these
open
issues
and
see
if
we
can
come
to
a
resolution
that
works
for
everybody
and
how
we
can
make
things
better,
feel
yeah.
B
B
If
you
know
we
could
have,
we
could
revisit
the
past
through
design
and
make
sure
that
that's
going
to
be
working
for
people
who
are
using
UNIX
sockets,
but
I'm
also
afraid
that,
if
we're
all,
if
we
need
to
do
any
kind
of
work
on
the
UNIX
connection
for
like
if
we're
working
with
like
socket
credentials
or
any
kind
of
unique
specific
features,
we're
going
to
always
be
using
pass-through
and
always
kind
of
bypasses
bypassing
the
dialer
stack.
So.
F
F
B
Not
all
they're,
not
all
sorry
I
mean
basically
like
the
dialer
model,
is
really
good
and
go
and
I
wish
that
the
the
G
RPC
go
package
would
kind
of
subscribe
to
the
model
of
a
net
package
a
little
bit
more
closely,
so
that
you
know
we
could
leverage
that
common
infrastructure
rather
than
having
to
like.
We
almost
have
to
like
fight
against
it
in
places
so
yeah.
F
I
mean
I
agree
with
some
of
the
sentiments
there.
Unfortunately,
you
know
I
mean
we
are
trying
to
maintain,
even
though
you
know
it
does,
it
does
break.
In
some
cases.
We
are
trying
to
maintain
backward
compatibility
as
much
as
possible,
which
means
that
we're
sort
of
stuck
with
some
of
the
existing
API
decisions
that
were
made
in
the
past,
and
you
know
we're
doing
the
best
that
we
can
to
work
around
those
to
add
features
to
them.
Instead
of
you
know,
changing
things
and.
F
Probably
there's
a
there's
a
PR
about
this,
but
probably
what
we're
going
to
do
is
Institute
or
implement
a
new,
a
new
way
of
creating
a
client
that
more
mimics
the
way
that
the
other
languages
do
G
RPC
clients,
which
is
you
know,
like
a
new
client
method
that
doesn't
actually
you
know,
start
even
dialing
at
that
time
right
and
that
that
makes
using
G
or
Christine
go
a
lot
more
like
using
jira,
see
in
other
languages,
and
it
does
make
it
a
little
bit
less
like
using
the
net
package.
But
really
a
Geo.
B
F
B
F
E
B
Than
because
like
for
I
mean
like
one
big
example
is
like,
we
can't
use
the
balancer
infrastructure
at
all
in
India,
our
PC
go
just
because
the
way
it's
implemented
I
mean
the
original
one,
and
so
I
haven't
looked
at
the
new
one.
Even
so
we
would
just
our
current
balance
or
infrastructure
and
so
on
fit,
for
example,
is
just
completely
custom
because
we
can't
use
it
so
yeah
I
mean
and
I
don't
know.
B
If
we're
gonna
be
able
to
do
the
work
to
get
that
done
and
I
know,
other
projects
are
having
trouble
integrating
it
I'll
send
you
the
I'll,
send
you
the
239
issue
and
that
that
covers
that
that
stuff
and
I'll
just
try
the
any
more
work,
more
involvement.
But
you
know
it's
at
the
same
time:
it'd
be
good
to
kind
of
open
communication
channels
between
us
a
little
bit
better
yeah.
A
A
B
Talks
about
the
source
of
these
problems
and
it's
down
I,
don't
it's
down
somewhere
in
it
in
it,
talks
about
that.
If
I,
you
know,
if
I
rewrote
package
like
what
would
you
do
you'd,
you
know,
you'd
make
new
client
instead
of
dial
and
then
most
confusions
and
any
and
I
understand
backwards
backwards.
Compatibility
as
king,
but
I
would
have
rather
had
a
compiler
break
than
have
the
behavioral
break.
It's
just
in
turn
like
weighing
those
dude.
B
A
A
A
E
E
Just
maybe
a
general
comment,
and/or
question
for
best
practices
kind
of
day,
so
the
problem
there
essentially
is
when
they're
not
actually
restarting
a
server
right,
then
all
points
are
connecting
back.
You
know
or
there's
a
spike
in
connection
area
and
that
choose
a
lot
of
CPU
cycles.
Where
I
was
wondering
if
this
best
practices
situation.
A
E
A
So
I
actually
am,
and
this
is
for
everybody
I,
just
added
it
to
the
doc
I
also
planned
on
sending
out.
You
know
some
sort
of
like
survey
thing
or
whatever,
but
I
am
trying
to
compile
like
our
hit
list
of
the
main
documentation
we
want
so
like.
What's
the
wish
list
for,
if
I
suddenly
had
you
know
the
ability
to
have
someone
write
anything
I
want
to
you
know
what
would
y'all
want
to
be?
A
F
E
The
problems
with
the
forks
with
kindness,
you
came
up
whoa.
Would
you
help
obviously
suggest
where
it's
fine
I
should
right
now
we
just
want
to
see
what's
the
long-term
vision
for
supported
properly,
maybe
one
day,
I
think
just
an
officiai
story.
You
know
if
you
could
document
at
a
certain
place
that
should
be.
F
B
H
G
A
we
have
a
we
have.
No
interpretation
is
a
very
comprehensive
implementation
tool
for
Zico
and
pure
JavaScript.
So
you
may
want
to
look
at
that.
If
you're
talking
about
browser,
we
have
a,
we
actually
have
been
working
on
it
for
a
while
DP
repo
Cote
RPC
web
is
just
private.
For
the
moment
we
have
been
going
through
internal
processes
to
get
approvals
to
make
it
public.
So
in
there
we
are
in
do
this.
G
You
have
question
about
0s
in
distress
and
we
are
using
law
for
that
so
consistent
with
other,
clear
PC
languages
so
on.
If
you
want
access
to
the
privately
github
repo
for
now
our
please
email
me
and
Stanley
term
at
a
standard
can't
get
up
or
Spanish
on
at
Google,
so
we
but
put
in
there.
We
are
very
close
well.
It
should
be
a
couple
of
more
weeks
to
get
all
the
approvals
for
making
the
github
repo,
okay.
H
G
So
in
there,
basically
we
or
browsers
and
JavaScript-
we
provide
a
very
GRCC
like
API
on
the
client-side
ahead.
I
go
through
a
different
protocol.
Is
that
your
PC
web
protocol?
It
doesn't
directly,
isn't
a
directory
RPC
protocol
and
it's
a
proxy
and
that
proxy
were
translated.
It
back
to
a
GP
s--
back
end.
So
the
trustee
story
there
we
are
using
nginx,
but
in
the
future,
or
there
might
be
more
in
competition
on
the
side.