►
From YouTube: 2019-06-10 Golang Core Dev Weekly Sync
Description
https://cryptpad.fr/code/#/2/code/edit/sLR7IWXiCbXc-GQOL+DPTDc-/
Updates on gateway performance, Badger, provider records, graphsync, and core API constructor.
B
B
A
B
D
A
A
D
So
they
are
like
one
fix,
I
think
because
apparently
there
is
some
other
way
to
make
things
work,
but
I
think
it's
quite
weird
because
it
involves
closing
and
opening
butter
to
first
compaction
to
run
garbage
collection
and
I
need
to
read
more
about
that.
It's
weird
so
yeah
I
just
need
to
like
give
it
a
day
of
work
to
see
where
it's.
A
So
when
we're
thinking
about
things
like
okay,
ours
or
other
initiatives
that
we'd
like
to
to
land,
alright,
we
have
limited
limited
time
on
our
hands.
Let's
make
sure
we're
prioritizing
it
too
to
get
things
over
the
finish
line.
If,
if
that
is
possible,
and
just
at
the
very
least
what
we're
gonna
need
to
have
in
our
minds
of
like
where
what
is
left
on
this
thing,
is
it
something
we're
continuing
on
into
q3
to
continue
pushing
on,
or
are
we
abandoning
this
or.
B
Great
thanks
for
that
Molly
moving
on
to
bit,
swap
if
everyone's
okay,
good
yeah
I,
know
Hannah's
out
last
weekend
this
week
on
vacation,
she
did
put
up
a
PR
use,
realistic
test
network
with
DHT
and
which
needs
review
so
Imperium
subnet.
Can
you
check
that
check
that
out
this
week?
If
you
have
time,
there's
no
data
session,
moving
on
to
core
API,
Lucas
I
think
you're
working
on
this
as
well?
Yes,.
D
Oh
I,
spam
half
of
at
least
half
of
last
week,
debugging
pops
up
issue
where
if
pops
of
SAP
was
running
under
nyan,
and
you
tried
to
custody
money,
it
would
just
hang
for
whatever
reason
and
yeah.
It
turns
out
the
banking
multiple
context
and
like
5
thirteenth's
at
one
time
is
hard.
So
it
took
some
time,
but
it's
fixed
now
and
the
constructor
for
request
is
ready
for,
if
you
know.
B
B
D
So
that's
on
the
topic
of
who
cares.
That
is
one
of
my
okay
as
for
q2,
but
there
is
very
little
hope
of
finishing
it
with
in
q2.
So
as
we
probably
carry
over
to
Q
free,
but
yeah
I
started
working
on
some
bunch
to
have
something
to
compare
against
I,
probably
know
what
I
will
do
to
improve
things,
so
it
mainly
will
be
just
abandon
any
use
of
amethyst
in
Korean
each
other,
because
it
creates
like
three
times
or
four
times
overhead
for
some
calls.
When
I
didn't
meant
many
files
and
directories
so
yeah.
B
So
provider
record
so
I'm
working
on
that
I'm
working
on
the
provider
itõs
a
little
not
distracted,
but
I
had
I
was
multitasking
last
week.
Basically,
the
JFS
team
has
some
stuff
going
on
that
I
reviewed
looked
at
have
been
having
some
conversations
around
I
went
through
I
introduced
a
provider
tagging
went
through
the
issues
list
and
tried
to
either
review
or
closed
org
and
tag.
Anything
that's
provider
related.
B
A
Yes,
you
know
I'm
people
using
the
provider
records
fix
that
we
you
know
turning
off
providing.
Is
that
something
that's
already
like
accessible
to
people,
or
is
that
gonna
be
it.
B
Is
it
it
is
accessible
I
haven't
I
had
I
was
coordinating
with
Aaron
for
from
infra
for
the
for
rolling
it
out
in
to
get
to
some
of
the
gateways.
I
need
to
pick
up
that
conversation
again.
I
just
yeah
I'll
try
to
get
on
that
I.
A
B
George,
like
carbon
5
George,
oh
great,
ok,
yeah,
well,
I
will
start
that
conversation
with
both
with
both
of
them
thanks
family,
ok,
moving
on
to
the
base
three
to
CID
v1
work
Lydell
can't
join,
but
it
has
a
question
about
next
steps
and
has
a
link
to
an
issue
if
any
security
was
realistic
to
ship
new
Co
ipfs
with
this
before
camp.
Or
are
we
moving
this
to
July?
E
Sure
not
much
to
report
this
week.
I
have
mostly
been
spending
time
on
camp
related,
stuff
kind
of
preparing
in
a
way
to
speak
with
the
team
about
that,
and
then
the
following
week
speak
to
I.
Guess
the
community
about
that
kind
of
stuff.
So
yeah
a
lot
of
a
lot
of
prep
work,
not
a
lot
of
presentation,
work!
Yep!
That's
it
for
me!
Do.
A
You
want
to
just
give
like
a
teeny
I
know,
you're
gonna
probably
be
telling.
Is
this
next
week
as
well,
but
just
like
very,
very
high-level
picture
on
like
where
we're
at?
What's
remaining
and
like
kind
of
like
the
next
steps
on
it,
and
then
we
can
dive
deeper
into
that
next
week
that
we're
going
in
at
least
with
a
little
bit
of
of
context
in
a
big
big
picture.
E
Totally
so
something
I
have
been
doing
recently
has
been
like
looking
at
other
file
systems
and
then,
like
other
things
that
are
like
file
systems,
one
of
those
being
IPL
D,
so
something
that
they
were
talking
about
at
their
last
meeting
was
kind
of
like
the
metadata
around
I.
Guess
like
node,
metadata
and
stuff
like
that,
and
that's
something
that's
kind
of
undefined
in
my
space
at
the
moment.
E
So
it
seems
I'm
kind
of
undefined
or
loosely
defined
they're,
totally
undefined
in
my
domain
and
I
kind
of
have
the
same
problem
that
they're
having
where
we
can
kind
of
do.
Basically,
whatever
we
need
to
do,
but
we
don't
really
know
what
that
is.
So
it's
a
matter
of
like
asking
developers
kind
of
what
do
you
want
to
see
out
of
this
like?
How
do
you?
How
would
you
use
this
API?
E
That's
that's
kind
of
the
thing
I'm
trying
to
focus
on
next
I,
guess
and
then,
like
the
state
of
where
things
is,
is
that
we
have
an
implementation
up.
That
can
do
ace
a
specific
thing,
so
it
mimics
a
UNIX
file
system
and
then
we
can
create
this.
Other
thing
that
doesn't
necessarily
have
to
be
UNIX
like
it
can
be,
go
like
or
if
we
coordinate
with
the
Jas
team
it'll
be
similar
to
a
j
s-
library.
Something
like
that.
I
don't
know
if
that's
paints,
a
clear
picture,
but
that's
where
I'm
at.
E
So
in
the
file
system,
we
have
basically
the
same
kind
of
abstraction
of
that
we
don't
care
what
it
is.
We
just
want
to
know
that
it's
a
node
and
have
a
certain
common
interface
with
that,
so
something
I
saw
in
plan
nine
was
they
have
this
concept
of
like
every
node
has
a
version,
and
it's
not
really
important
for
anything
other
than
caching.
So
it's
it's,
this
abstraction
of,
like
things
like
that,
which
are
not
very
typical
in
UNIX
systems,
where
it's
basically
a
bowl
of
like.
E
B
You
know
we
have
some
discussions
around
metadata
for
like
providing
and
garbage
collection
and
stuff
too
I'd
be
curious.
If
you
could
point
me
to
I'll,
go
looking
myself,
but
if
you,
if
you
know
of
anything,
that's
written
down
that
the
IP
LD
folks
are
talking
about
I'd
love
to
love,
to
look
at
that
generalize,
it
may
not
be
exactly
the
same
metadata.
A
You
definitely
do
I,
hoping
that
I
guess
I
can
give
a
really
high-level
overview
and
then
maybe
folks,
who
are
more
tactically,
so
I
think
as
people
notice.
The
gateways
have
been
struggling
for
the
past
couple
of
weeks,
so
we've
rolled
out
421
to
a
number
of
gateway,
nodes,
I,
don't
quite
know.
If
it's
it's
a
hundred
percent
yet
and
have
been
doing
a
lot
of
investigating.
A
We
spun
back
up
the
the
Gateway
tiger
team
and
renamed
it
the
Gateway
fire
squad,
because
we're
dealing
with
with
fires
and
tigers
at
the
moment
and
have
been
kind
of
jumping
in
to
address
so
roll
has
been
doing
a
lot
of
instrumentation
and
adding
more
metrics
to
the
DHT
I.
Don't
know
whether
those
all
got
pushed
out
over
the
weekend.
But
if
not,
there
are
a
couple
of
more
that
we
can
push
out
to
to
prod
on
the
gateways
and
get
some
feedback
on
weird.
A
The
slowdowns
are
happening
and
there's
already
been
some
some
information
collected
through
that
and
so
I
think
roll
already
has
closer
ideas
of
where
the
problems
are
and
Steven
was
pushing
on,
like
an
actual,
an
actual
fix
that
he
thought
was
a
root
cause
of
some
of
the
issues
that
people
have
been
seeing
and
then
I
think,
kuba
and
or
other
Jeremy
you
might
have
jumped
in
on
helping
instrument
some
of
the
stuff,
the
high
level
goal
being,
if
we
add
better
instrumentation
to
gateways
into
the
the
Lib
p2p
network
and
we're
gonna
have
a
better
sense
of
where
the
problems
are
happening
and
it's
gonna
be
faster
for
us
to
debug
issues
and
prevent
issues
in
the
future.
A
And
so
that's.
We
have
kind
of
a
tactical
track
on
this,
like,
let's
just
figure
it
out
right
now,
and
then
we
have
a
more
forward-looking
track
of
what
are
the
things
that
we
want
to
be
alerting
on
over
time
and
the
Gateway
team
can
is
helping
set
up.
Some
of
that,
we
have
a
meeting
I.
Think,
right
after
this
one
too,
to
talk
about
our
status
on
that
and
how
how
those
changes
that
were
ended
does
anyone
else
have
a
more
tactical
update
to
me.
C
So
the
the
other
thing
that
will
help
the
network
with
404
21
is
more
people
upgrade.
That
will
only
help,
because
that
includes
a
few
fixes
for
some
people
address
lotion.
We
were
basically
some
of
the
keynotes
will
remember
all
the
dresses
and
won't
forget
them,
as
you
give
them
new
addresses,
so
that
when
I
wish,
I
was
going
to
try
to
help
20
addresses
or
4,800
instead
of
just
like
one
or
two
good
addresses
which
make
battling
take
longer.
C
C
So
that's
thing:
we
can
do
an
IPS
s,
the
questions
like
do
we
want
to
change
this?
You
see
white
people
there's
a
config.
Currently,
unfortunately,
we
build
all
of
these
people
into
the
config
so
like
we
have.
No
there
told
me
if
the
user
explicitly
said
I
want
to
be
the
HT
node
or,
if
that's
just
the
default
to
the
Gudrun.
Recreational
interest,
kicked
em
off
with
the
HD
and
we
say
hey
when
you
upgrade.
We
run
this
migration.
C
If
you
want
to
rejoin
the
DRC's,
please
do
this
I
really
don't
want
to
do
that,
but
I
think
that
may
be
the
best
way
of
doing
that
at
this
point,
yeah,
because,
like
otherwise
just
like
upgrading,
maybe
some
issues-
you
don't
select
the
partners.
So
if
you
have
a
lot
of
partners
that
are
controlling
one
of
these
nodes
or
they
may
be
able
to
just
switch
on
some
things,
I
want
to
make
sure
that
Excel
and
opals
are
not
putting
all
the
notes,
like
all
the
mobile
nodes,
the
ICC
service,
but
yeah.
A
Sounds
worth
starting
to
a
conversation
about
that
I.
Don't
know
whether
that's
this
week,
but
that's
like
we
could
have
a
conference
and
then
we
feel,
like
you
know,
we're
gonna
have
them
all
they're
doing
like
the
best
hand,
that
might
be
a
good
opportunity
to
like
get
a
lot
of
Acts
and
something
out
quickly.
Cuz
I'm,
just
worried
this
thing
is
gonna
end
up
in,
like
it'd,
be
a
really
good
idea.
A
B
A
Yeah,
nothing
like
me,
I
should
have
updated
the
note.
So
we
had
a
meeting
on
Monday
with
Alex
north
on
the
Falcon
teams.
You
just
get
a
sense
of
kind
of
what
our
timeline
as
Aslan
grassing
Tiger
team
is
looking
at
like
and
what
our
first
first
date
user's
timeline
is
like
and
whether
that's
gonna
fit
well.
A
So
that's
kind
of
where
everything
is
where
we
want
to
have
a
conversation
about
this
Wow
lots
of
people
are
sync
with
each
other
in
Barcelona,
and
so
we
should
keep
that
on
our
agenda
and
make
sure
you
find
time
during
an
uncommon
to
you
know,
be
looking
through
Eric's
selector
stuff
and
making
sure
that
we're
super
good
to
go
on
it.
So
right
when
July
starts,
we
can
hit
the
ground
running.
A
B
A
B
Appreciate
that,
though,
so
garbage
collection,
I
added
this
on
here
until
the
provider
stuff
gets
closer
to
wrapping
up.
This,
probably
not
gonna,
be
huge
updates
on
this,
but
I'm
gonna
try
to
push
the
low-hanging
fruit
forward
steadily
as
much
as
I
can
turns
out
that
the
JSI
PFS
teen
is
also
doing
stuff
around
garbage
collection,
I
kind
of
sort
of
randomly
and
talking
about
the
provider
stuff.
B
With
someone
learned
that
and
so
I've
been
trying
to
read
up
on
on
some
stuff
that
they're
doing
and
also
having
some
discussions
there
I
have
started
putting
together
the
bigger
garbage
collection
issue
and
I'm.
Also
gonna
do
the
same
thing
where
I
go
through
and
try
to
find
everything,
that's
related
and
try
to
bring
it
all
together
into
one
issue.
I
started
that
process
I
just
have
them,
I
just
haven't,
got
it
all
together,
yet
so
so
I'm
still
working
like
I'm
still
working
on
and
that's
it
for
the
initiatives.
B
C
F
Know
a
little
bit
so
who
is
owning
the
I
penis
to
DNS
for
like
within
the
web,
browsers
working
group
and
so
is
goal-
is
to
have
that
completed
still
this
week
or
next
week,
because
workshop
on
identity
will
rely
on
it.
I
don't
believe
there
is
a
spec
I
think
these
like
work
is
very
develop
on
on
the
description
that
pilars
wrote
a
while
back
and
so
I
think
right
now,
it's
very
G
eccentric,
like
I.
F
C
Okay,
just
having
this
some
document,
I
can
look
at
it
because,
like
if
there's
some
know
this
Kalia
v1,
but
it
doesn't
sound
like
it-
should
be
that
hard
to
go
because
we
have
a
bunch
of
good
DNS
libraries.
You
can
just
throw
and
pull
the
data.
The
server-side
it's
like
more
complicated,
apply
inside
so
I
think
it'd
just
have
a
hacky,
well
yeah.
F
C
A
A
C
F
Think,
like
that,
it's
also
like
I
think
I
professed
desktops
like
they
do.
Just
the
IPTS
movie
team
I
spend
a
lot
of
working
out,
updating
or
enabling
users
just
click
on
one
box
and
I
get
to
know
that
updated
for
a
professor
stuff,
which
which
apparently
helps
a
lot
and
they're
like
right
now,
like
50k
downloads
or
something
like
of
course
like
not
all
of
them
are
always
on
all
time.
Bam.
It's
a
lot
of
notes.
Coming
on
there.