►
From YouTube: ⚡️ⒿⓈ Core Dev Team Weekly Sync 🙌🏽 February 18, 2019
Description
A
For
that.
We
also
added
that
to
the
ipfs
NPM
registry
mirror
on
that
alex
is
doing.
Some
work
on
also
got
the
initial
release
of
pull
in
Plex
done
currently
working
on
running
that
up
the
chain.
I've
got
a
work-in-progress
pull
requests
for
JSI
PFS
tests
are
passing
there,
so
I'm
just
working
on
Interop
test
suite
right
now
to
make
sure
everything
is
good
with
go
so
hopefully
we'll
have
that
ready
to
go
this
week.
I
have
to
do
migrating
some
stuff
over
from
Travis.
A
If
you
are
working
on
the
Jas
ID
of
things,
and
you
have
projects
and
you're
working
on
them,
make
sure
to
get
those
migrated
over
to
Travis
at
your
earliest
convenience,
you
can
follow,
along
with
the
ipfs
asier
ticket
number,
to
see.
What's
going
on
there
and
see
what
stuff's
already
been
worked
on
today,
also
released
a
fix
for
the
PDP
circuit,
where
the
source
and
the
destination,
pure
validation,
wasn't
really
happening,
so
that
should
be
improved
now.
A
A
Has
the
ability
already
for
I
if
s
to
be
able
to
specify
your
actual,
listen
addresses,
which
is
super
convenient
if
you're
running
inside,
like
a
docker
container
or
on
a
z2
instance,
and
you
want
to
run
behind
the
ELB,
you
can
actually
specify
that
you're
listening
or
you
can
announce
on
the
ELB
and
though
you're
not
technically
listening
on
it,
which
is
super
helpful.
So
we've
been
running
into
some
issues
with
that
and
we'll
need
it
for
the
Lippe
to
be
testing
environments.
A
B
But
I
found
it:
okay,
hi
everyone
nice
to
see
you
I'm
here,
I
didn't
mean
to
be
here,
but
I've
got
my
dates
wrong
anyway,
I
what
we're
done:
I,
okay,
first
of
all,
Travis
sorry
Jenkins
went
away
rather
abruptly
and
we
switched
to
Travis
and
so
I
did
quest
to
do
that.
For
the
HTTP
client,
then
I
updated
a
couple
of
examples
for
the
HTTP
client,
because
they're
really
out
of
date
and
I
didn't
realize
and
I
think
that
was
part
of
something
else.
I
was
doing,
but
I
just
noticed
it.
B
B
Okay,
so
we
renamed
the
HTTP
client
module
a
while
ago
now
and
we
took
a
snapshot
of
the
people
who
who
were
currently
depending
on
it
and
at
the
moment
there's
a
whole
bunch
of
people.
We
were
depending
on
the
old
version,
so
I
made
a
little
script
to
grab
to
take
that
data
or
that
snapshot
data
and
turn
it
into
a
bunch
of
like
maintain
as
github
user
names
who
we
consider
to
be
the
maintainer
because
it
so
we
get
reposed.
B
But
they
are
not.
You
can't
just
grab
the
oh,
the
great
I've
using
that
amount
of
them
always
because
sometimes
they're
an
org,
and
so
what
the
script
does.
Is
it
if
it's
a
github
user,
then
that's
cool
they're,
probably
the
maintainer
for
that
particular
package.
But
if
they
are,
if
it's
an
org,
it
looks
at
the
contribution
stats
for
that
particular
repo
and
sees
who's
done.
The
most
commits
and
they're,
probably
the
maintainer,
as
per
that's
good.
B
But
if
anyone
has
any
other
suggestions
on
that,
then
I
am
open
to
open
to
hearing
them,
because
that
was
just
a
metric
I
made
up
and
so
yeah.
We
can.
We've
now
got
a
list
of
people
who
wish
you're,
probably
paying
more,
send
a
pull
request
to.
But
the
important
thing
is
we
know
who
they
are
and
we
know
what
repo
is.
That
currently
depends
on
it
and
it
takes
out
all
of
the
ones
that
have
updated
since
the
snapshot
were
taken.
So
that's
all
good
cool
debugging,
an
issue
with
Jess.
B
B
So
that's
a
quarter,
debugging
well,
I'm,
really,
but
I've
just
found
out
what
the
problem
seems
to
be
so
anyway.
I
think
that
needs
some
go.
People
there's
some
help
from
the
go
people
and
that
side
I
started.
I
started
work
on
a
on
implicit,
which
is
a
kind
of
async
iterators
implementation
of
M
Plex
because
call
em
Plex
exists
now.
This
is
way
easier
because
it's
way
easier
to
look
at
the
pool,
koan
and
figure
out.
What's
going
on,
but
yeah
I,
so
I
just
made
a
quick
started
on
it
over
the
weekend.
B
Actually-
and
it's
currently
so
currently
uses
vo
to
avoid
unnecessary
buffer
can
cats,
which
is
good
because
I'm
getting
buffers
together.
It
takes
a
long
time
and
I
did
an
experiment
using
so
variance.
We
use
that
quite
a
lot
too
so
to
encode
and
decode
things,
but
it
needs
a
it's
meant
to
be
to
be
given
a
buffer
but
I'm,
giving
it
a
BL
instead
and
I'm
using
a
proxy
to
to
get
the.
So
it
like
variant
only
asks
like
for
the
length
property
and
like
the
square
bracket
kind
of
zero.
B
Okay,
that
was
really
long.
Sorry,
I'm
blocked
on
that
bits.
What
thing
I
was
talking
about
and
I
am
not
in
on
myself
first
day
this
week,
so
I
feel
like
she
got
tomorrow
to
do
more
work,
so
I
would
be
doing
more
work
on
em
exit
or
the
CEO
TV
one
base
ready
to
work.
That's
me.
Thank
you
very
much
for
listening
any
questions.
C
So
I
mostly
spend
my
time
in
the
IPL
de
world,
but
the
interesting
bits
for
people
also
in
this
meeting
is
I've
created
a
tool
called
tech
builder,
which
is
you
describe
your
deck
in
a
text
file
and
in
crates,
then
the
deck
for
you
in
does
the
linking
and
hashing
and
the
links
match
with
the
hashes
and
so
on
so
yeah.
And
so,
if
we
ever
need
to
experiment
with
some
deck
stuff,
it
might
be
useful.
C
C
So
that's
really
work
in
progress,
but
if
anyone
wants
to
check
out
what's
going
on
there
and
just
to
get
out
and
see
but
crazy
things
we're
doing
and
I'm
not
blocked,
and
hopefully
next
week,
I
am
not
fully
into
the
selector
stuff,
so
I
can
finally
get
this.
I
is
I,
feel
the
API
change
merged,
hopefully
yeah.
D
C
C
So
basically,
yes,
so
the
idea
is
that
M
L,
unless
you
were
building
the
Explorer,
for
example
the
ipod
explorer.
We
probably
had
the
problem
of
you-
don't
have
a
deck
to
test
it
on.
So
the
idea
is
not
really
big
so
busy.
It's
truly
still
one
way
thing.
So
it's
not
that
you
create
text
file
out
of
your
deck.
C
But
it's
really
you
do
some
testing
and
you
need
some
weird
shaped
deck
somehow
and
before
basically
pit
in
those
see
things
manually
then
copy
and
paste
the
CDs,
so
you
link
them
properly
and
so
on,
and
you
can
just
put
them
in
a
in
a
file
the
structure
and
then
it
will
figure
out
the
CDs
automatically
until
thinking
and
everything
and
you
don't
have
to
care
upon
it
basic.
That's
the
idea.
I
see.
Oh.
B
B
C
Works,
no,
no
there's
the
problem
with
me
like
in
the
FBI
stuff.
It
somehow
didn't
always
work
for
me
that
well
and
also,
if
you
really
want
to
see
the
full
graph
so
the
big
score,
you
can
click
through
things,
but
if
you
really
want
to
have
a
bigger
one,
you
want
to
see
the
full
shape.
For
example,
then
em
you
can
use
at
the
graph
with
thing.
But
three
is
super
simple
scriptures.
You
get
an
idea
of
the
data
and
also
what
it
does
is.
C
So
you
write
a
function
which
then
tells
you
which
properties
to
use
as
a
name
for
the
node.
So
you
get
a
structure
of
a
thing
without
all
the
properties
and
everything.
But
again
it
was
more
just
like
a
quick,
heck
and
I
think
that
might
be
used
for
someone
else.
So
I
just
published
it.
But
it's
not
really
like
that.
Properly
supported.
C
E
Everyone
and
so
I
got
some
progress
on
interrupt
during
the
last
days.
Most
importantly,
I
shipped
today
the
HP
go
dependency,
which
is
in
the
NPM
Dola
to
pay
the
repo
IRA's
support
for
Linux,
Mac,
OS
and
Windows
binaries.
Then
I
also
got
ready
today
to
interrupt
test
setup
and
the
connection
for
request,
then
to
Jacob
will
soon
review
and
then
I
got
some
fixes
or
improvements
in
both
Liberty
J's
lab
experiment
and
just
like
a
speed
demon
client
for
making
everything
work
in
danger,
oxides
then
I
also
got
Liberty
Vegas,
observe
initial
implementation.
E
Reviews
I
provided
a
lot
of
feedback
to
my
dear
she
so
that
she
had
some
movements
in
the
PR.
I
have
not
checked,
but
maybe
she
already
fixed
everything
there.
It's
a
report,
but
I
will
try
to
have
another
look
this
way.
Then
there
was
a
problem
with
the
defaults
for
the
lipids
become
fake
and
the
DHT
was
not
being
disabled
in
browser
now.
E
E
F
Last
week,
spent
all
of
the
week
in
London
with
the
London
crew
was
great,
really
fun
time,
so
our
work
cut
and
most
of
it
was
rather
it
to
the
mouth
to
Travis
fixing
problems
with
some
ripples
and
more
problematic
like
the
lead,
p2p
and
JSA
PFS
repo,
almost
all
of
its
God
Souls.
We
are
only
missing.
The
windows
brought
one
in
linkage,
P.
Hopefully
that
will
not
be
on
our
side
and
it's
something
that
the
widow's
machine
from
travels
is
missing.
We
will
get
that
sorted
out
as
soon
as
you
can.
F
F
F
Also,
we
did
a
lot
of
conversations
with
a
few
team
about
the
benchmark
set
up
machines.
Are
we
deploy
stuff,
the
pillow
lers
bootstrappers,
making
Jesse
BFS
wineries,
so
it's
easier
for
the
infantry
to
deploy
our
staff
to
the
to
the
there's
infrastructure
and
some
actions
got
out
of
it
and
we
will
follow
up
in
the
next
couple
of
weeks
on
diets
and
this
week,
I'm
hoping
to
finish
the
bonuses
PRS
during
the
bonuses
decoration
to
basically
add
an
extra
check
to
at
least
some
of
the
repos.
F
F
A
F
It
depends
on
your
needs
if,
by
default
we
shall
hope
that
the
windows,
the
we
know
the
tests
running
in
Windows-
will
pass
for
now.
We
are
only
getting
problems
on
lipika
P,
but
more
stuff
can
come
up
because
the
windows
VMs
on
Travis
are
still
in
beta,
so
yeah.
If
you
run
into
problems
either
ping
me
first
so
to
see
if
we
can
fix
them,
because
some
of
some
stuff
is
already
in
the
link.
I
I
put
there
some
stuff.
We
already
know
that
Windows
like
breaks
and
we
need.
F
F
I'll
have
the
config
or
the
Yamal
block
to
the
to
this
issue
on
Asia,
so
people
that
people
that
find
problems
we
do
Windows
know
how
to
add
the
allow
failure
option
to
their
Travis
country,
or
probably
just
to
the
readme,
where
the
the
instructions
to
I
try.
These
are
one
of
those
spots
all
that
that
documentation.
H
A
D
Right
so
this
would
be.
We
just
gone
up
women
doing
right.
So
the
first
thing
was
I
investigated
a
memory
leak
because
everything
seemed
to
be
filling
up
on
MPO,
no
ifs
and
it
would
fall
over
after,
like
half
an
hour,
turns
out
that
wasn't
a
memory
leak
good
like
great
news.
What
was
happening
was
writing
all
this
stuff
to
us.
Three
was
taking
too
long,
and
so
the
pool
stream
was
kind
of
like
the
it
was.
It
was
downloading,
all
the
stuff
from
NPM
and
then
having
no,
where
to
send
it.
D
Everything
was
just
buffering
up
in
memory
and
eventually
the
prices
fell
over
so
I
just
got
it
I,
just
tuned
it
a
bit
and
got
it
to
be
less
aggressive,
we're
just
downloading
and
that
willed
that
problem
went
away.
So
that
is
cool,
really
good
news.
A
few
minor
pr's
started
doing
the
Travis
migration
in
Phoenix
fairs,
em,
FS
rim,
full
log
post,
all
about
the
NPM
on
my
gfs
registry
and
added
some
monitoring
too,
as
well,
which
is
great.
D
I
asked
I
spoke
to
Jacob
and
Basco
about
it
earlier
last
week
was
that
it
seems
that
the
PIAT,
the
pub/sub,
doesn't
work
for
more
than
an
hour
on
IBM
on
NPM
anak
EFS,
which
is
really
weird
so
I
create
a
you
know.
Just
a
little
replica
of
it
and
I
haven't
been
able
to
replicate
the
problem
and
I
don't
Reno
how
to
debug
it
either
so
really
appreciate
a
bit
of
help
on
that
gonna
open
an
issue
so
that
we
can
talk
about
it,
but
yeah,
very
weird.