►
From YouTube: ⚡️ ⒿⓈ Core Dev Team Weekly Sync Jun 25, 2018
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
It's
got
a
new
name,
so
that's
cool.
We've
got
a
note-taker.
That's
good
I!
Remember
that
for
once,
shall
we
do
a
round
of
updates
like
we
normally
do
yes,
good,
okay
from
me,
I
have
whatever
done
this
week
investigated
in
race.
Yes,
so,
okay,
you
might
have
noticed
that
our
JSI
PFS
api
tests
have
been
failing
on
Windows
for
about
a
month
or
two
I've
sort
of
been
investigating
this
off
and
on
for
awhile
and
I've
come
to
the
conclusion
that
I
should
just
open
an
issue
won't
go
out
FS.
A
So
if
you're
wondering
why
that
was
that
that
was
happening,
then
that's
that's.
Why
I
rip
in
the
PR
you're
welcome
to
he'll,
never
look
yeah!
That
was
that's
been
interesting
to
to
track
down
and
also
I'm,
not
a
native
Windows
user.
So
it's
been
fun
I've,
we're
done
so
I.
Did
it
oh
yeah
right,
so
the
live,
p2p
connection
manager,
stuff
got
merged
in
it's
not
yet
been
released.
I've
added
a
PR
to
expose
that
configuration
so
that
when
chase
ipfs
gets
released.
A
Next
people
can
start
using
the
connection
manager,
so
Jeff's
ipfs
API.
There
is
a
fix
for
getting
a
block
that
has
no
data.
I've
just
released
a
new
version.
Yes
I
profess
API
to
fix
that
I
added
tests
and
found
that
that
was
also
a
problem
in
chess
ipfs.
So
I've
also
created
a
PR
to
J's
I
profess.
To
fix
this
issues
with
getting
a
block
with
no
content
in
it,
the
pin
the
pin
stuff
got
merged
in
hooray.
A
Yes,
the
problem
that
David
has
noticed
was
that
we
had
there
were
some
tests
that
were
running
in
JSI
PFS.
Well,
there
are
some
tested,
yes
I
confess,
but
they
weren't
running
they
weren't
being
required,
so
it
all
was
not
lost
because
the
interface
like
the
first
core
tests
were
still
wearing
one,
but
the
one
said
that
had
been
written
and
were
in
JSI
PFS
mysteriously
were
never
being
run
so
david
open.
A
The
PRT
enable
those
and
I
tweaked
to
fix
it
and
most
it
and
no
and
we'll
never
know
oops
anyway
and
I'm,
not
I'm,
not
blocked
on
anything.
I've
got
a
few
next
up.
I
I've
got
a
few
things
to
do
on
the
PR
for
modular
interface
IVFs
core
tests.
It
looks
as
though
everyone's
kind
of
happy
with
that.
So
that's
great.
Thank
you
for
commenting
if
you
did
reviewing
whatever
I
think
so.
The
main
easy
to
do
on
that
is.
A
David
has
requested
that
when
we
skip
tests
there
being
some
sort
of
description
printed
to
the
console
for
wires
being
skipped
and
the
thing
that's
quite
sensible,
so
I'm
gonna
see
what
I
can
do
to
to
enable
that
and
then
what
else?
So
it's
0
30
is
the
thing.
I'm
sort
of
focused
on
connection
manager
got
merged;
it
just
needs
a
liberal
version.
A
David
requested
that
we
also
try
and
get
in
the
p2p
contact
is
PR
for
updating
the
wider
config
works
for
the
p2p,
so
I've
got
to
have
a
look
and
maybe
get
back
in
and
then
again
for
a
second
we
running
I
might
get
around
to
reviewing.
Imagine
me
and
the
fester
for
the
next
release.
A
That's
cool:
who
is
who's
next
err,
so
you're
up?
Yes,
so.
B
I
was
working
on
the
crafting
stuff
in
a
so.
The
exciting
thing
is
that
I
kind
of
get
it
working
that
I
can
get
one
one
file
like
with
one
call
getting
a
full
file
without
any
back
and
forth
between
the
little
peers,
and
it
was
a
quick
heck.
So
it's
not
that
useful,
but
at
least
I
get
an
idea.
So
now,
let's
talk
about
I,
don't
know
like
10x
faster,
so
not
that
much
and
socially.
The
graph
sync
won't
make
things
faster,
because
we
have
lots
of
other
brunettes
in
JSF
FS
anyway.
B
It
gave
me
a
good
understanding
of
like
how
all
the
cool
stream
stuff
anything.
So
it
was
worth
the
effort
and
I
finally
think
I
know
what
grafting
is.
The
quick
summary
is
what
I
would
want
to
concentrate
next
quarter
on?
Is
it's
not
about
reversing
graphs
is
not
about
making
things
fast.
It's
about.
You
can
send
one
C
ad
with
some
meta
information
and
it
will
return
several
blocks
and
that's
the
difference
between
bit
swap
using
bits.
Puppies
city
requests
one
blog
getting
it
in
one
book
bag.
B
So
the
difference
is
between
sending
one
block
with
some
meta
information
getting
several
blocks
back.
This
is
what
I
will
concentrate
on
it's
someone
super
straightforward,
but
it
isn't
yes.
So
that's
the
core
of
crafting
I!
Think
and
yes-
and
the
next
thing
I
will
do-
is
writing
this
down
and
also
playing
a
round
a
bit
with
really
p2p
to
see
how?
B
Well
it
will
work
with
graphs
and
stuff
I
will
work
on
and
another
exciting
thing
is
that
you
Richard
Snyder,
is
working
on
the
IP
of
the
format
changes
that
I
did
so
the
spec
was
updated
quite
heavily,
but
nothing
of
it
was
implemented
and
Richard
now
works
on
the
implementation.
So
that's
pretty
cool.
C
It
was
a
simple,
simple
hag
makes
things
a
little
bit
faster,
at
least
when
we
have
all
the
blocks
already
in
the
in
the
browser,
because
when
we
are
getting
locks,
nothing
else
matters,
but
the
overhead
for
crypto
I
try
to
improve
in
several
ways.
But
everything
comes
down
to
two
Krypto.
That's
why
I
try
to
focus
a
little
bit
more
on
that
stuff.
C
Did
a
couple
of
hag's
trying
to
make
web
crypto
worth
a
try?
It
also
some
webassembly
stuff
kind
of
in
the
middle
in
the
middle
of
things,
but
yeah.
That's
it.
What
I
have
done
the
previous
week
this
week,
I'll
try
to
all
do
all
the
retrospective
stuff
on
planning
and
also
at
least
make
the
web
crypto
work
with
a
local
Lehman
I'm.
Having
a
couple
of
issues
with
that,
and
also
I
will
look
into
one
issue
that
someone
posted
on
Jess
ipfs,
epi
not
being
able
to
end
all
kind
of
small
file.
C
C
C
C
This
one
not
much
to
see
just
like
some
broth,
it's
faster
I
can
run
it
with
different
blocks.
This
is
actually
getting
at
least
one
walk
and
running
the
benchmark
with
the
two
implementations.
So
I
can
run
with
all
these
these
files
and
it
will
be
consistent
at
being
like
a
tree
at
least
3x
times
faster.
C
C
The
the
plot
above
Jas
class
I
think
we
also
do
that
in
some
of
the
the
repos
that
we
use
protobufs,
but
even
though
I
think
we
don't
do
this
in
the
in
UNIX
ipfs,
it
will
still
be
faster,
even
if
we
add
at
least
that
so
this
will
probably
a
good
addition
to
an
exact
TFS
and
the
other
demo
I'd
like
to
show
you.
We
just
get
the
task
manager.
C
C
E
C
Another
one
is
the
using
this
new
lab,
it's
not
new,
but
it's
a
different
library
for
and
Link
auto
works
and
the
other
one.
It's
the
download
thing,
because
you
know
the
demos
that
we
have.
We
we
we
couldn't
handle
big
files
when
we
get
the
blocks
and
assemble
everything
into
one
buffer
and
then
we
just
throw
it
to
the
user
in
with
one
big
blob.
C
If
it's
like
Andrew
bites,
it
doesn't
work
right.
So
it's
a
couple
of
different
people,
different
things,
this
one
about
the
Downloads.
It's
just
about
streaming
it
everything
and
using
a
serviceworker
to
catch
the
request
and
build
up
a
stream
and
respond
through
the
service
word
to
the
to
the
main
thread
and
the
download
is
it's
in
streaming
mode.
Basically,.
E
Is
there
any
changes
to
be
worried
about
or
concerned
about,
in
terms
of
like
how
malformed
data
is
handled
between
the
old
protobuf
tool
and
the
new
library
you're
using?
Can
you
beat
that
please?
Are
there
any
changes
to
how
malformed
data
is
handled
like
have
you
been
testing
that
between
the
two
protobuf
libraries
yeah.
C
C
C
The
schema
class
this
auto
generates
auto
generated
by
the
the
BBF
package.
So
it's
basically
the
same
thing
and
I've
been
testing
with
the
two
implementations
and
everything
works.
I
constantly
check
if
I
mostly
work
with
the
videos,
because
it's
the
easiest
way
it
like
to
check
if
the
final
assemble
file
works
and
it
it's
been
working,
so
I
didn't
turn
in
any
any
issues.
I.
E
Guess:
I
guess
what
I
mean
or
what
I
was
trying
to
ask
is
since
we're
using
a
different
library
with
different
tools,
do
the
errors
and
things
you
get
back
in
the
case
of
malformed
or
bad
data
change
at
all,
because
that's
part
of
the
public
interface
that
users
ultimately
see
if
anybody's
paying
attention
to
a
particular
type
of
error,
rather
than
just
I
got
an
error.
Does
that
change,
or
do
we
have
anything
more
specific
or
better
to
say
about
errors
now
than
we
used
to
oh
I.
C
A
F
What's
up
guys
last
week,
I've
been
continued
with
my
yard
of
the
errors.
I've
never
smoked.
Both
requests,
I
bring
some
error,
messages
and
I
see
with
the
vid,
and
he
said
to
open
a
big
issue
like
a
road
map
that
there's
everything
that
is
related
to
diamond
gracious
and
error
handling.
So
we
have
like
a
table
with
multiple
issues
and
we
can
track
them
if
it's
possible
to
reproduce
the
errors
manually
or
if
you
have
as
well
flipping
those
those
errors
and
stuff
like
that.
F
A
A
A
D
D
I've
also
created
the
new
package
called
GSI
Venice,
which
is
kind
of
a
disgusting
name
right
now,
because
it's
called
oh
I
P&S
that
package
the
same
packaging
go,
but
we
don't
know
that
it's
the
best
name
or
not
so
I'm
discussing
it
with
you,
the
riddance
there
Walden
in
order
to
figure
it
out,
then
I've
merged,
two
peers
for
CTO
and
I'm
blocked
in
that
I
can
as
modules
in
naming
definition,
and
for
this
week,
I
need
to
add
a
nano
seconds.
Precision
to
the
records
because
actually
go
I
been
s
uses
it.
D
But,
yes,
the
native
data
does
not
have
nanoseconds
so
I'm,
trying
to
figure
out
where
it's
the
best
way
to
do
that.
I've
checked
some
miracles
and
big
number
and
I'm
thinking
about.
Where
is
the
best
way?
I
will
try
to
get
PR
tomorrow.
I
guess
then:
I
need
to
finish
some
tests
in
the
ipfs
in
Jessup
FS
for
the
my
IP
and
s.
D
A
D
So,
basically,
when
you
have
a
penance
working
locally,
you
can
have
publish
and
resolve
I
penance
names
locally.
To
that
note,
there
is,
there
is
no
routing,
and
if
you
connect
a
new
peer,
it
will
not
work
the
next
step
after
having
the
I
penis
working
locally
and
which
was
like
my
second.
We,
a
cure
for
a
penis,
is
to
add
the
DHT.
A
D
G
I've
added
interrupt
tests
last
week
for
the
private
network
that
found
some
issues
both
with
order
of
go
dialing,
first
or
Jas
tiling
first,
so
I
fixed
some
issues
with
that.
The
only
outstanding
item
I
have
is
when
there
are
bad
network
connections.
The
J's
poll
reader
doesn't
handle
over
reading
very
well.
G
It
will
sit
there
and
wait
60
seconds
so
in
an
effort
to
not
have
60
second
test
timeout
I'm
working
on
a
fix
for
a
poll
reader,
so
that
will
just
hang
up
into
an
error
when
it
should,
instead
of
just
soccer
typing
out
so
I
think
I
have
that
mostly
working
I
just
need
to
do
some.
Some
final
tests
on
that
I
demoed,
the
private
network
at
the
IP
all
hands.
G
So
if
you
want
to
see
any
of
that,
it's
basically
just
Wireshark
in
action,
but
that's
there
and
then
this
week,
I'll
fix
the
the
poll
reader
stuff
finish,
submitting
all
the
PRS
for
private
networking
and
then
work
on
some
q3.
Okay,
our
stuff,
as
well
as
try
to
finish
up
the
rest
of
the
Lib
p2b
config
updates
with
David
questions.
H
H
I
started
by
just
trying
to
use
against
the
running
demon
and
got
loads
of
errors.
I
thought,
oh
no
I
know
if
there
are
tests,
run
the
tests
and
there's
any
one
test
that
actually
tests
the
implementation,
the
rest
of
lots
of
stubbing,
and
you
know
that
test
doesn't
work
either.
So
I
was
hoping,
there's
going
to
be
a
simple
case
of
just
switching
between
go
and
J
s,
implementations
at
the
demon,
but
the
project
itself
needs
coil
or
work
to
get
it
working
again.
First,
so
I
was
trying
to
do
some
of
that.
H
Interestingly,
when
it,
when
it
first
spins
out,
it
tries
to
contact
a
node
that
doesn't
exist
anymore
and
download
a
IP
NS
thing.
There
is
a
mapping
of
all
NPM
nodes
too,
so
all
NPM
modules,
two
content,
hashes
and
that
doesn't
exist
either
and
I'm
not
sure
how
you
recreate
that
there
doesn't
seem
to
be
anything
in
the
rebate
about
how
you
would
generate
and
publish
that.
But
it's
been
done
so
somebody
must
know
so.
I
was
going
to
try
and
get
a
bid
David's
time,
so
you
can
take
me
through
how
it
works.
H
H
B
H
E
Not
a
whole
lot
to
report
I'm,
you
know
it's
been
I've,
been
spending
the
last
couple
weeks
pushing
on
people
to
try
and
redefine
some
of
the
language
around
IPL
DS.
So
it's
much
more
clear
and
to
do
things
like
comment
on
a
lot
of
pull
requests
and
push
towards
having
better
in
code
documentation
in
Godot
generation
for
the
ghost
stuff
add
diagrams
to
the
various
builders,
and
things
like
that.
E
E
I
A
I
A
J
A
I
don't
know,
but
presumably
that's
where
you
typed
your
words,
so
how
you
fought
it
will
win
and
I'm
sure
it
went
swimmingly
for
everyone,
but
do
write
what
you
feel
in
there.
So
then
there
is
also
the
second
thing.
If
you
can,
please
update
your
final
score
in
the
spreadsheet,
for
your
four
quarter:
okay,
ours-
that
should
take!
You
know
a
long,
hey!
Well,
hopefully,
won't
take
you
very
long
anyway,
and
then
the
final
third
and
final
thing
is
to
maybe
have
a
think
about
what
you
would
like
to
do.
A
What
you
feel
david
says
think
very
deeply,
so
we
should
all
do
that
about
what
your
goals
for
the
next
quarter
are
and
then
add
them,
as
infamy
like
as
I
guess,
a
comment
on
the
rest
on
the
existing
progress.
That's
there
yeah
and
then
I
think.
Unless
anyone
has
any
other
questions
or
agenda
items
or
I
didn't
go
down
to
the
one
but
yeah
no
there's
nothing
else
added
anything.
Does
anyone
have
anything
else,
any
other
business?
Are
we
done.