►
A
Again,
hi
everyone
welcome
to
the
I
profess
corporations,
weekly
sync
for
the
8th
of
february
2021..
I
have
aiken
brain.
I
will
be
your
host
we're
going
to
go
through
our
high
priority
initiatives.
Our
other
initiatives,
questions
pokemon
q,
a
all
that
good
stuff.
It's
going
to
be
amazing,
I'm
looking
forward
to
it
the
most.
A
We
have
a
note
taker.
It
is
lidl
we
can
start
with
the
high
priority
initiatives
releases.
Who
can
talk
about
releases.
B
Go
ipfs
rc2
is
out
assuming
no
bugs
are
found
specific
to
the
rc
that
will
turn
into
the
final
release,
but
try
it
out.
Let
us
know
if
there
are
things
we
missed
or
am
I
messed
up.
B
This
will
also
this
will
also
have
to
include
binary
releases
to
things
like
ipfs
ds,
convert,
because
it
requires
that
when
you
do
new
releases
of
things
with
repo
migrations
in
them,
so
that'll
be
out
when
the
release
is
out.
A
There
was
a
release
of
jsipfs
last
week
in
the
middle
of
hackweek.
54
was
released.
The
standout
feature
is
a
nat
traversal.
So
if
you
have
a
router
that
supports
upnp,
it
will
negotiate
an
external
port
and
forward
it
appropriately,
so
you're
externally
valuable.
This
is
amazing.
A
I
had
no
idea
if
we
hadn't
been
externally
valuable
this
time.
How
does
it
ever
work?
I
mean
crazy
anyway,
so
now
it
does,
which
is
nice.
There's
the
other
standout
thing
that's
worth
mentioning.
Is
it
upgrades
level.js
in
the
browser?
This
is
browser
only
node
is
not
affected
by
this.
In
the
browser
there
is
an
upgrade
for
level
chest
to
the
latest
version.
It
changes
how
it
stores,
keys
and
values.
It's
now
buffers
only
or
you
an
arrays
only
even
so.
A
There
is
a
repo
migration
that
will
take
place
when
you
first
run
jsipfs54
in
the
browser
in
node.
No
migration
will
take
place,
for
it
is
not
necessary.
I
say
that
it
will
change
the
version
number
of
the
repo,
but
it
will
not
do
any
significant
work.
This
will
only
happen
in
the
browser,
so
be
aware
if
you
are
deploying
js
ivfs
in
the
browser,
there's
also
a
patch
release
earlier
today,
with
the
fix
for
a
problem
with
the
abort
controller.
A
There
is
an
issue
I
will
link
in
the
notes
if
you're
interested,
please
do
read
it.
That
is
it.
C
C
E
B
Yeah,
that's
yes,
we
can.
We
can
close
up
shop
on
local
pinning
for
now
it
may
return.
It
may
return
in
the
future
with
more
features,
but
for
now
it's
it's
it's
done.
Does
the
thing
will
be
in
0.8.
F
A
G
Yes,
so
the
data
transfer
last
last
week,
we
did
we
were
mostly
doing
hack
week,
but
before
that
we.
G
Out
beyond
bit
swap
repo,
it's
looking
pretty
like
you
know,
completed
and
refactored
and
like
lovely
for
running
benchmarks
against
and
alex,
has
started
to
do
some
of
that
to
build
some
metrics.
We've
got
some
initial
stuff,
that's
a
little
funky
and
interesting,
so
we're
trying
to
figure
out.
What's
going
on
with
that,
I
mean
generally,
the
performance
of
a
graphic
is
looking
really
good,
but
there's
some
bizarre
inversions
with
bitswap
it's
certain
latencies
and
networks.
So
we
want
to
figure
out.
G
What's
going
on
with
that,
so
we'll
be
doing
that
this
week
and
yeah.
That's
it.
A
Awesome
next
up
just
increase
discoverability
and
connectivity.
H
Yes,
there
is
a
proposal
for
the
connection
manager
overall,
linked
in
the
notes,
according
to
the
planning
this
week,
we'll
see
when
we'll
actually
work
on
this,
but
the
chainsaw
folks
are
also
interested
in
some
of
these
features
for
lobster
which,
for
those
of
you
might
not
know,
is
the
jazz
client
for
it
too,
so
yeah,
they
might
probably
help
with
some
of
this.
If
this
is
a
priority
for
them,
we
are
still
talking
about
it
and
that's
it.
For
this
week,.
A
Fair
enough,
I
just
added
a
little
section
back
in
for
the
jstht,
because
last
week,
jacob
vashko
and
myself
spent
the
week
hacking
on
the
jsdht
and
we
added
types
which
is
the
thing
but
whatever
but
more
excitingly.
We
added
the
pre
like
seeding
k,
buckets
with
imaginary
piers,
which
are
uniformly
distributed
over
the
shah
2256
address
space,
which
is
cool,
ported
straight
from
go
and
also
improved
the
algorithms
around
the
eviction
of
of
peers
from
the
k
buckets.
B
Well,
the
cause
one
thing
I
brought
up
to
cause
all
the
the
nonsense
with
like
the
huge
like
break
a
few
bites,
a
few
bits
of
of
shot.
Two
isn't
required
for
client
mode
because,
like
basically
because
you
only
care
about
the
first
two
buckets
and
the
last
one
and
everything
in
the
middle
is
like
irrelevant
if
you're
a
client.
I
I
think
one
of
the
most
notable
things
if
you
haven't
watched
the
video
at
all.
We
took
the
phantom
drift
work
that
happened
like
last
year.
That's
the
libpdp
observation
deck
that
lets
you
hook
a
new
node,
and
that
was
one
of
the
the
things
that
we
started
working
on
there
so
that
you
can
actually
inspect
a
routing
table
inspect
connections.
This
will
be
reusable
in
go
as
long
as
we
implement
the
hooks
for
that
and
I
think
the
reple
repo
for
go.
I
Let
p2p
also
export
some
of
that,
but
one
of
the
in-progress
graphs
that
we
added
for
one
of
the
widgets
to
that
was
the
piezor
library
that
petar
put
together
last
year
for
inspecting
the
go
queries,
and
so
we've
started
trying
to
include
that
so
that
we
can
debug
our
our
js
queries.
Ui
still
needs
some
work
there,
but
super
excited
to
be
able
to
debug
the
queries
live
there.
A
J
A
big
release
coming
up
as
soon
as
I
finish,
integrating
it
with
the
jspfs
monorepo,
which
improves
a
lot
of
aspects
to
especially
typescript
real
typescript
ripples
like
the
php
noise,
there's
a
new
bundler
that
handles
the
typescript,
so
even
like
node
tests
in
lipid
pinoys,
which
used
to
register
a
hook
and
run
bubble
at
runtime
now
do
it
with
a
new
bundler,
which
is
super
fast.
So
even
the
mocha
tests
in
the
php
nodes
are
going
to
be
faster.
J
The
browser
tests
are
going
to
be
faster,
but
I'll
do
a
blog
post
or
something
with
the
new
stuff.
That's
coming
up
on
azure
you
all
can
read
it.
There's
no
need
to
explain
everything
right
now,
because
this
this
is
just
that
trick,
but
yeah
a
lot
of
stuff
is
coming
up,
but
during
this
week
I'll
like
the
types
in
our
reports
need
to
be
finished.
J
A
Try
and
finish
the
unix
of
vesta
this
week,
which,
like
it's
kind
of
getting
into
protons,
which
is
really
hairy,
and
I'm
going
to
look
at
like
there
are
pro
cool
buffer
modules
out
there.
That
will
generate
types
from
for
all
the
wire
transport
types.
So
I'm
going
to
look
at
maybe
getting
rid
of
protons
and
just
using
someone
else's
module.
So
we
don't
have
to
maintain
it
anymore,
but
that's
going
to
have
knock-on
effects
for
everything
else,
because
protons
is
used
in
the
p2p
for
doing
all
the
like
everything.
A
So
we're
going
to
have
to
then
like
roll
out
elsewhere,
and
also
things
like
the
new
multi-format
stack
has
like
its
own
bits
of
protocol
buffer
stuff,
like
for
reasons.
So
it
would
be
nice
to
not
have
multiple
types
of
protocol
buffer
module
in
the
dependency
tree
when
we're
making
browser
bundles.
So
yeah,
that's
going
to
be.
J
Interesting,
oh
yeah,
another
thing
to
mention
is
actually
me
and
alex
discovered
this
and
kind
of
fixed
it
in
the
interface
core.
We
we
have
a
doc
in
asia,
talking
about
all
the
typescript
stuff,
some
best
practices
that
we've
been
using
in
lipitp
ripples
and
ipfs
reports.
J
One
of
them
is
to
write
some
of
the
more
complex
types
in
a
external
file
like
types.ts
file,
and
we
we
started
doing
this
along
some
time
ago
and
now
we
have
like
multiple
repos,
but
we
actually
discovered
that
using
a
dot
ts
file
is
not
the
best
option
because
when
you,
some
of
you
may
be
familiar
with
this
when
you
regenerate
the
types
and
then
another
another
user.
J
Another
report
using
your
package
as
a
as
a
dependency
when
you
try,
when
you
run
the
typescript
compiler,
if
it
finds
a
dot
ts
file,
even
if
it
just
types,
no
call
there,
it
tries
to
compile
it
and
then
it
tries
to
resolve
the
imports
and
exports
and
stuff
like
that.
Then
some
errors
may
occur
with
the
imports
paths.
J
So
we
figure
that
out
and
the
the
easiest
solution
is
to
actually
stop
using
dot.
Ds
files
for
types
and
use
dot,
d,
dot,
ts,
the
type
declaration
files
for
the
types-
and
we
have
a
new
feature
in
azure
that
actually
copies
any
type
declaration
file
to
the
the
this
folder
that
the
the
one
that
gets
published
and
then
no
no
problems.
J
We
stopped
having
problems
with
the
wrong
paths
being
outputted
by
the
compiler
and
stuff
like
that,
and
actually
the
the
type
of
corrections
files
are
a
little
bit
better.
If
we
only
write
types
there,
because
we
can
define
classes,
abstract
classes
and
the
compiler
will
not
tell
you
that
you
can't
have
an
abstract
class
without
an
implementation.
J
So
the
all
that
stuff
that
we
create
interfaces
and
then
create
an
extra
interface.
So
so
we
can
define
the
constructor
there's
no
need
for
that,
because
we
can
declare
it
as
a
class.
So
everything
is
there,
the
prototype,
the
interface
and
the
constructor.
J
So
we
kind
of
change
the
change,
the
the
express
the
best
practices
in
the
in
the
repo.
That's
one
of
the
things
that
is
coming
up
in
the
new
age
here:
release
the
documentation
we
updated,
but
just
a
reminder,
because
this
a
lot
of
a
lot
of
reapers
already
use
these
files,
and
they
will
need
to
be
changed
to
avoid
these
types
of
errors.
With
the
the
pass
the
import
pass.
A
B
B
In
order
to
get
this
into
go
ipfs,
we
would
need
to
modify
some
of
the
existing
work
for
that
we
were
using
for
badger
2
so
that
we
have
like
a
more
sane
path.
So
we
have
things
like
ipfs,
init,
badger
ds
gives
you
the
latest
version
of
badger,
not
badger
one,
which
is
the
thing
that
you
would
get
right
now
and
then
and
then
just
have
the
repo
versions
and
the
data
store
spec
track
that
changing
over
time.
B
F
As
far
as
migration
is
concerned,
we're
going
to
auto
migrate,
or
I
mean.
B
I
don't,
I
don't
think
so.
I
think
we'll
probably
you'll
support
both,
because
I
I
would
hope
that
badger
3
is
better
and
more
stable
than
badger
one,
but
I
would
like
to
know
that
for
our
use,
that
our
users
experience
that
as
well
and
they
have
the
active
sds
convert
tool
that
should
help
them
migrate.
I
Yeah,
no
major
updates
harsha's
still
working
on
getting
the
existing
work.
The
whole
punching
coordination
work
merged.
So
that's
that's
some
progress
there
and
then,
at
the
same
time,
working
on
getting
some
some
scripts
implementations
together
so
that
we
can
do
some
live
testing
on
the
network
with
just
the
group.
It
appeals
home
routers,
so
we
can
play
around
with
that,
but
yeah.
A
Sweet
unix,
fsv,
1.5
and
co
ipfs,
I
don't
know,
will
ever
be
done.
A
No
update,
kypfs,
gc
improvement.
F
All
right
so
taking
a
new
approach
to
the
gc
improvement,
which
does
actually
involve
doing
reference
counting
for
blocks
and
that's
sort
of
a
big
change
from
what
the
previous
prototype
design
was
for.
F
So
that's
that
is
the
the
current
latest
thought
on
what
would
be
the
best
approach
to
this,
and
so
just
and
trying
to
get
this
prototyped
out.
So
we
can
other
proof
of
concept.
F
Mffs
improvements
that
was
actually
what
I
worked
on
for
the
most
part
of
the
hack
week
and
had
some
interesting
progress
there.
It
made
that,
for
the
most
part,
be
able
to
recognize
sorry.
Current
commands
can
recognize
mfs
namespace,
not
necessarily
any
impressive
for
demo,
but
one
of
the
things
that
was
did
get
implemented
was
the
ability
to
import
files
directly
into
mfs.
F
Again,
it's
not
terribly
impressive
because
it
allows
you
to
ipf,
has
add
and
add
files
directly,
but
the
nice
thing
is
that
you
don't
see
is
that
it's
all
gc
safe.
So,
instead
of
doing
an
ipfs
add
an
ipf
has
files
cp,
you
can
just
do
an
ipfs,
add
and
then
add
it
directly
to
an
mfs
namespace
and
in
a
gc
safe
manner.
It
transfers
all
the
files
directly
into
it
and
it's
so.
F
The
api
is
a
lot
nicer
that,
like
yeah,
the
user
experience
is
a
lot
nicer
because,
instead
having
to
go
through
two
commands,
you
go
through
one
simple,
more
intuitive
just
add
all
these
files
now
to
mfs
namespace
and
it
does
it
within
the
gc
lock.
So
there's
no
garbage
collector.
That's
going
to
interfere
with
that
that
process
of
getting
an
mfs,
so
that's
only
implemented
for
one
of
the
commands.
There's
a
few
more
to
go.
The
document
is
there:
it
describes
what
the
work
is.
F
Three
main
remain
points
are
the
name
spacing
stability
and
modification
of
commands,
but
I'm
not
going
to
go,
but
I
would
like
to
get
some
feedback
on
that
if
you
want
to,
if,
if
people
want
me
to
finish
that
effort
or
go
ahead
and
just
get
back
to
the
gc
work,
so
so
call
on
prioritization
leave
that
too
leave
that
for
for
later,
if
we
need
to,
but
that's
where,
but
that's
the
the
motivation
that
and
we
can,
the
the
stability
changes
are
again.
It's
change.
F
It's
a
minor
changes
than
how
locking
and
and
and
channels
are
used
for
signaling
within
the
mfs.
So
again
nothing
extraordinarily
outwardly
facing,
but
it's
all
sort
of
internal.
F
Oh
yeah
I'd
love
to
well.
Actually
I
don't
want
much
to
report
from
last
week.
It's
done.
There's
four
pr's
out
there.
I
it's
there's
probably
going
to
be
a
little
bit
of
you
know
there
might
be
some
discussion
or
some
changes
within
within
the
context
of
a
review,
but
until
those
reviews
happen,
I
would
welcome
anybody
to
look
at
it
and
see
how
that
stuff
works.
Try
it
out.
If
you
want
it's,
it's
a
complete
overhaul
of
the
migration
system.
So
how
about
it.
A
Awesome
so
next
is
the
ipfs
pub
sub
api
revamp.
D
Yeah,
so
it's
still
working
progress.
I
have
a
readable
stream
across
node
and
browsers.
Now
I
have
a
working
progression
that
builds
on
top,
but
I
still
need
to
finish
it
up
and
do
the
tests
before
I
can
submit
the
pull
request.
I
guess
I
could
submit
it,
but
before
I
consider
it
ready,
yeah
it'd
be
nice
to
see.
What's
going
on,
yeah
I
was
holding
off
because
that
readable
stream
sync,
I
was
kind
of
published
as
a
separate
library.
D
A
Super
exciting
memory,
leak
and
js
ipfs
says
no
update
and
we
denying
it
it
exists.
It's
convenient.
J
No,
I
just
paused
until
I
heard
back
from
them.
I
maybe
pick
it
up
when
I
finish
the
other
stuff
just
to
make
sure
we
don't
miss
anything
yeah,
fair
enough.
H
P2P
testing
set
up
yeah
no
no
play
this
week.
I
think
we
can
probably
take
all
these
together
with
the
dht
work,
as
the
dhtv
is
one
of
the
affected
ripples
by
that,
and
we
can
decide
what
to
do
for
pub
so
another
hippos
from
there.
A
Excellent,
this
brings
us
to
the
end
of
the
other
initiatives.
Now
we
move
on
to
all
the
any
other
business
kind
of
stuff,
so
design
review
proposals.
Does
anybody
have
anything
to
propose
for
a
design
review.
F
Well,
I
haven't
posted
it
officially,
but
if
anybody
wants
to
look
over
the
mfs,
what
I
propose
for
mfs
and
see,
if
that
makes
sense,
I
can
the
link
is
there
to
the
html
version,
but
I
can
also
put
the
hack
md
link
somewhere.
People
can
comment
on
it
or
whatever,
but
just
get
some
some
feedback.
It
should
be
a
really
light.
Read
it's
not
like.
It's
not
really
designed
it's
more
of
a
feature
proposal.
F
Yes,
there's:
oh,
I
can
put
the
hack
md
on
the
on
slack
if
you
want
to
make
comments,
direct
panned
out,
otherwise,
the
html
dock
is
linked
there
with
the
ipfs
link.
B
That's
cool,
I
just
put
it
in
the
just
copy
paste
the
hackmd
link
into
the
markdown
into
this.
The
meeting
that's
marked
on
file
away
yeah,
whatever
permissions
you
you
need
to,
but
that
way
anyone
else
who
is
watching.
This
call
can
also
take
a
look
all
right.
I
will
do
that
right
now,.
A
I
A
general
general
ask,
as
we
are
looking
at
reorg
stuff
coming.
New
teams
coming.
This
meeting
will
likely
get
dismantled
in
the
future
in
favor
of
other
forums,
and
so
as
an
ask
like
we'll
evaluate
this
as
we
get
closer
because
we
want
to
make
sure
that
we
have
appropriate
roll
off
time
should
count
on
starting
to
wrap
up
the
work
that
we
have
in
progress
over
the
next
few
weeks
and
not
taking
on
new
work
as
we
look
to
transition
into
those
groups.
So
keep
that
in
mind
as
you're
doing
stuff.
I
Don't
pick
up
stuff
start
wrapping
up
work.
We
should
stop
a
few
weeks
to
be
able
to
roll
off,
so
this
doesn't
need
to
be
abrupt,
but
there's
something
to
keep
in
mind.
A
D
I
have
a
question
so
during
the
hack
week
I
was
trying
to
do
something
that
I
was
surprised,
didn't
work
and
I'm
wondering
if
I
was
doing
it
wrong
so
idea
was
that
I
would
get
a
dog
node
out
of
for
my
cid
and
then
gonna
clone
it
and
replace
some
links
and
write
it
into
back
to
ipfs
to
get
a
new,
updated
version
of
a
tree.
D
D
Yeah,
I
know
so,
I
I
get
a
through
the
deck
api.
I
get
a
dog
pb
node,
so
I
then
call
remove
link
things
and
then
I
call
add,
link
things
to
replace
those
and
then
I
say
doug
put
and
that
you
know
that
seems
to
I
didn't
have
time
to
like
dig
into
what
was
not
working,
but
it
was
not
working
and
then
I
end
up
encoding
the
dag
node
as
a
block
and
saving
it
as
a
block
which
worked.
K
K
D
D
D
Okay
and
the
other
related
question
was
which
was
kind
of
based
on
this.
It
seems
to
me
that
it
would
make
a
lot
of
sense
to
be
able
to
create
a
mfs
out
of
cid
and
then
kind
of
do
the
all
your
file
system
using
insert
and
try
to
do
the
block
tag
back
and
forth.
Is
there
a
reason
why
there
should
be
one
mfs,
oh,
and
why
you
shouldn't
be
able
to
control
what
it's
mapped
to
in
terms
of
cid.
A
F
Yeah,
you
know
feel
free
to
propose
features
or
even
just
say,
discuss
them
as
something
that
might
be
related.
You
know
it's
not
directly
part
of
it.
Absolutely
if
we're,
if
we're
going
to
do
a
bunch
of
work
on
it,
now
is
the
time
to
figure
out
what
what
we
want
to
do,
or
at
least
what
what
we
should
do
at
some
point.
A
There
should
be
no
mfs;
it
should
just
be
like
here
is
a
cid.
I
happen
to
have
like
persisted
the
root
of
it,
and
I
have
that,
as
you
know,
there's
a
convenient
way
that
I
can
reference
that
in
the
future,
but
it
should
just
be
like.
Let
me
let
me
change
my
dags
in
a
way
that
I
am
familiar
with.
Like
a
cli,
you
know,
I
know
how
to
use
the
like
bash
to
make
a
directory
like.
Why
can't
I
do
that
with
everything
anyway
got
conscious
of
time,
but.
B
We
have
a
place
already
to
put
up
like
pointers,
tips
things
you've
run
into
when
trying
to
use
like
you
know
the
js
ipfs
http
client
with
electron,
or
I
don't
know
almost
anything
with
electron,
because
electron
is
special
right
like
these
are
things
that,
like
people
who,
whether
it
was
in
the
hack
week
or
just
otherwise
have
like
run
into
and
like
they
know
the
answers
to,
but
now
maybe
want
to
to
tell
other
people
about
the
answers
they
don't
have
to
discover
them
on
their
own.
A
And
the
docs
directory
in
the
js
ipfs
repo,
like
consider
good
place
for
like
electron.md
like
here,
are
the
here:
are
the
gotchas
that
kind
of
thing
cool
thanks.
A
Also,
a
link
from
loyal
in
the
chat.
B
Yes,
this
is
more
of
and
and
ask
or
or
something
like
that,
it
turns
out
that,
after
after
some
poking
around
snap
has
a
bunch
of
quirks
and
how
it
operates
and
scopes
things
and
what
you're
allowed
to
do.
That
makes
deploying
a
strictly
confined
snap
as
a
cli
tool
kind
of
annoying,
because
you
generally
expect
your
cli
tool
to
just
work
on
any
path
you
give
it.
B
But
snap
is
like
nope,
you
don't
exist
unless
you're
in
one
of
the
places
that
I
have
been
told
in
advance,
I
am
allowed
to
access
right
now,
we're
probably
just
going
to
put
some
docs
explaining
how
this
all
works
on
the
website
until
you
figure
out
exactly
how
you
want
to
do
this.
But
if
you
have
opinions-
or
you
know,
experience
with
working
with
with
things
that
live
in
snap
and
how
you
expect
them
to
work,
then
your
feedback
would
be
appreciated.
F
F
F
B
E
In
in
the
notes,
though,
I
mean
generally
having
some
design
plan
is
useful,
because
we'll
also
end
up
with
this
happening
on,
for
instance,
the
mac
sandboxing
and
ios
sandboxing
and
stuff
right,
like
you're
you're,
going
to
see
likely
more
and
more
cases
where
you
run
up
against
running
in
an
environment
where
you're
by
defaulting
an
isolated
thing.
So
having
the.
E
What
is
the
plan
for
how
these
things
work
when
they
find
themselves
in
that
environment
both
like
it's
probably
a
similar
set
of
communication
to
users
in
all
of
those
cases
of
hey
you're
running
into
sandbox?
You're
gonna
need
to
do
this
as
a
way
to
access
stuff
on
the
rest
of
your
file
system.
B
Yeah,
I
guess
I
guess,
there's
some
there's
some
like
user
diagnostic
things
there,
which
you
know
come
in
various
forms,
whether
it's
hey,
I'm
my
note,
is
behind
a
notch
or
you
know
hey.
I
can't
access
any
of
the
files,
because
my
thing
is
only
my.
My
thing
is
restricted
to
only
use
you
know
the
home
directory
or
the
home
mounted
media
directories
or
whatever.
A
Lot:
okay,
we're
done
amazing,
wrapped
up
thanks
for
coming.
Everyone
don't
forget
to
fill
in
your
async
updates.
I
know
it
says
people
are
very
naughty
and
they
don't
do
it.
You
should
do
it
I'm
naughty.
I
don't
I'm
gonna
do
about
alex
everyone
should
do
it.
Okay,
cool!
Thank
you
for
coming.
Everyone
we'll
see
you
next
week,
or
maybe
we
won't
bye.