►
Description
Meeting description here:
https://github.com/ipfs/team-mgmt/issues/992#issuecomment-576331029
For more information on IPFS
- visit the project website: https://ipfs.io
- or follow IPFS on Twitter: https://twitter.com/IPFS
Sign up to get IPFS news, including releases, ecosystem updates, and community announcements in your inbox, each Tuesday: http://eepurl.com/gL2Pi5
A
Okay,
it
records
the
had
is
in
the
notes,
if
you're
here,
please
put
your
name
on
the
attendees
list.
If
you
have
a
update
for
this
week,
please
add
it
to
the
bottom
of
the
PAP
for
asynchronous
review.
So
people
can
see
what
you're
up
to
we
won't
go
through
everything
you've
been
working
on,
but
if
you
have
initiatives
or
big
big
projects
and
things
that
I'm
not
listed
in
the
initiatives,
then
please
feel
free
to
add
them,
and
we
will
do
a
quick
update
on
those
alright.
So.
A
A
B
A
Nice,
the
other
it
for
me,
is
that
the
CIA
is
my
BFS.
0:41
is
now
green.
The
goal
of
our
tests
are
passing
four
browsers
for
Mac
OS
for
Linux
alex
has
got
the
majority
of
the
examples
tests
passing
so
that's
awesome
as
well.
I'm,
currently
working
on
just
getting
interrupt
straightened
out,
and
then
we
will
be
good
to
go
so,
hopefully
a
RC
this
week.
C
Those
tests
are
entering
tests
in
real
chromium
web
browser,
just
making
sure
you
can
like
add
files
to
idea
files
using
web
interface,
and
it
tells
a
lot
of
stuff,
including
HTTP
client,
just
a
different
go
idea:
files
in
the
repo
itself,
so
I
just
reused
those
end-to-end
tests
and
added
them
to
existing
CI
in
go
repo
and
enjoy
this
repo.
That
means
now,
when
you
make
a
change
in,
go
ID,
FS
or
Jess
idea
files,
one
of
tests.
That
means
to
pass
in
your
PR
our
regression
tests
against
ipfs
web
UI.
C
A
B
E
Some
of
us
specifically
describing
what
he's
doing
there.
It's
implementing
disjoint
paths
right
and.
B
Multiple
things
so
yeah
I
destroyed
paths
implementing
like
the
correct
kid
Omega
query
logic
he's
actually
making
a
pluggable,
so
you
can
have
like
three
different
like
remotely
query
logic
or
whatever,
because
like
they're
different
ways
of
actually
finding
Emily,
it's
not
like
paper
describes
it
sort
of,
but
they're
different
optimizations.
You
can
apply.
I
say
try
to
make
it
so
you
can
just
like
plugging
a
new
query
function
but
try
it
out
different
things.
It
also
gets
rid
of
things
like
that.
B
A
A
C
So
I
like
drafting
those
config
dogs
for
copy
first
repo
and
like
listening,
some
like
how
would
convict
look
like
for
those
common
use
cases
that
we
want
to
support,
don't
like
look
at
the
variable
names
will
probably
change
them
like
three
times
before
it's
ready,
but
it's
like
the
approach
I
want
to
do
here
is
ensure
we
don't
like
create
too
convoluted
configuration.
So
I
think
the
thing
that's
already
on
the
PR
I
had
seen
with
Steven
and
kind
of
too
complex
already
I
want
to
like
dumb
it
down,
make
it
even
simpler.
A
A
This
is
essentially
what
is
what
he's
just
to
summarize
that
for
the
people
who
are
watching
the,
this
is
configuration
for
like
if
you're
setting
up
a
gateway,
if
you
want
to
run
it
as
a
gateway
that
resolves
resolves
paths,
paths
based
gateways
so
like
ipfs,
the
bios
/
IDF
s,
or
if
you
want
it
to
be
a
like
hash,
dot,
I
professed
Ohio
gateway.
So
it's
resolving
it
as
a
subdomain
and
then
like
what
domains.
A
You
actually
want
it
to
be
working
on
and
stuff,
and
so
it's
configuration
allowing
you
to
set
up
your
goai
vessel.
Jess
I'll
be
first
node
as
a
in
as
a
gateway
in
your
own,
like
for
yourself
in
your
own
way
and
making
that
configuration
nice
and
easy
and
simple.
Is
that
fair,
yeah,
okay,
cool
all
right,
subdomain
gateway?
If
there's
no
questions
on
that,
then
we'll
move
on
to
distributed
signaling
still
holding
its.
A
B
G
Said
I've
been
looking
at
improving
pinning
performance
historian
pins
in
the
datastore,
and
my
update
here
is
that
I
discovered
that
the
muscle
memory
for
locking
your
keypad
on
a
laptop
is
not
the
same
as
on
a
keyboard,
and
if
you
don't
know
when
I
keep
on
your
laptop,
it
actually
ends
up
quitting
the
web.
Browser
that
you
have
open,
which
happen
to
have
my
proposal
for
the
right
of
my
proposal
for
the
pinning
performance
in
it.
So
that
was
really
I
will
rewrite
that
this
way.
A
Firefox
only
alright,
anyway,
that's
the
side.
The
migration
tamati
hash
keys
in
the
block
store
on
the
jer
side
still
on
hold,
because
we're
still
focusing
on
getting
async/await
out
the
door
and
apparently
go
still
needs
a
migration
to
be
written
for
this
change.
But
everything
else
is
in
place
right.
A
Things
are
in
place-
probably
yes,
excellent.
Okay!
So
that's
that
this,
what
updates
Dirk?
You
have
an
update
here.
Yes,.
E
The
last
week
I've
been
working
on
a
test
suite
in
test
ground,
so
I
want
to
get
to
the
point
where
we
can
test
out
a
few
different
kinds
of
environments
like
a
data
center,
internet
style
or
an
internet
kind
of
environment.
Just
so
we
have
confidence
when
we're
making
releases
of
bits,
one
so
yeah.
It's
probably
gonna.
Take
me
another
week
or
so,
but
making
progress.
A
F
We
are,
there
were
some
issues
with
the
browser
examples.
I
was
actually
hitting
issues
that
other
folks
weren't,
because
I
didn't
have
some
of
the
bootstrap
node
intermediary
certificates.
All
he
was
tackling
those
and
think
everything
is
starting
to
look
much
much
better,
so
I've
got
a
browser.
Example
is
almost
done.
F
I
revamp
tit
because
it
was
kind
of
annoying
in
crappy,
and
so
it
should
be
better
now
so
I'm
just
finishing
that
up
a
few
PRS
for
that
then
I'm
gonna
redo
the
do
final
pass
in
the
getting
started
guide
and
then
I
think
we
have
one
bug
with
pub/sub
in
Libby,
TP
Interop
I
should
be
able
to
get
that
fixed
and
hopefully,
tomorrow
and
then
I
think
we
should
be
hopefully
good
to
release.
Sometime
later
this
week,
fingers
crossed.
A
You
know,
establishing
connections,
SEC
IO,
sending
files
receiving
them
all
of
the
PT
stuffs
and
it
works,
and
what
I
was
seeing
on
my
laptop
was
that
between
the
current
version
that-
and
this
version
we
have
like
things-
are
faster
and
when
the
file
is
bigger,
things
are
a
lot
faster.
So
at
the
moment
we
have
this
problem
where
we've
got
this
kind
of
exponential
slowness
curve
where,
as
you
add,
bigger
files,
things
get
slower
and
slower,
but
now
it's
more
it's
way
more
linear.
So
as
you
well,
it's
kind
of
a
constant.
A
You
can
say
that
like
JSI
BFS
is
like
10
times
slower,
but
it's
not
like
10
times
slower
when
when
it's
got
win
at
one
make
file
and
then
100,000
times
slower
when
you've
got
a
hundred
Meg
file,
it's
like
ten
times
the
way
whatever
so
there's
there's
a
constant
there,
and
so
that
makes
me
feel
a
lot
better
about
the
performance
because
it
seems
like
there's
there's
some
like
constant
thing
that
we
need
to
just
eradicate,
and
then
we
can
bring
that
down.
So
things
are
already
faster.
A
According
to
the
tests
on
my
laptop
that
we
wish
I'll
see
as
I
finish
off
the
interrupt
test
and
start
running
it
in
COI.
So
in
those
particular
examples
things
are
looking
good.
So
that's
cool
and
that's
the
kind
of
async
await
update
I
have
for
JSI
BFS
other
than
like
we're
we're
getting
there
slowly
getting
there
and
interoffice
happening
yesterday.
A
I,
so
this
is
these:
are
tests
that
test
between
go
to
go
and
go
to
j
s
and
j
s2
j
s.
I
think
they're
all
done
over
this
over
tcp,
your
WebSockets
I
can't
remember.
Quite
it's
probably
TCP
in
the
browser
there
will
be
WebSockets
yeah
they're,
pretty
tests.
I
realized
that
just
saying
compared
to
the
ones
that
were
running
earlier.
They
are
better.
A
A
A
And
seeing
as
I'm
recording,
it's
probably
lost
for
everyone.
Okay,
right
now,
I've
lost
the
window
with
the
things
in
it:
okay,
cool
cool
Peter
stream
stream
content
based
chunking
research,
him
previous.
H
H
H
Plan
is
to
write
up
a
write,
write
all
of
this
up
into
a
preliminary.
This
is
what
looks
about
right
and
then
the
plan
after
that
is
to
let
this
thing
run
continuously
over
what
seems
interesting
and
see.
Basically
what
what
was
out.
That's
why
the
preparatory
work
to
be
able
to
an
experiment
as
fast
as
possible.
Modernist
I
also
discovered
some
problems,
as
I
was
going.
You
know
through
through
that
with
GSA
be
present.
Why?
H
A
H
Mean
even
after
my
work
eat,
so
we
definitely
will
look
at
different
defaults,
because
there
are
things
that
we
could
be
doing,
that
we're
not
doing
right
now
like
before
doing
any
kind
of
size
order.
Transcendentalism
like
that.
Look
for
any
kind
of
padding
like
8
or
16,
or
more
identical
bytes,
and
only
after
that.
Do
everything
also
to
8
to
12
inch
right
now,
which.
H
H
B
What
we're
building?
Yes?
Yes,
so
it
was
really
to
make
a
new
chunker.
It's
gonna
like
the
ultimate
content
or
early
content-based,
chunker,
sorry
ultimate
general
purpose,
content-based
junker,
then
switching
this
the
default
is
probably
going
to
come
along
with
switching
which
other
things
the
default
we've
discussed
several
times,
having
like
a
importers,
version.
D
B
Of
thing
or
like
you
can
supposed
to
get
pro
version,
zero,
which
is
current
system
work
for
import
over
from
one
which
would
be
like
all
the
nice
features
we
have
right
now
turned
on,
and
you
quarter
version
to
be
like
that
plus
UNIX
ms2
and
all
that
kind
of
stuff.
Then
we
use
this
could
just
like
say:
okay,
like
I've,
missed
with
a
checkpoint
use.
This
version
I
was
thinking.
A
D
Wonder
I,
don't
I,
don't
think
I've
sinned,
this
list
of
endeavors
change
very
often
now
might
be
a
good
time
to
review
the
list
with
the
roofless
focus
on
content
resolution
speed
and
just
as
a
quick
pass
I
thought
maybe
distributed
signalling,
IDNs
and
migration
to
multi
hash
keys
and
box
or
might
all
be
endeavors.
You
might
consider
pruning
with
more
fluid
flow
time
to
content
resolution
and
anything
blocking
it.
B
Here
I've
been
waiting
for
the
finish,
the
plan
for
a
conic
resolution,
so
we
actually
have
a
set
of
tasks
like
okay,
and
these
are
down
lists.
I'll
talk
to
world
later
on
today,
to
like
figure
out,
we
can
you'd
like
at
least
an
early
version
of
that
ASAP
in
terms
of
like
removing
different
tasks.
I
like
actually
I
recently
added
the
migration
thing,
primarily
cuz.
B
B
Yeah
I'm
gonna
go
ahead
and
remove
all
okay
yeah.
We
should
probably
remove
your
eye.
Penis.
Add
performance
is
probably
not
as
important
with
moment
migration
of
all
tasks
easier.
We
don't
there's
no
reason
to
actually
keep
this
on
the
list.
It's
kind
of
a
back
for
anything
that
might
happen.
We
can
have
to
connect
hock,
update
it
swap
updates,
that's
an
in
progress
thing
to
be
kind
of
Al
and
basically
obviously
UNIX
fest
1.5
yeah.