►
From YouTube: go-ipfs Call - Q4 Planning
Description
This was the October 10th IPFS go-ipfs call that delved into Q4 Planning. See https://github.com/ipfs/pm/issues/217.
A
Could
you
guys
mute,
please
Richard?
Thank
you
all
right,
so
40
I
bffs
in
q4.
We
we
want
to
finish
up
our
endeavors
from
q3,
which
includes
IPL
d
and
a
few
other
things
like
the
core
API
and
some
GX
stuff
and
some
bandwidth
reduction
stuff
beyond
that.
What
we
really
want
to
do
in
q4
is
make
sure
that
we
clean
up
dwive
EFS,
make
things
work
better,
a
documentation,
make
it
easier
to
get
into
things
like
adding
more
tests.
All
over
the
board.
A
We
made
a
endeavor
for
q3
to
get
coat
higher
code
coverage
on
every
repo.
We
haven't
quite
made
that
time
constraints
are
difficult,
but
I
want
to
definitely
work
on
finishing
that
in
q4.
A
Oh
also,
directory
sharding,
would
be
really
cool
to
have
done
in
q4,
just
because
it's
mostly
done
already
yeah
so
trying
to
avoid
focusing
on
new
features
in
q4,
because
I
think
it'll
be
better
overall
for
everybody
involved.
If
we
just
try
to
make
it
better,
as
opposed
to
tacking
on
all
these
fancy
things.
So
that's
going
to
be
the
primary
motivation.
That
said,
we
don't
have
to
not
do
new
things.
I'm
just
protocol
labs
in
general
is
going
to
be
focused
on
just
improvement
over
all
the
code
base.
B
Yeah
I
think
there's
a
distinction
between
say
completely
new
features
and
say
things
that
are
sort
of
in
flight
that
have
been
understanding
stuff
for
a
while.
So,
for
example,
there
is,
let
me
put
something
like
private
networks
like
the
shared
private
network
thing
is
something
that
has
been
sort
of
in
client
has
been
discussed.
That
would
be
a
big
deal.
It
is
sort
of
a
new
feature,
but
it's
sort
of
in
a
gray
area.
B
Another
one
that
comes
to
mind
is
not
like
the
the
data
store
fast
search
in
the
file
system,
thing
that
has
been
implied
for
a
while,
and
we
need
to
do
a
lot
of
review
and
landing
that
and
getting
is
getting
this
ship.
So
that
is
something
else
up
for
consideration.
I
feel
like
with
no
counting
as
part
of
its
Nicole
right
now
and
I
think
he
he
would
want
to
do.
B
Work
on
this
quarter,
there's
other
things
going
on,
for
example,
I'm
talking
to
people
about
different
transport,
so
you
EP
and
click
cancel
potentially
being
able
to
ship.
Those
bills
would
be
mostly
integration
on
our
side.
They
would,
they
would
do
them,
and
then
we
would
be
like
when
it
rains
and.
B
So
these
are,
but
these
are
sort
of
like
new
things.
I
do
think
that
it's
do
that.
We
spent
a
good
chunk
of
time,
just
focused
on
improving
what
we
have
now
and
fixing
bugs
and
yeah
like
doing
things
like
the
massive
battle
is
reduction,
getting
a
more
stable
and
that
sort
of
thing
like
getting
happens
to
be
rock
solid.
A
D
Focus
on
I
mean
honing
in
on
figuring
out
priorities,
but
also
get
a
sense
of
what
what
it's
going
to
take
right
if
things
are
going
to
be,
if
it's
high
priority
and
it's
going
to
be
really
hard
versus
its
moderate
priority.
But
it's
really
easy
those
kinds
of
things,
because
that's
what?
But
it's
those
kinds
of
assessments
that
then
impact
where
things
land
on
the
road
map.
D
Right
I,
can
you
hear
you
know
yeah
I
can
hear
you
now.
So
basically
just
go
ahead
and
use
this
time
as
you
as
you
need
just
keep
in
mind.
The
main
thing
you're
wanting
to
get
out
of
this
is
figuring
it
out
priorities
honing
in
on
priorities
and
also
figuring
out
how
hard
certain
things
are
going
to
be
so
you
can
figure
out
where
and
how
to
slot
them
into
the
road
map.
But
right,
okay,.
A
D
A
Basically,
once
we're
done
with
these
calls,
so
hopefully
how
that
today
pushed
it's
the
what
the
fix
itself
is
already
merged
into
master,
and
then
we
have
a
patch
that
we're
applying
on
0
for
3,
and
that
may
not
be
the
only
thing
applied
there,
so
people
can
get
it
out
sooner.
I
also
have
a
tool
that
is
capable
of
potentially
recovering
lost
pins
if
you
haven't
found
a
garbage
collection.
A
So
if
you've
encountered
this
pin
bug
and
you
haven't-
run
a
garbage
collection,
you
can
run
this
tool
and
it
will
print
out
things
that
look
like
pins.
Sorry,.
B
Gone,
we
probably
should
leave
that
to
a
post
that
describe
all
this
sort
of
stuff
and
right
so
you're
talking
we're
diving
deep
into
like
discussing
like
tools
and
so
on.
We
probably
would
just
want
to
say,
like
super
high
level,
all
right
I
just
would
like.
A
To
make
sure
that
this
is
clear
for
anybody
watching,
because
it's
fairly
important
just
in
general,
it's
Oh
grab
you,
okay,
yeah,
so
that's
a
possible
thing.
So,
just
putting
that
out,
there
you'll
read
more
in
the
blog
post
when
we
have
that
out
yeah.
So
now
next
is
I.
A
Pld
I
build
us
coming
soon
we're
getting
things
prepped
and
ready
switching
over
to
different
libraries
for
addressing
things
and
then
going
to
have
the
Seaboard
IPL
the
nodes
available,
hopefully
quite
sick,
at
least
in
some
development
builds,
so
that
will
be
cool
beyond
that.
Kevin
has
been
working
on
the
file
store
code,
we're
getting
pretty
close
on
that
pushing
a
lot
of
compatibility
and
like
interrupt
changes
that
they're
making
it
easier
to
integrate
that
so
we're
hoping
to
have
that
in,
and
you
know
fairly
soon,
not
sure
the
exact
timeline
there.
A
F
F
A
So
yeah
we'll
get
that
done
next
up
is
just
in
general,
the
code
coverage
endeavor
that
we
have
before
q3,
we're
probably
I'd,
say
maybe
halfway
done
with
that
overall,
which
is
good
and
bad.
It's
a
lot
of
work,
there's
a
lot
more
work
than
we
expected
I
think
and
that's
partially,
why
it
hasn't
gotten
done
yet
so
we're
going
to
keep
pushing
that
in
q4
and
just
try
and
ship
more
tests,
which
is
never
a
bad
goal.
And
yes,
so
that's
roughly
it
for
the
q3
carryovers.
A
So
moving
over
to
like
Q
for
specific
things,
documentation
is
going
to
be
a
big
thing.
I
want
to
make
an
effort
to
well
make
it
easier
for
people
to
figure
out.
Oh,
how
do
I
make
something
with
ipf
s
in
various
different
modes
of
that
and
I.
Think
we'll
probably
talk
more
about
specifically,
would
want
to
see
there
later,
but
that's
going
to
be
kind
of
a
big
goal
for
us
and.
A
Yeah
so
having
that
that'll
be
big
on
our
timeline,
this
quarter
for
the
next
is
in
general,
just
clean
up.
So
that's
going
through
issues
fixing
bugs
trying
to
take
our
massive
473
issues
and
trim
that
down
and
close
things
out,
clean
things
up,
make
things
better
and
generally
just
pay
off
technical
debt
that
we
accrued.
A
This
will
be
really
really
really
really
good
for
the
code
base
and
I
think
it'll
provide
a
nice,
a
nice.
You
know
ramp
for
q4,
leading
up
into
q1,
where
we
can
really
start
doing
some
new
cool
stuff.
You
know
clean
up
the
codebase,
make
it
nice,
get
it
ready
and
then
moving
forward
from
there
it'll
be
a
good
jumping
boy,
I
think
kuba.
Do
you
want
to
provide
any
updates
on
anything
I?
Don't.
A
Pi
yeah
so
with
the
new
rescore,
kick
because
the
IPF
s
no
longer
we'll
just
spawned
thousands
and
thousands
of
threads
and
and
also
we
used
fewer
file,
descriptors
per
connection,
because
previously
we
would
use
to
file
descriptors
/
network
connection
because
of
the
way
we
use
part
was
working
now
we're
just
using
one
reused
port
for
the
whole
thing,
and
then
you
know
per
connection
one
file
descriptor.
A
D
G
A
Ok,
yeah
well
I
mean
it's
time
takes
long
to
go
over
this
list.
So
what
we
won't
want
to
see
in
particular,
is
better
test
coverage,
not
testing
of
networks.
Oh,
we
want
to
see
a
continuous
test
net
running
so
like
a
set
of
nodes
of
different
versions
and
different.
You
know
scenarios
running
user
loads
automatically.
A
We
want
to
start
looking
at
work
on
I'm,
actually
working
on
private
networks.
I,
don't
know
if
we're
going
to
prioritize
that
or
not
yet,
but
it's
definitely
on
board.
We
want
to
reduce
bandwidth
usage.
This
is
both
DHT
content.
Writing
changes
as
well
as
fixes
to
bit
slob
and
potentially
looking
at
using
pub/sub.
For
some
of
these,
we
do
definitely
want
to
merge
the
file
star
code.
We
are
moving
along
pretty
nicely
on
that
I.
Think
Kevin.
Do
you
want
to
say
anything
on
that
real,
quick
or
I'm.
C
Gonna
heal
me
yeah.
No,
not
really
I
would
kind
of
like
to
get
this
in
before
too
much
I
pld
stuff
gets
in
because
the
two
are
really
conflicting
with
each
other
and
every
time
you
mode
something
you
just
make
my
walk
for
me.
That's
just
like
a
I'm.
Sorry,
that's!
Okay!
That's!
Okay!
I'm!
Just
constantly
rebasing,
we
may
be
mixing
conflicts.
Yeah.
C
A
It's
it's
pretty
clean
now
so
liking.
It
sounds
good,
cool,
ok,
so
yeah
all
right.
No
under
the
most
requested
things
from
people
community
partners
in
general,
as
private
networks,
reduction
of
bandwidth,
reduction
of
resource
usage,
so
memory
and
CPU
reduction.
We
have
a
few
different
things
that
we
can
definitely
hit
with
this,
for
example
pierced
or
information
here
in
a
disc-
and
you
know
other
similar
things,
so
that's
somewhat
low
hanging
fruit
I
was
still.
A
So,
there's
a
lot
of
good
stuff
there
that
we
can
hit
and
lower
memory
usage,
improving
connectivity,
so
improving
connectivity
right
now
means
testing
connectivity.
We
need
to
figure
out
which
situations
we
actually
need
to
fix,
where
we
go
blindly
trying
to
fix
them.
The
guide
PFS
api,
so
we
need
we
need
to
figure
out
who
wants
to
captain
this
and
then
make
a
clear
path
forward.
I
think
the
best
thing
best
way
forward
for
this
is
to
get
the
core
API
defined
and
then
just
implement
the
go.
A
I
PFS
HTTP
API,
using
the
core
API
like
make
it
even
implement
that
interface
tour
integration.
So
this
means
working
with
open
bizarre
and
looking
at
the
tour
code
that
they
have
right
now
and
auditing
this
and
then
figuring
out
what
all
the
different
security
implications
are.
So
that's
definitely
on
the
two
lists:
datastore
and
file
store
yeah.
So
we
talked
about
that.
A
The
is
that
one
wrote
that
did
you
have
anything
in
particular
or
do
we
discuss
this
already
know
you
discuss
okay,
cool
and
then
sharding,
so
sharding
is
reviewing
and
then
getting
stuff
merged
and
ya
doing
that.
So
that's
yeah,
so
that's
a
not
a
huge
list,
but
it's
a
very,
very
important
list
and
so
I
think.
If
we
work
on
prior
cut
prioritizing
those
things
that
we
are
going
to
make
some
awesome
progress.
B
In
may
be
worth
sitting
that
whoever
can
going
to
reach
this,
or
maybe
we
can
make
a
general
note-
that
would
be
a
good
time
to
start
like
surface
think
we
will
go
through
the
issues
and
so
on
before
we
figure
out
like
the
final
milestones
list,
for
though
I
business.
But
if
people
have
certain
things,
they
really
want
us
to
remember
to
post.
G
Bosal
that
maybe
we
start
having
an
issue
like
q4
or
desires
or
goals.
I
can
just
be
sort
of
like
people
just
start
putting
in
stuff
or
should
I
do
issue
specific
things.
Hey
I
want
this
because
I'm
its
kind
of
duplicating
I,
don't
know
where
I
don't
know
how
the
user,
where
I
could
best
save
those
things.
Besides,
I'ma
get
my
receipt,
that's
what
I'm
asking
so
having
like
a
quarterly
issue
where
people
can
comment
might
be
a
good
idea.
G
B
We
could
do
that
I
alternately.
We
could
figure
out
a
way
of
trying
to
source
some
notion
of
like
importance
from
the
issues
list.
I,
don't
know
there
are
tools
that
do
this,
that,
like
will
parse
your
issues
and
then
tell
you.
These
are
the
ones
that
have
most
comments.
These
are
the
ones
that
have
most
like
you
know
like
the
reaction
of
votes
or
whatever
it's
something
like
that
might
be,
might
give
us
a
very
clear
view
until
it's
like
project
reasons,
horses,
something
like
project
real
is
so.
G
I'll
look
into
seeing
if
I
can
find
something
that'll,
actually,
log
reactions.
I
know
that's
a
new
get
in
feature:
I,
don't
know
if
that's
available
through
their
API,
but
that's
a
that's
a
really
good
point
and
a
good
question
so
I'll
just
say
that
all
again
to
that-
and
we
can
go
forward
from
that
I'll-
make
an
issue
on
I
PFS
community,
about
where
the
best
place
to
weigh
in
and
go
I.