►
From YouTube: 📦 Package Managers WG Weekly Sync August 6, 2019
Description
Weekly meeting and sprint planning for the IPFS Package Managers working group - https://github.com/ipfs/package-managers/issues/1
B
B
B
Fair
I'll
get
this
together
for
next
week.
Don't
worry
the
so
this
recording
is
the
first
inaugural
combination,
package
manager,
task
force,
sprint
planning,
along
with
the
usual
weekly
that
we
were
having
after
so
that,
hopefully,
all
the
information
comes
together
into
one
place
and
everybody
gets
it
more
efficiently
and
we
have
a
little
bit.
You
know
half
hour
less
meetings
every
week,
which
is
great.
B
We
already
talked
about
how
andrew
is
going
to
move
to
you
back
to
the
original
package
managers,
perfect
concept,
research
role,
but
he
had
filled
last
quarter
to
discover
amazing
new
things
about
how
and
distributed
systems
might
work
better.
With
our
might
help
package
managers
do
their
job
more
efficiently,
effectively
and
happier
maybe
and
Steven
and
Molly,
then
will
be
spending
more
time
in
this
in
our
group.
Helping
DRI
helping
lead,
leave
the
work
otherwise,
so
that
Andrew
has
time
to
go
back
and
do
the
work.
B
Okay,
cool
well
also
I'm
sure
iterate
a
little
bit
on
the
process
for
the
stylet.
These
are
combined.
Molly
suggested
that
we
get
a
note
stock
together
or
some
sort
of
agenda,
so
people
who
drop
in
on
the
call
will
know
what's
going
on
or
what
to
expect
and
then
maybe
know
if
they
should
drop
in
on
the
call
at
all
so
we'll
see
about
that
for
next
week.
B
In
the
meantime,
stick
with
what
we
have
been
doing,
which
we
needed
to
change
a
little
bit
anyways,
but
we'll
take
a
peek
which
was
looking
at
our
design
hub
board,
which
is
sort
of
Fez
most
of
the
stuff
that
people
are
working
on,
except
some
hubs
strange.
So
we'll
see
how
it's
going
check
in
the
review,
so
I
want
to
check
in
the
review
QA
column
first
and
go
down
those
and
just
see
where
we're
standing.
B
If
any
of
that
needs
to
be
reviewed
or
something
can
be
finished
off
and
moved
to
done,
I
know
the
package
manager,
q3.
Okay,
ours
are
not
done
because
that's
on
me
to
work
with
Molly
and
other
people
this
week
to
take
a
look
and
finish
out
whatever
it
needs
to
be
finished
out,
because
I
think
we
left
those,
maybe
not
completely
done
and
Andrew-
is
that
your
understanding
also
I,
think
it
looks
where
we
left
it.
Yeah,
okay,
so
I'm
not
gonna,
even
open
that
one,
because
we
know
it's
not
done.
B
B
B
C
Hi
and
so
Michael
did
his
report
and
that's
up
and
available,
and
so
there's
a
recording
from
it.
But
I
didn't
see
any
additional
issues
created
out
of
that
work
and
so
I
think
it
might
be
worth
it.
E
C
A
F
So
I
guess
jumping
ahead
here
to
kind
of
were
on
a
the
current
plan,
for
that
kind
of
stuff
is
to
basically
get
out
an
experimental
read-only
version
and
get
that
merged
into
ipfs
somewhere
and
then
basically
to
get
us
up
to
speed
to
where
we
would
need,
or
what
we
would
need
to
be
able
to
do.
Our
sync
stuff
would
basically
be
write,
support
in
a
not
super
complicated
way,
but
not
in
a
trivial
way
either.
G
A
A
I
mean
the
the
most
in
yeah,
the
most
important
one
that
I'm
working
on
is
the
first
one
which
links
to
more
PRS
the
other
site.
The
second
one
is
is
sort
of
back
burner
for
now,
basically
tommix
two
other
priorities
and
Hugo
was
gonna,
keep
working
on
his
stuff,
but
he's
on
an
under
prep
ne.
We
haven't
put
him
under
any
pressure
to
do.
Spec's
work
for
Dominic
is
not
just
more
important
and
then
once
either
and
maybe
a
little
bit
in
parallel,
but
mostly
once
I'm
done
with
this
first
thing.
A
H
G
B
A
B
Policy,
that's
right.
We
talked
about
at
retro
last
Friday,
making
sure
that
world
we
all
think
we're
doing
the
most
important
things
I
wish
be
doing
so
we'll
keep
an
eye
on
that
too.
If
you
feel
like
you're,
being
pulled
in
50
different
directions
and
need
more
like
no
really
drop
that
other
stuff
and
do
this
stuff
over
here
we
can
talk
about.
F
I
F
Then
yeah
it's
just
at
the
moment.
I
mentioned
it
earlier
that
we're
just
trying
to
get
this
like
file
system
exposed
on
the
daemon
in
a
read-only
mode
that
we
can
extend
gradually
so
that
we
can
work
towards
features
like
one
for
our
sink
and
and
other
things
that
come
up
as
we
need
them
and
just
I
guess
I
should
mention.
There
was
a
small
setback
this
week,
just
because
we
were
targeting
one
library.
F
B
F
B
F
B
F
Should
I
should
mention
I,
guess,
jumping
back
to
what
a
Dean
said
about
some
of
the
DNS
stuff
and
all
this
back
work
and
anything
else.
My
work
like
motive
at
the
moment,
is
just
to
kind
of
race
title.
So
just
take
this
one
task
and
get
it
done.
If
something
does
seem
important,
please
do
let
me
know
at
all
reprioritize.
F
A
F
That
got
handled
yesterday,
but
I
didn't
write
about
it.
Basically,
there
was
an
issue
where
the
the
transport
between
the
the
client,
the
server,
didn't
function
properly
on
Windows.
That
has
since
been
resolved.
So
now
we
have
transport
on
that,
and
things
should
go
as
originally
expected
is.
F
B
Feel
free
also
to
write.
You
know
this
is
what
I'm
trying
to
do
blah
blah
I.
Think
it's
gonna
be
this
and
then
you
know
give
us
an
update
when
you
actually
have,
though
I
was
ridiculous,
because
you
know,
as
you
know,
it's
useful
also
to
have
the
the
track
of
what
folks
are
thinking
about
and
how
they
were
thinking
about
it,
and
then
that
helps
let's
not
do
the
same
thing
later.
So
if
it
was
the
wrong
path,
gotcha.
B
Updates
are
better
than
less
updates.
Related
I
have
a
ridiculous
car
I
have
a
Trans
Am
Firebird
Trans
Am,
more
Firebirds,
better
than
less
Firebird,
so
you
can
get
like
everything
in
the
world
with
like
Firebirds
on
it.
Just
absolutely
everything,
because
clearly
more
is
better
than
less
so.
You
can
think
about
that
with
github
updates,
where
fire
hurts
better
than
less
Firebirds
as
I
think
you
know,
but
I'm
sure
I,
don't
do
I
need
to
give
you
all
more
updates
as
well.
Let's
see,
I
have
anything
else
on
that.
One
questions
comments.
F
F
A
good
question
I'd
like
to
get
read
support
by
the
end
of
the
the
week.
So
what
that
would
mean
is
that
we
have
some
kind
of
program
up
that
basically
either
it'll
be
like
a
shell
or
it'll,
be
an
implementation
of
fuse
or
something
that
demonstrates
that
you
can
start
an
ipfs
node
and
read
from
it.
Using
this
thing
that
we've
made
that
sound,
sensible
and
I
should
have
that.
Probably
by
the
end
of
the
week.
D
B
E
B
B
J
J
G
C
C
B
B
B
J
B
K
Find
oh
there,
it
is
yeah,
so
you're,
working
on
figuring
out
how
we
can
improve
performance.
So
at
the
moment
it's
a
lot
of
just
kind
of
like.
Firstly,
I
was
exploring
and
figuring
out
how
it
works
today.
So
I
kind
of
wrote
up
an
issue
explaining
that.
That's
the
first
link
that
and
then
I
suggested
like
three
different
areas
where
we
can
look
at
improving
performance
and
I'm
kind
of
going
back
and
forth
mercy
with
Steven
has
been
very
responsive
and
has
also
put
in
some
comments
in
there
as
well.
K
So
I
guess
this
week.
Gonna
continue
kind
of
iterating
on
the
ideas
there
and
then
I'm
also
gonna
try
to
get
the
test
finances
in
a
state
where
I
feel
feel
confident
about
making
changes
and
measuring
what
they're
actually
having
the
right
impact.
So
there's
already
some
good
test
harnesses
there.
Some
benchmarking
stuff
I
just
want
to
kind
of
clean
it
up
a
little
bit
and
alter
it
for
a
specific
use.
Cases.
B
I'll
get
them
for
now.
Let's
see
okay,
anything
else,
I
don't
want
to
talk
about
or
any
questions
for
anybody.
Any
comments
on
your
like
breakfast
for
the
week.
Whatever
we
have
time
now,
cuz,
we
made
it
through
our
plans
for
the
week.
Thank.
E
B
C
C
I'm
Michael
burns.
Okay,
if
you
got
are
we
like?
No
we're
good,
you
should
be
fine
right,
I'm,
a
little
bit
worried
that
I
have
like
a
left
to
me,
just
like
my
constant
terror,
no
we're
good
something
that
might
be
interesting
for
this
group
and
might
overlap
with
what
Alan
is
benchmark
work
or
is
there
anywhere
where
we're
like
other
than
the
wonderful
bit
swap
benchmark
work,
that's
happening
where
we're
running
kind
of
regular
package
manager
of
focused
benchmarks
to
get
a
sense
of
their
performance
feedback
loop
on
PRS
we
submit.
D
So
I
believe
and
that
the
bits
that
were
like
functional
enough
to
be
worthwhile
being
benchmark
was
the
work
that
Dirk
was
doing
mostly
and
didn't
turn
into
like
a
repeatable
ongoing
benchmark.
Yet,
but
the
the
code
that
he
wrote
to
be
able
to
do,
those
pentatonix
could
potentially
be
turned
in
something
for
the
generation
of
particular
shapes
or
sizes
of
trees,
and
there's
also
like
our
index
of
the
different
repositories
that
are
rather
large.
D
How
long
does
it
take
that
out
of
terabyte,
mostly
because
the
infrastructure
suddenly
gets
a
little
bit
more
difficult
once
you
get
to
that
kind
of
size,
but
the
it
sound
like
right
now.
There's
we're
not
at
the
point
where
we
have
the
holistic
full
package,
manager'
path,
we're
working
on
individual
pieces
that
make
that
work.
So
there's
each
one
of
those
pieces,
these
benchmarks,
rather
than
the
full
path
of
setup
and
then
also
installation
via
the
thing
that
has
been
said
but
didn't
have
anyone
agree,
disagrees
so.
C
In
that
case,
like
all
of
the
existing
kind
of
IQ
fest
benchmarks
that
are
benchmarking,
ads
and
gets,
and
and
all
of
the
awesome
stuff
that
Alan
and
Jim
worked
on
kind
of
spinning
back
up,
that
kind
of
meets
the
use
case
of
hey
I
can
validate
that
ads
are
getting
faster.
Even
if
it's
not
like
package
manager
ads.
It's
like
ads
at
a
small
scale,
yeah.
D
So
that
that,
as
long
as
I
said,
there's
visibility
on
ads
of
a
good
number
of
files
gonna
go
and
which
Dirk
was
already
documented
in
his
issue
that
there's
I
forget
it
was
polynomial
dope
nodding.
D
B
C
It's
like
do.
Do
we
have
that
feedback
loop?
We
talked
about
being
really
useful
for
us
in
order
to
evaluate
that
we're
moving
in
the
right
direction,
yeah
and
have
we
set
it
up
such
that
we
don't
have
to
worry
about
it
anymore
and
it's
just
gonna
keep
making
our
lives
happy
and
give
us
that
information
I'm.
I
Okay,
so
I
think
this
will
become
more
visible
to
people,
but
after
the
meeting
later
today,
but
the
benchmarks
that
are
qfs
team
is
in
the
process
of
being
revived.
There
are
tests
there,
which
do
things
like
adding
a
small
unit,
1k
file,
a
full
megabyte
file,
a
64
megabyte
file
and
there's
tests
there
for
adding
100
1k
megabyte
files
to
ipfs,
and
so
that
information
can
be
graphed
over
time,
and
that
is
currently
happening.
I
Every
night,
Rajesh
ipfs
and
it
will
shortly
be-
we
soon
be
being
also
done
for
go
ipfs,
there's
also
other
tests
for
adding
files
and
which
will
be
relevant
from
package
managers
like
adding
files,
but
also
then
having
them
replicated
to
other
nodes.
And
honestly,
the
the
the
whole
idea
here
is
that
we've
got
this
benchmarks.
I
Framework
in
place
has
been
around
for
a
while
and
any
tests
that
you
want
to
see
in
the
four
that
are
relevant
for
package
managers
just
need
to
be
written
or
suggested
in
an
issue
on
the
repo,
and
then
we
can
get
that
done
and
get
that
added.
At
the
moment.
The
files
that
are
being
added
are
generated
and
they're
just
random
bytes
we
can
generate,
you
know
we
could
have
it
generate.
You
know
a
home
home,
no
more
files
or
bigger
files.
Whatever
you
need
the
possibility
is
there,
it
just
isn't
written
here
so.
D
In
the
benchmark,
school
and
I
want
to
say
a
couple
of
weeks
ago
and
I
think
I
reported
back
that
we
were
basically
in
a
place
where
we
would
wait
until
there
was
the
infrastructure
in
place
before
putting
any
effort
into
and
to
push
those
benchmarks,
because
there's
already
a
good
number
they're,
just
not
being
tested
against,
go
ipfs
yet
and
hopefully
we'll
be
able
to
see
the
same
kind
of
improvements.
They
add
performance
gains
with
the
existing
benchmarks,
without
necessarily
adding
any
more
benchmarks,
but
it'd
be
interesting
to
see
that
as
well.
K
Yeah
I
think
is
also
very
specific,
so
I'm
not
entirely
sure
if
it's
going
to
finish
the
benchmarking
framework.
So,
for
example,
one
thing
we
found
one
thing
I
drew
notice
was
that
when
you're
using
an
external
hard
drive,
we
don't
see
the
same
kind
of
improvement
again.
So
we
we
could
probably
use
a
lot
of
the
ipfs
benchmarks
and
then
we
might
need
to
write
a
few
of
our
own
that
are
extremely
specific
of
those
kind
of
use
cases.
H
So
we
can
actually
do
with
that.
Using
a
percent
marks
that
go
FS
has
an
equal,
a
store,
it
did
store
and
we
committed
insert
arbitrary
Clinton
C's
and
we
could
even
had
a
like
a
bad
one
thing.
Good
thing
so,
like
you
can
just
like,
spend
that
first
and
say:
okay,
like
I,
wanted
this
to
have
to
say
to
you
throughput
and
then
benchmark
that
and
see
where
the
bombs
are.
I
Think
the
I
think
it's
a
good
idea
to
see
these
two
not
exist
yet
I
like
that.
That's
fine,
because
we
are
just
getting
us
up
and
running
anything
that
you
think
might
not
fit
in
the
benchmarks
tool
then
like
at
least
open
an
issue,
and
let's
talk
about
it
rather
than
just
discount
it
straight
away,
because
there
may
well
be
ways
of
doing
this
stuff.
I
We
might
we
control
everything
that
gets
set
up
like
the
there
are
tests
there,
which
swap
out
different
different
muxes
in
J's,
ipfs
and,
and
you
enable
or
disable
like
SEC,
IO
and
stuff
like
that
for
tests.
So
we
can
swap
out
anything
if
we
need
a
data
store.
That
is
that
you
know
has
some
kind
of
delay
or
kind
of
bandwidth
limitation
to
it.
Then
we
could
create
one
for
the
purpose
of
the
test.
K
A
I'm
wondering
any
thoughts
on
how
the
the
benchmarking
framework
would
deal
with
like
I
want
to
run
our
sink
on
top
of
fuse
kind
of
things
like.
Is
that
reasonably
into
how
things
are
set
up,
or
is
that,
like
totally
on
a
different
path?.
I
H
Well,
ideally,
we
didn't
be
able
to
run
or
retrieve
a
script
in
just
like
what
that
VHS,
Runner
I
think
we
feel
that's
operation,
potentially
yeah
yeah,
and
then
you
just
be
able
to
drop
off
scripts
into
the
folder,
and
then
one
of
the
tests
would
certainly
just
like
farm
out
to
some
bastards.
We
could
even
have
the
JavaScript
tests
run
Charnas
and
then
it
won't
start,
especially
which
aren't
so
sweet
we
go
and
actually
give
the
few
stuff
Pollock
use.
Is
it
tends
to
like
lock
up
criminals
and
do
nasty
things
to
nature
system.
H
I
A
F
B
I
It
will
take
us
to
a
point
where
it,
the
tool
is
actually
useful
for
the
NGO
ipfs
and
then
then
we
can
maybe
start
writing
tests
that
are
specific
to
package
managers.
There
is
a
call
later
today,
I
think
it's
on
the
codes
calendar
for
for
testing,
where
I
think
Jim
is
going
to
demo
demo
the
graph
on
boards
and
the
work
he's
been
doing,
and
it
will
make
way
more
sense.
If
you
come
to
that
and
and
have
a
look
at
it
yeah,
you
can
I
say
yeah
if
you'd
like.
I
A
D
No
I
don't
believe
so
it's
mostly
a
case
of
working
out
what
like
how
close
things
should
be
to
the
current
state
of
IPS
versus
more
like
dream
code,
futuristic
like
what
would
this
look
like
if
I
get
asked
to
do
these
things
in
the
future
and
and
also
kind
of
looking
to
where
how
much
of
a
human,
the
human
parts
of
it?
I
I've
been
talking
to
Eric
a
little
bit.
Eric
comes
with
at
the
same
problem
from
completely
the
opposite
direction.
B
B
B
C
C
My
sons
are
doing
a
weird
thing
today
and
but
if
folks
I
just
took
notes
on
the
package
manager
meeting
issue,
it
has
lots
of
the
motifs
because
that's
easier
to
do
when
typing
into
this
than
just
typing
into
Google
Docs,
because
Google
Docs,
sometimes
locks
me
from
using
emojis
and
so
folks
want
to
take
a
look
at
those
or
if
you
want
to
copy
them
to
wherever
you
want
to
take
notes
from
in
nature,
that's
fun.
But
once
we
end
the
meeting
I'll
stick
the
recording
there
too.