►
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
A
A
That's
produced
against
the
state
that
is
basically
accepted
as
the
blockchain
state,
so
that
regression
test
is
ongoing
and
we've
found
some.
Some
interesting
things
that
we
have
to
go
back
over
and
sort
out
so
there's.
That's
that's
been
a
lot
of
the
work
one
of
the
other.
So
that's
two
things.
One
is
the
the
regression
test
for
the
for
the
EVM.
The
second
is
performance
which
I
think
I'll
ask
Alan
to
just
tell
us
how
it's
going
in
a
minute
and
then
the
other
was
with
regard
to
fast
sink.
A
Adam
was
working
on
basically
a
way
to
to
speed
up
fasting,
try
and
figure
out
why
it
was
just
that
little
bit
noisy
in
terms
of
losing
its
connections
and
maybe
blacklisting
a
couple
of
connections
that
maybe
it
shouldn't
have
so
just
sort
of
studying
that
in
detail,
and
what
he's
come
up
with
is
some
is
some
throttling
that
seems
to
help
a
lot
on
ec2.
So
that's
gone
through
as
well.
A
That's
in
the
in
the
in
the
current
code
base
and
then
the
thing
that
we
haven't
tackled
is
the
yes
is
the
level
DB
Java
error
which
we
see
occasionally
so
that's
still
outstanding
and
then
beyond
that.
What
we're
looking
to
also
do
is
plan
for
the
next
release.
So
the
next
release
will
be
integrating
with
the
Daedalus
wallet
in
some
form.
So
that's
that's
the
the
planning
for
that
has
begun
and
we're
looking.
We
keeping
you
updated
on
I
guess
when
the
what
timeline
we
come
up
with
it
depends
on
a
few
things.
A
It
depends
on
what
we
have
much
of
a
chunk.
We
decide
to
bite
off
and
what
else
we
decide
to
put
in
so
that
is
I
guess,
basically
an
update,
the
other.
The
other
housekeeping
issues
that
I
would
mention
is
that
the
next
meet
the
next
broadcast
for
this
meeting
will
probably
be
the
fifth
of
September,
because
most
people
aren't
holidays,
including
myself,
so
I've
scheduled
the
next
broadcast
for
the
fifth
of
September,
and
then
we
should
know
a
lot
more
about
our
that
week.
A
So
Alan,
you
were
looking
at
some
of
the
measuring
some
of
the
performance
of
our
basically
the
the
main
cycle,
which
is
the
which
is
taking
reinf
lating.
The
blocks,
deserializing
them
from
the
database
and
executing
them
against
the
current
state
and
coming
up
with
a
new
state
and
rewriting
I
guess
whatever
needs
to
be
written,
the
receipts
and
so
on
to
the
to
the
DB
how's.
That.
B
Again,
yes,
actually,
there
are.
The
the
desk
was
based
on
an
issue
that
you
are
not
reported.
Related
to
sinking
was
slowing
down,
Iran
block
a
fifty
thousand,
which
is
where
transactions
start
to
appear
on
the
main
net.
So,
based
on
that,
we
started
grabbing
some
benchmarks
on
information
data
from
running
the
node,
with
a
lot
of
debug
and
Delta
a
capture
code
just
to
get
information
and
figure
out
what
what
was
going
on
effect.
B
You
were
right,
it
was
slowing
down
around
that
block
and
then
we
started
from
a
different
angle
and
we
discarded
on
the
network
and
all
the
noise
that
might
be
appear
from
there
and
we
started
benchmarking,
just
processing
blocks
from
one
database
and
trying
to
insert
them
in
another
database
which
which
is
basically
like,
we
processing
the
whole
blockchain
without
the
network
staff,
and
we
have
found
potentially
the
issue.
We
are
still
doing
the
final
test
with
our
fix.
B
Basically,
we
found
that
we
had
some
places
where
we
were
recovering
pooling
calories
from
a
transaction
that
wasn't
necessary,
and
that
was
a
not
helping
the
performance
of
the
client.
So
that
was
a
lot
of
time
instead
doing
some
crypto
math.
That
was
that
is
too
expensive
and
we
will
provide
on
a
pull
request.
B
Hopefully
today
and
there
you
will
see
the
analysis
in
terms
of
the
data
and
the
web
mark
that
we
ran
and
they
all
the
information
regarding
to
that,
the
different
approaches
we
have
to
fix
the
issue
and
why
we
picked
one
and
just
to
make
it
clear.
Where
was
the
performance
issue
and
how
we
decided
to
fix
it
in
order
to
not
to
introduce
a
lot
of
banks
in
other
places
of
the
call,
but
also
keep
it
clean
for
developers
that
might
be
looking
at
the
source
code
and
for
for
ourselves
too.
So
that's
it.
A
Okay,
yeah
so
I
mean
the
the
as
he
said
there
was
an
expensive
operation
being
called
unnecessarily
really.
So
that's
that's
probably
the
first
of
the
low-hanging
fruit
that
we
were
that
we
were
hoping
to
pick
I.
Guess
that's
what
you
do
with
low-hanging
fruit
when
we
went
back
and
started
to
look
at
at
some
of
the
performance.
So
that's
good
because
there
was
a
fairly
I
mean
assuming
it
all
works
out
and
we're
probably
we
don't
want
it.
A
B
I,
don't
want
to
say
any
number
you
know,
even
if
I
have
one
just
to
not
to
create
high
expectations,
but
it
seems
that,
yes,
we
will
have
a
really
a
performance
boost.
We
related
to
this
issue
and,
as
the
fix
is
going
to
be
like
surgery,
a
fix,
it
will
have
a
low
impact
on
the
quality
and
the
security
of
the
other
code
base
so
of
their
classes
and
stuff.
So
I
think
it's
a
really
really
nice
trade,
oh
cool.
A
Okay,
so
we're
kind
of
open
I
mean
there
are
I
haven't.
There
are
probably
a
few
other
things
that
we
found
during
our
own
beta
testing
that
that
still
need
to
be
sorted.
I
know
we
have
to
generate
more
databases,
we
found
a
problem
with
one
of
the
receipts
and
that's
gonna.
That's
resulted
in
it
in
a
peor.
A
That's
that's
gonna,
basically,
do
a
more
rigorous
check
for
that
and
beyond
that,
as
I
said,
I
think
the
only
the
the
only
other
thing
we
have
that's
sort
of
outstanding
functionally
is
the
is
the
level
DB
error
right
that
I
can
think
of
I.
Think
that's,
that's
that's
pretty
much
it,
but
having
said
that,
it
would
be
nice
to
I
mean
we'll
still.
A
C
Yes
sure
well,
I've
been
looking
at
several
minor
issues
that
we
had
a
fund
for
combat
that,
while
using
Morton
and
asking
for
the
network
version
with
a
JSON
RPC
method,
we
were
reporting
a
broad
number
here
issue
that
they
were
working
in
is
that,
as
you
were
well
said,
that
that
when,
when
we
were
doing
the
regression
test
on
our
bootstrap
database,
the
idea
was
comparing
the
story
taped.
With
the
visit
obtained
during
decryption
of
the
blogs.
C
We
found
that
there
was
a
difference
and
the
it
was
due
to
a
problem
that
we
were
not
a.
It
was
not
done
on
the
receipts
obtained
during
passing,
so
that's
already
been
fixed.
You
said
there's
a
little
question
for
that
and
apart
from
that
I'm
emotional,
about
several
minor
issues
that
suitable
reporting,
a
properly
errors
found
on
the
bootstrap
nodes
that
were
provider
kind,
or
also,
when
changing,
for
example,
from
fasting
to
revelers.
In
that
that's,
not
something
that
we
were
able
to
do.
C
A
C
What
well
well
this
this
changes,
it's
in
approval,
if
that
turbine,
so
it's
going
to
change,
but
what
was
changed
that,
yes,
we
could,
as
think.
As
you
said,
we
we
are
logging
that
this
change
is
you're,
not
able
to
do
so
as
a
cyst
on
in
our
client.
You
can,
which
has
private
and
with
she
knew
with
first
think,
although
it
was
the
claim
was
computed,
do
regular
sync:
okay,.
A
A
No,
probably
not
okay,
so
for
anybody
who's
wondering
where
everybody
else
is
today
is
a
bank
holiday
in
Poland.
So
all
of
our
polish
contingent
are
often
joining
themselves,
I
hope
and
just
to
recap,
then,
the
last
week
we've
spent
doing
really
two
things:
one
sorting
out
the
really
the
gold
information,
the
golden
information
that
we
found
out
from
her
from
our
testing
last
week
in
terms
of
observing
performance
and
observing
issues
by
just
running
it
over
and
over
again
in
different
scenarios.
A
The
only
issue
that
I
saw
was
related
to
too
verbose
logging
and
that's
controllable
through
log
back
and
then
the
third
thing
is
trying
to
come
up
with
a
schedule
and
a
roadmap
for
the
next
release
and
even
even
releases
beyond
that.
So
that's
ongoing
at
the
moment
and
we'll
know
more
about
that
and
after
the
break.
So
yes,
the
final
point,
I
guess,
is
that
the
the
next
meeting
the
next
one
of
these
will
be
in
two
weeks,
because
I'll
be
out
for
two
weeks
and
I.