►
From YouTube: Crimson/Seastar OSD Meeting 2019-01-01
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
So
I
sent
yesterday
a
pull
request
with
we've
changes.
The
monitor
play
to
the
monitor
client
thanks
for
interpolation,
I
also
spotted,
can
actually
be
another
problem
with
with
descent
machine
of
rocket
connection.
If
I
recall
correctly,
the
problem
is
in
messenger
f.
We
need
to
have
some
kind
of
message,
x-ray,
CrossRef
or
something
or
something
like
that
at
the
moment.
It's
it's
just
instance
of
steps.
A
A
Well,
maybe
we'll
need
to
extend
it.
Maybe
I
was
poking
a
bit.
I
was
in
the
pool
request.
I
was
probably
a
bit
with
messenge
F
F
Messenger
right
path.
Maybe
it's
another
issue
that
makes
me
headache.
Actually
there
is
no
language
way
to
ensure
there
will
be
no
modifications
on
to
the
object
remote.
Even
if
we,
if
we
would
like
to
rework
all
the
code
and
which
to
Const
message,
FF
somebody
can
define
a
mutable
member
of
the
object
and
cons
doesn't
prevent
that.
A
A
A
Breaks
I
have
any
SEC
fault
because
of
that
I'm,
not
our
how
the
memory
pool
unlock.
Oh,
the
new
and
delete
are
overridden
in
sister,
and
maybe
it
would
be
just
memory
leak
or
something
like
that.
I'm
just
taking
the
rule
that
the
memory
should
be
gala
should
be
delegated.
On
the
same
now
on
the
same
engine
that
allocated
it.
B
A
A
A
B
B
A
A
B
But
but
I
have
one
trouble
when
when
implementing
this
patch
is
because,
if
you
also
tested
this,
you
can
find
out.
There's
two
segments
fault.
Not
it
is
not,
it
is
not
happened,
but
it
is
just
random
happened.
So
if,
if
you
run
test
messenger
knew
that
the
new
unit
has
you
can
see,
sometimes
it
will
pass,
but
sometimes
it
won't.
So
that
must
be
erased
conditions
in
in
this
implementation
and
the
only
the
only
issue
I
thought
because
there's
a
segment
of
fold
and
then
there's
arrow,
stack
and
I
can
see.
B
C
B
A
B
B
B
B
A
A
B
A
A
C
A
A
You
might
be
also
interested
in
Adam
coupe
trickster
wall
clock.
It's,
it
uses
it
based
on
the
same
principle
as
marks
gdb
PMPs.
The
main
difference
is
speed.
Wall
clock
isn't
using,
doesn't
use
gdb
inside
its
native,
though
it's
more
useful
in
workloads
with
mark
is
more
useful
in
workloads
with
high
timing
requirement.
I
will
provide
you
with
a
link.
Let
me
fight.