►
From YouTube: Jitsi Live Stream
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
Okay,
the
live
stream
is
life,
and
let
me
just
have
a
look
at
the
situation
there,
because
last
week
it
was
registering
yep
there
we
go
it
so
cool
all
right,
hi.
Everyone
welcome
to
this
edition
of
the
GT
community
called
I'm,
going
to
start
with
a
quick
round
of
the
table
and
and
then
we
can
move
to
open
mic
as
usual.
The
parts
that
I
can
that
are
probably
the
most
important
right.
Now
we
have
this
crash
and
started
appearing
with
chrome
58.
A
So
it
sounds
like
this:
indie
might
be
the
issue,
we're
not
certain,
because
there
were
also
stack
traces
containing
access
to
stats
and
also
the
fact
that
there
have
actually
been
changes
in
stats
in
58.
So
we're
not
sure
exactly
how
these
relate
to
each
other,
but
it
might
simply
be
a
separate,
a
separate
bug.
The
one
that
we
were
able
to
reproduce
with
automated
testing
seems
to
disappear
as
we
rewrite
picture
IDs.
So
that's
that's
one
important
thing.
George.
Would
you
like
to
add
something.
B
Issues
the
other
I
worked
on
to
other
things
last
week,
so
one
was
currently
well
in
this
in
this
deployment
and
in
if
you
have
installed
this
specific
version
of
the
bridge
that
we're
running
now,
like
average
from
the
last
couple
of
weeks,
you
might
have
a
absolute
end
time.
Header
extension
broken,
so
you
are
advised
to
upgrade
your
bridges
and
this
this
affects
probing
and
adaptivity.
So
if
you
have
those
two
features
enabled,
then
you
should
definitely
have
upgrade
george.
B
You
know
what
ranger
versions
are
affected:
I-I
dunno,
not
by
hard,
though
so
we
can
we
can
discuss
about.
I
think
this
deployment
is
broken,
but
I
think
we
kind
of
know
that
already
so
yeah
that's
and
then
I
also
worked
on
on
on
what
turned
out
to
be
a
performance
issue
in
the
bridge.
B
So
now
I'm
working
on,
like
we
need
to
increase
the
UDP
buffer
sizes
and
the
UDP
socket
backlog,
sighs.
We
do
that
in
our
deployments.
We
don't
do
that
in
the
Debian
installer,
so
I'm
going
to
update
the
installer
to
fine
tune,
the
UDP
socket
configuration
so
that
performance
is
not
effective.
It's
not
affecting
the
bridge
in
in
a
negative
way.
So
yeah,
that's
that's
my
update.
Thank.
A
You
George
another
thing
that
we've
been
working
on
is
peer-to-peer
4121.
This
is
currently
employed
here,
we're
seeing
and
that
there
there
it
might
be
worth
looking
at
DIA
and
our
stats
a
little
bit,
but
we're
still
seeing
forty-two
percent
of
our
sessions
around
forty
and
fifty
percent
go
to
peer-to-peer.
A
I'm
not
sure
if
that's
in
this
made
sense
around
40
and
50
between
forty
and
fifty
percent
of
our
sessions
go
to
peer-to-peer
now.
This
includes
all
sessions,
including
the
moody
party
ones.
So,
for
example,
normally
this
call
should
generate
18
sessions.
Obviously
none
of
those
are
peer
to
peer
because
they
cannot
go
that
way,
but
then,
on
the
other
hand,
we
are
all
we
on
average
have
about
sixty
percent
of
the
causal
me
GT
as
one
to
one
calls
and
those
should
be
go
going
predominantly
through
peer-to-peer.
A
D
D
A
Speak
right
now,
if
you're
here
Karthik
usually
joins
these
coast,
but
oh
yeah,
there's
Karthik
Karthik.
A
A
Or
maybe
it's
a
microphone
issue?
Well,
let
me
update
what
I
known
Coast
ads.
They
push
the
new
version
that
takes
post
peer
connection
status
into
account.
So
now,
basically,
when
we
do
switch
to
a
peer
to
peer
connection,
we
pass
the
direct
connection,
so
the
aggregates
for
call
and
in
general,
for
the
service
that
cost
us
presents
should
be
coherent,
should
be
showing
the
right,
the
right,
quality
inn
and
other
stats
Karthik.
We
still
cannot
hear
you
that
remains
to
be
able
to
be
observed
a
little
bit
we'll
see
how
it
behaves.
A
Okay,
there
was
a
lot
of
work
for
release
on
on
mobile
last
week
and
Google.
Would
you
like
to
update
us
on
this
yeah.
F
So
I
released
your
voice
on
to
the
app
stores
and
I
realized
with
web
RTC
it
before,
because
we
didn't
get
enough
stability
from
way
back
to
see
57.
So
we
find
every
day's
toes
out.
He
only
calls
Bob
back
switching
of
the
color
front
back
and
vice
versa
and
I'm
sharing
the
URL
so
nicely
interested.
We
got
into
the
canvas
tents
of
whoever's,
testing
yup.
F
So,
as
was
previously
said,
we
have
those
efforts
to
our
comfort,
web
and
mobile,
so
go
starts
in
peer-to-peer
learning
that
direction.
We
should
land
in
soon
next
week.
Okay-
and
we
are
continuing
with
the
district-
will
stop
texting
by
Betsy
58
on
mobile
as
well,
so
that
he
wouldn't
join
today.
F
A
Thank
you
for
that.
By
the
way
the
audio
might
mode
on
on
mobile
is
something
that
I
find
very,
very
useful,
so
I'm
very
glad
that
this
is
not
part
of
the
release.
So,
let's
move
on
Damien
has
been
working
on
load
balancing
G
gases
in
G
Co
fo
Damien.
Would
you
like
to
talk
a
little
bit
about
that.
A
A
B
H
I
Can't
go
you're
here,
yella.
E
Okay,
so
not
not
much
from
me
directly,
but
I
worked
closely
with
Lenny
on
the
device,
selection
and
video
audio
preview.
E
So
we're
going
to
it's
already
on
me,
but
we
are
going
to
push
some
more
changes
there
and
fixes
all
so
that's
better
so
and
Lenny
have
worked
on
how
the
only
mode
for
web-
and
this
is
currently
this
code-
is
currently
in
review
and
will
land
soon
on
meat
as
well
other
than
that
I've
personally
been
working
on
the
iframe
API
and
trying
to
make
this
device
settings
and
the
dialog
available
in
film
strip
only
mode
and
they'll
continue
that
and
it
should
be
Oh
probably
available
in
two
weeks
deployed.
I
Like
implementing
a
scheme
to
to
be
able
to
add
some
different
types
of
transports,
let's
say
that
for
communications
with
external
applications
like,
for
example,
we
have
this
electron
wrapper.
So
we
are
communicating
through
post
messages
with
it
for
the
remote
control.
Let's
say
also,
we
have
the
iframe
API,
which
uses
post
messages
again.
I
So
now
I
have
a
PR
about
it,
which
adds
one
more
abstraction
layer
and
makes
it
possible
to
to
use
another
different
transports.
Let's
say
like
web
sockets
or
something
else
for
different
communications
with
external
applications,
because,
for
example,
for
depending
on
what
kind
of
wrapper
are
you
using,
you
may
not
be
able
to
use
post
messages,
so
he's
thinking
help
it's
not
much
scared,
but
it's
in
review.
Yet
it's
in
the
review
now,
maybe
soon
this
week
it's
gonna
be
merged.
I
I
A
We
have
recently
also
started
an
effort
to
try
and
see
to
try
and
work
on
cascading
bridges
little
bit
and
potentially
going
I
try
to
solve
multiple
issues.
With
this,
such
as
geolocation
or
big
conferences,
we,
it
is
already
possible
to
handle
geolocation
with
jitsi,
but
you
basically
end
up
that.
Basically
optimizes
where
you
host
your
conference,
it
doesn't
do
much
for
conferences
that
are
actually
multi
multi
region,
multi
continent,
so
Boris
leasing.
They
started
looking
at
these
things
Boris.
Would
you
like
to
talk
a
little
bit
about
this
I?
Don't.
J
We're
still
working
out
what
the
architecture
that
we
want
to
use
is
going
to
look
like,
but
the
idea
is
to
interconnect
more
than
one
bridge
together
and
distribute
the
conference
on
more
than
one
machine
so
that
they
can
scare
well
and
I
can
support.
As
you
said,
relocation
potentially
other
other
use
cases.
So
far.
We
don't
have
much
to
share.
A
Alright,
thanks
for
that,
Boris
brian
is
also
looking
a
little
bit
into
the
whether
or
not
there
are
any
performance
gains
that
we
can
make
on
in
various
places.
In
the
bridge.
Many
places
have
held
up
well,
which
is
I,
guess
good
news,
so
we're
now
diving
a
little
bit
more
into
the
detail.
Brian
Orser
took
an
effort
on
on.
A
Improving
the
hammer
so
friend,
would
you
like
to
talk
a
little
bit
about
all
these
things?
Sure
yeah.
K
So
the
hammer
has
been
updated
to
work
with
the
more
recent
version
of
smack
that
can
release
its
Mac,
which
got
rid
of
a
60-second
delay
when
he's
joining
calls.
So
that's
a
bit
better
and
then
it
also
supports
now
or
just
be
mocked
and
bundle.
So
those
are
some
improvements
on
the
hammer.
I
also
took
a
look
at
the
crash.
A
meal
I
was
a
little
late,
so
I
miss
some
of
the
context
here.
K
Discussion
I
might
spend
some
more
time
on
that
today
we
did
still
see
a
crash
even
on
George's
server
deployment,
so
there
might
be
some
more
that
we
need
to
look
into
their
one
of
the
problems
is
that
I
was
having
some
trouble
getting
stacktrace
symbols
and
chrome.
So
we
were
having
a
hard
time
distinguishing
one
crash
from
potentially
multiple
different
crashes,
but
am
I
trying
to
get
that
working
today.
K
So
we
can
find
distinguish
them
and
then
yeah,
and
then
it's
not
that
that
I'll
be
spending
some
more
time
on
ice
for
Jade
performance
and
I.
Have
some
good
test
results
from
last
week,
but
I'm
going
to
wait
a
little
bit
on
kind
of
sharing.
Those
because
of
f4
is
to
kind
of
put
another
set
of
eyes
on
the
test
bed
and
bet
that
a
little
bit.
So
we
can
make
sure
that
the
results
were
getting
are
accurate
and
our
kind
of
its
not
analogous
to
what
we're
doing
in
the
bridge.
A
Thanks
friend
right
would
the
point
that
I
was
making
is
that
we
did
eliminate
a
major
ease.
I
mean
we
do
see
greatly
reduced
number
of
crashes
with
picture
IDs
rewritten.
So
yes,
it's
probably
not
conclusive
to
talk
about
all
this
right
now
and
we
should
wait
for
to
play
with
it
a
little
bit
more,
but
it
did
seem
to
me
like
rewriting
the
picture
IDs
eliminates
at
least
one
problem
arm,
because
it
was
behaving
significantly
better.
That.
H
A
Yeah
we
could
I,
don't
know,
I,
don't
know
what
the
other
crashes
are
and
whether
or
not
they
are
the
same
problem,
just
we're
hitting
it
less
for
some
reason
or
or
if
it's
completely
different
problem,
but
I,
guess
that's
what
you
were
saying.
Do
we
have
to
see
right
George,
you
have
anything
to
add
there
yeah.
B
B
A
Hello,
okay,
I'm,
not
sure,
if
is
my
mic
working.
Yes,.
D
A
I'm
having
this
very
I'm
going
to
write
a
very
bad
read
about
this
headset,
so
Boris
also
worked
a
little
bit
on
that
something
I
should
have
mentioned
not
a
little
bit.
He
worked
on
on
adding
an
alternative
transport
to
data
channels
for
the
video
bridge.
So
now
you
can
also
do
them
over
web
sockets
and
the
reason
for
that
is
well,
it's
twofold.
First
of
all,
are
we
being
seen
a
lot
of
crashes
with
us?
Are
sctp
and
we
haven't
been
able
to
to
fix
them
in
the
library
itself
arm.
A
We
thought
we
might
kill
two
birds
with
one
stone
in
I'm,
just
realizing
now
how
cruel
that
sounds,
we're
not
killing
anything,
but
we
might
achieve
with
a
single
implementation.
A
Solve
two
different
problems
and
we
added
support
for
web
sockets
due
to
the
video
bridge
for
that
now,
none
of
this
is
implemented
in
meet
yet
data
channels
through
web
sockets
are
usable
on
the
bridge
and
Boris
ago,
when
you're
talking
a
bit
to
say
how
you
want
can
enable
them,
but
in
yeah
key
baskets,
do
is
currently
working
on
an
implementation
for
meat
for
edge
and
he's
going
to
also
add
the
support
for
daily
channels
there
Boris.
Would
you
like
to
to
tell
people
how
they
can
experiment
with
this.
J
Yeah,
actually
I
meant
to
write
some
documentation
for
this,
but
haven't
done
so
yet.
Basically,.
J
G
G
A
Okay,
so
the
bottom
line
is
that
Boris
is
going
to
write
documentation
on
how
to
enable
support
for
data
channels
in
web
sockets
and
he's
going
to
submit
that
as
part
of
the
bridge
and
then
finally,
I
wanted
to
mention
that
Sergio
has
been
working
on.
Adding
support
for
perk
in
in
a
in
a
fork
of
chrome
and
boris
has
been
working
on
adding
the
necessary
parts
in
GT
video
bridge.
A
So
this
is
a
conjoint
effort
between
between
us
and
Sergio
and
an
analogous
and
symphony,
and
last
week
I
said
you
had
his
first
successful
multi-party
end-to-end
encrypted
call.
There's
still
a
lot
of
work
left
there
still
a
lot
of
decisions
to
be
made,
we're
not
even
sure
whether
or
not
we're
going
to
end
up
with
something
that's
compatible
with
with
the
perk
specification
on
the
ITF,
because
that's
a
tough
one
to
you
in
terms
of
in
terms
of
well
politics
weenie.
So
we're
going
to
see
where
this
this
entire
thing
ends
up.
A
But
the
important
part
is
that
as
a
community
we're
getting
closer
to
a
solution
that
has
into
an
encryption
show.
So
that's
cool
and
I
think
this
summarizes
are
pretty
much
everything
that
we've
been
working
on
lately.
If
I
have
forgotten
something.
Please
speak
now
and
if
not,
let's
move
to
open
mic.
A
Why
I
think
I
gave
the
update
I
was
asking
about
if
you
wanted,
to
give
an
update
on
on
the
whole
peer-to-peer
versus
bridge
static.
A
M
F
E
F
J
J
L
Hey
guys,
can
you
hear
me
yes,
so
I
have
a
question
somewhat
like
I
want
to
create
like
a
scenario
like
suppose,
like
I
joined
this
conference,
I'm
receiving
everyone's
video
enroll
low
resolution
and
and
want
the
one
I'm
pinned
it's
full
resolution?
Is
it
possible
to
stop
stop
the
video
bridge
of
sending
the
video
for
a
person
and
only
to
start
if
they
tell
it
to
do
it's
like
simulcast,
but
not
send
the
video
by
default.
B
We
don't
have
a
data
Channel
message
that,
but
if
your
neighborhood
activity
supposed
to
do
that
automatically,
we
have
been
thinking
about
ways
to
provide
a
more
fine-grained
control
over
simulcasts
in
the
in
the
bridge.
But
we
haven't
decided
yet
how
to
proceed
on
how
to
implement
that.
So
that's
not
currently
possible,
but
definitely
yes,
meal.
A
A
L
A
A
So
the
only
or
demo
audio
only
mode
basically
gets
everyone
cuts,
all
video
going
to
you.
It
cuts
your
video
and
then
you
go
out
of
that.
Only
after
you
click
on
the
audio
only
button.
Currently,
this
is
only
available
on
mobile,
but
it
is
our
plan
to
provide
it
probably
the
month
of
May
over
web
as
well.
L
A
C
It's
great,
it's
very
informative
about
this
chrome
58
thing
we're
running
a
pretty
old
version
of
jitsi,
it's
about
9,
months
old
and
we're
in
the
process
of
updating
it.
We
haven't
seen
a
crash
like
I've,
experienced
several
times
during
this
meeting,
but
we
have
seen
an
uptick
in
impossible
exception
from
FMJ,
so
chrome
is
sending
a
lot
more
packets.
That
apparently
are
unrecognized.
I
was
just
wondering.
Are
those
two
related?
Do
you
know,
or
should
we
be
looking
for
this
crash,
maybe
even
even
on
our
old
version
of
jitsi,
it's
possible?
J
A
You
deniers:
if
we
had
fun
with
that
one
yep
are
any
questions,
I
mean
other
questions.
N
I'm
going
as
it
go
ahead,
oh
just
in
time
before
you
call
it
somewhere.
In
the
past,
she
deja
I
submitted
a
PR
to
the
jetty
video
bridge
project
and
I'm
already
building
on
top
of
that
for
my
local
project,
so
I'd
like
to
see
if
what
I
submit
it
has
any
chance
of
making
it
into
a
merge
state.
N
Basically,
what
I
did
is
I
attached,
the
native
library
said
ship
with
the
J
video
bridge
as
a
maven
artifact,
and
then
I
did
a
follow-up
where
I
updated
the
the
open
fire
plug
in,
but
that's
of
lesser
concern
to
me.
I
needed
a
way
to
get
to
the
need
to
fill
it.
Libraries
for
for
the
HD
video
bridge
Damien.
Have
you
seen
that.
A
No
I
they
ended
yet.
Would
you
be
able
to
have
a
look
at
it
this
week
and
see
if
we
have
a
problem
with
this
yep?
Thank
you.
Excellent
thanks.