►
From YouTube: WebPerfWG triage call - June 26th 2019
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).
C
B
D
A
A
A
B
You
know
I
on
my
own
set
of
frustration
with
respect
in
the
past
two
weeks,
because
of
because
of
that
I
think
at
this
point
we
either
had
a
choice
to
say
you
know
what
we're
going
to
keep
with
respect,
and,
yes,
we
have
to
basically
suffer
a
little
bit
for
the
until
tea
back
and
and
and-
and
you
know,
that's
at
the
price
rather
than
having
to
rewrite
all
of
our
specs
or
we
decide-
you
know,
respected
tourists
able
and
we
need
to.
We
need
to
change
too.
B
D
B
E
D
The
underlying
issue,
really
there,
the
problem
is
there's
some
set
of
documents
which
we
won't
manually
validate
every
time.
A
new
respect
version
rolls
out
and
there's
plot,
there's
probability
that
it
will
break
something
and
it
caught.
It's
a
cause
of
constant
frustration
for
us,
because
it's
not
something
that
we
do.
It's
like
Marcos
is
making
positive
improvements
in
respec,
but
sometimes
they
have
a
backwards,
incompatible
failures,
but.
A
I
think
this
time
around
it
wasn't
respect
changes,
but
something
will
to
Shepherd,
which
is
a
service
that
collects
cross
references
from
various
pecks,
and
there
were
some
failures
there
that
caused
these
failures.
Similar
failures
in
bike
shed,
but
those
are
compile-time
failures
which
enable
me
to
stay.
Ok,
I,
won't
use
the
cross-reference.
I
just
use
a
manual
link
where,
with
respect
it's
a
production,
failure
and
yeah,
but
we
could
freeze
the
HTML.
We
could
create
some
sort
of
a
build
step
for
respect
all.
D
A
D
Think
even
we
do
that
with
puppeteer
and
just
and
then
even
we
could
manually
update
it
like
Revit
on
our
own.
Well,
if
in
one
respect,
version
changes-
and
we
think
it
offers
significant
benefits
right
episode
of
like
then
we
control
what
bugs
we
introduce
or
not,
because
we
can
manually
validate
what
improves
up.
F
B
I'm
fine,
either
way
I
mean
you
guys
are
the
one
who
are
spending
most
time
on
those
special
occasions
or
not.
You
know
III
I'm
willing
to
try
to
help
fix
the
references,
the
references
issues
it
takes
a
little
bit
more
time
than
I
was
hoping,
but
if
you
guys,
you
know
whatever
you
guys
decide,
your
work
flow
ought
to
be
that.
That's
that's
fine
by
me.
E
It
looks
like
most
of
them
are
already
fixed,
right,
spec
or
whatever
fix
for
this
time.
What
we're
talking
about
is
more
like
have
to
make
it
more
future-proof
in
the
sense
and
whenever
a
respek
or
Shepherd
shepherd
fails
again,
we
don't
immediately
start
displaying
websites
that
have
lots
of
errors
and
incorrect
links,
so
it
has
to
just
miss
the
version
we
present
instead
of
relying
on
the
respect
infrastructure,
which
has
been
flaky
historically
I.
D
Would
actually
agree
cream
even
from
it's
not
just
that
it's
flaky,
it's
that
it
changes
all
right.
Syntax
changes
like
Marcos
is
landing
improvements
in
changes
in
the
syntax
and
because
of
that
some
past
documents
just
have
errors
and
it
needs
to
be
updated,
and
unfortunately
we
are
always
chasing
this.
On
the
oh
now
it's
broken.
Somebody
reported
that
you
know
there's
a
bunch
of
issues
being
fly
or
does
it
work
before
and
we'd
like
to
move
so
world
where
we
are
in
the
control
of
rubbing
those,
as
opposed
to
being
reactive
to
these.
B
So
yeah
we
can
set
up
some
Travis
job
or
something
that
basically,
you
know
actually
takes
a
automatic
generation
of
respec
when
the
committee's
done
on
the
main
branch
of
of
the
repo
like
we
do
that's
what
we
do
is
back
shed
all
the
time
we
set
the
map,
so
that
yeah
you
push,
you
push
a
source
file
and-
and
there
is
a
job
that
will
take
care
of
generating
the
index
I.
D
B
A
B
D
B
E
A
A
D
A
A
D
A
H
Like
just
I
think
I'm
looking.
F
A
E
A
E
E
There
is
this
weird
dependency
performance
timeline
on
user
timing
on
the
ideal
analyst
so
that
we
can
actually
create
entries
and,
unfortunately,
because
user
timing
recently
changed
than
the
ideal.
Harness
is
failing,
because
of
that,
oh
okay,
so
dumb,
but
I
will
call
it
it
not
really
a
performance
timeline
issue,
at
least
the
ones
that
Firefox
is
feeling
yeah
so
far
is
very,
do
seem
to
be
relevant,
but
the
phone
is
mad
and
performance
yeah.
Those
are
supposed
to
be
about
easier
time.
So
it's
not.
A
G
E
G
E
D
E
D
A
I
D
E
B
D
Okay,
my
second
question
is
looking
at
the
web
platform
tests
on
Safari,
just
squinting
I
think
a
bunch
of
them
are
due
to
the
fact
that
performance
observer
is
not
supported,
which
we
know
but,
for
example,
there's
case
sensitivity.
Tests
Alex
that
are
failing
right
now
or
get
entries
by
type
is
I,
don't
know
an
issue
for
you
guys.
F
A
A
C
B
A
A
G
A
D
D
A
A
D
B
B
A
A
B
A
A
D
A
Okay,
so
there
are
a
couple
of
issues
for
request
idle
callback,
one
of
them
I
think
we
just
need
to
land
the
PR,
which
is
clarified
time
remaining
I
I
reviewed
it
once
I
need
to
do
a
second
sweep
like
there
were
a
few
comments
that
were
addressed.
I
need
to
do
a
second
sweep,
and
then
we
can
probably
lend
this
and
then
there
is
the
question
of
deadlines.
Commission
then,
for
where
we
made
some
progress
during
the
the
hackathon
after
the
face-to-face,
where
we
asked
ahead
some
proposals
and
I.
A
A
A
A
I,
don't
like
critical
I
think
was
the
wording
from
the
spec
and
but
basically
define
what
tasks
are
critical
in
terms
of
those
test
sources
from
HTML
and
then
use
that
in
order
to
determine
what
to
do
when
a
deadline
is
approaching-
or
you
know
to
calculate
the
women
remaining,
then
then
lock
gotcha.
So
in
terms.
A
F
A
A
A
A
D
A
D
D
A
G
The
status
is
that
I
push
the
button
that
says
that
I
will
be
the
assignee
for
it.
Okay,
but
yes,
it's
on
my
radar
and
I
have
one
just
to
land.
You
know
some
stuff
to
land
today,
but
then
I'm
gonna
pivot
to
that
pretty
quickly.
Since
it's
a
since.