►
Description
D
E
Sorry
so
I
submitted
a
doc
and
util
a
poor
quest
been
working
on
my
HTTP
bug,
trying
to
be
trying
to
help
some
of
the
packages
in
the
community
get
updated
a
v6
some
of
the
native
packages,
few
issues
here
and
there
then
just
been
working
on
this
commit
validate
or
tool.
They
can
hopefully
get
that
to
be
more
uniform.
Next
is
Jeremiah,
hey.
F
Hello,
buddy,
hey
Stephen,
we're
in
stand
up
so
I,
don't
actually
remember
for
the
other
CDC
mean
last
week,
but
I
did
a
bunch
of
stuff
for
v6,
including
a
breaking
changes.
Doc
from
that
this
week,
I've
worked
on
a
couple
of
sort
of
things
and
partially
cleaning
up
a
bit
of
a
mess
I
made
with
trying
to
get
unroofed
handles
to
have
a
check
on
them
that
you
could
actually
check
from
JavaScript,
but
I.
Don't
remember
what
else
significant
right
now,
I'm
working
towards
the
V,
6.1
I,
know.
G
I'm
we're
going
to
be
securely
last
week,
the
sitting
so
a
couple
of
regressions
from
that
one
of
which
we
already
dinner
to
discuss
today
fix
a
couple
of
bugs
in
you
till
inspects
have
to
do
it
up
proxy
objects
in
array,
length,
troubleshooting,
some
other
stuff,
so
there's
a
regression
in
buffer
indexes
and
last
index
of
that
apparently
traces.
All
those
active
before
working
on
error
code,
refactoring,
constant
refactoring
change,
loggers
are
curious
and
some
items
from
Vietnam
starter,
odd,
Josh,.
H
I
J
I
spent
some
time
investigating
issues
on
the
PPC
BBE
release.
Machine
turned
out
to
be
more
of
a
general
issue,
with
the
release
dub
configuration
now
working
to
validate
the
build
that
we
did
get
generated
from
the
release
job
for
a
nightly
for
the
big-endian.
Some
time
participating
in
a
sink
for
at
BP
review,
I
switched
over
the
benchmarking
builds
to
actually
run
off
the
new
v6
branch,
as
opposed
to
master.
Now
that
six
was
out,
I
wrote
and
submitted
the
stable
ABI
module
epe
along
with
Stefan
and.
K
L
A
M
Hey
guys,
let's
see,
the
main
updates
on
this
end
are
actually
Ali
is
visiting
the
v8
team
in
Munich
this
week,
which
is
I
believe.
Why
he's
not
on
the
call
right
now
continuing
to
figure
out
how
best
to
create
a
process
for
floating
back
patches
or
patches?
If
we
need
to
on
old
v8
branches-
and
you
know
basically
improve
our
waterfall
coverage
of
node
and
meanwhile,
the
the
dev
tools
team
is
continuing
work
on
on
getting
the
v8
inspector
working
directly
with
with
node
so
yeah.
That's
a
min
updates
on
Iron.
N
So
I
worked
on
a
couple
of
PRS
that
I
have
opened.
One
is
for
checking
in
ICU
into
master,
and
the
other
is
a
v8
break
it
earlier
to
throw
instead
of
crash
to
open
up
whole
to
get
the
International
Working
Group
born
again
try
to
get
it
we're
going
to
try
to
get
it
to
go
kind
of
a
monthly
cadence
and
next
is
server.
N
O
Right
now,
I'm
working
on
a
PR
so
that
I
rap
reports
because
right
now
we're
doing
a
lot
of
unverified
accesses
and
also
the
asa
crap
EP
review
and
comments
and
tuning
it.
I
haven't
responded
to
anybody
yet,
but
I
have
been
reading
it
and
I
do
plan
on
responding
as
soon
as
I
formulate
appropriate
responses.
Also,
while
writing
it
I
realized.
There
were
a
couple.
Laws
of
the
period
previously
mentioned
that
our
bug
fixes
that
need
to
be
fixed,
regardless
of
the
EP
itself,
so
also
been
working
on.
Those
next
is
rich
I.
B
Added
some
undocumented
for
now,
or
one
undocumented,
for
now
flag
to
tools.
Jslint
js2
automatically
fix
issues
that
are
automatically
fixable
I've
been
dogfooding
it
myself.
If
anybody
else
wants
to
give
it
a
shot,
please
do
and
I
also
opened
a
PR
in
the
last
24
hours
to
run
the
known
issues
through
CI
I
think
Jeremiah
discovered
that
that
it
was
not
being
run
in
CI
and
I,
so
got
that
happening.
B
P
So
we
merged
the
modules
EP
last
week
and
I
wasn't
at
the
CT
to
see
due
to
travel,
just
been
watching
the
response
to
that.
As
expected,
there's
kind
of
discussion
both
ways
on
it
started,
making
a
test
suite
to
ensure
whoever
tool
wise
creates
a
m
for
it
will
comply
with
it
and
looked
at
the
main
counter
proposal
that
popped
up
after
we
merged
the
EP.
Oh
it's
about
it
won't
discuss
more
later.
A
Okay,
thanks
everyone.
We
also
have
Bryant
else
in
here
from
microsoft,
/
tc39,
who
I
believe
will
be
joining
us
for
some
of
the
modules
discussion
we
won't
make
you
do
a
starter.
You
stand
up,
doesn't
make
a
whole
lot
of
sense.
Sorry,
let's
do
a
quick
review
of
last
meeting,
which
was
a
couple
of
weeks
ago.
We
added
new
collaborators
issue
number
six
to
eight
to
that
was
just
looking
at
the
list
to
sanity
check
the
list
issue
number
61
61
was
adding
a
dog
only
deprecation
for
you
to
log
the
issue.
A
N
That
is
qui
the
CIA's
clean
on
that
I
rebase
it
yesterday
I
guess
and
it
finished
the
build.
So
it's
clean.
Thank
you
to
see
lots
of
people.
Thank
you
for
a
job
and
others
for
fixing
the
cross-platform
issues,
so
we
actually
should
be
able
to
so
it's
ready
to
go
in
I.
Don't
know
if
there's
any
have
looks,
I'll
look
good
to
me
from
James
on
the
wire
gets
on
the
v8
one.
Any
other
comments
on
the
on
the
IP
one
together,
it's
a
reason
not
to
go
ahead
and
rode
it
I.
A
N
K
A
Okay,
if
your
number
60
63,
which
was
deprecating,
create
secure
pair
I,
think
that
went
ahead
planning
for
v6,
which
was
57
66.
We
did
the
legal
thing
which
I
think
we
skipped
introducing
staging
branch
for
stable
release
dreams.
We
actually
get
to
discussing
that
that
remember
the
outcome
of
that
and
the
end
of
es6
modules,
discussion
again,
which
I
don't
think
we
got
notifier
week
without
bradley.
P
A
G
Unfortunately,
none
of
our
CI
or
sitam
testing
or
and
pick
up
on
the
fact
that
this
book
few
other
things
just
to
go
ahead
and
revert
to
get
it
back
to
where
we
were
and
take
another
run.
It's
solving
the
problem
addressing
the
symlink
third
dependency
is
a
non
trivial
fix
at
this
point.
So
thank
reverting
for
the
best
option
coming
forward
and.
G
If
you
note
that
the
revert
commit
also
includes
two
additional
test
cases,
one
one
of
which
goes
into
the
known
issues,
actually
three
additional
at
this
one,
one
that
goes
into
the
known
issues
to
catch,
the
Sun
link,
appear
dependency
problem
and
the
others
to
add
new
tests.
That'll
catch,
the
regression.
E
Yeah,
I
agree.
I
think
we
should
regard
it
just
because
I
don't
think
that
we
can.
We
say
that
the
module
system
is
locked
and
we
actually
break
people
workflow
I,
don't
think
we
should
I
think
we
should
try
to
avoid
doing
that.
So
I
think
until
we
can
find
another
solution
revertant
it
would
be
a
good
thing.
J
G
O
Sure,
like
we
might
have
get
freaky
right.
Yes,
so
regardless
I
mean
we
are
breaking
things
so,
regardless
of
whether
it's
a
major
change
to
revert
inside
while
we're
reverting
a
major
breaking
change,
I,
don't
like
I,
don't
see
how
that's
an
issue,
especially
if
the
myth
itself
occurred
within
the
same
major
branch
that
in
which
is
being
referred.
Okay,.
F
P
G
A
couple
path
forward
that
are
being
invested,
dated
some
has
various
degrees
of
invasiveness.
Only
on
the
module
on
a
module
model,
be
probably
the
least
invasive,
would
be
to
put
the
new
behavior
behind
a
flag
command
lines
like
that
when
turned
on
it
switches
to
the
new
behavior,
but
by
default
it
uses
the
old
behavior.
The
other
changes
from
there
involve
a
variety
of
different
hacks
feed,
the
folks
that
have
been
on
both
sides
of
this
discussion.
I
kind
of
read
my
record.
G
The
discussion
last
week,
if
you've
been
give
these
folks
together,
hope
with
a
going
solution
and
they
started
put
their
heads
together
and
come
up
with
a
number
of
possible
coaches
that
are
being
investigated.
It's
still
too
early
right
now
to
see
if
to
have
a
real,
clear
path,
but
probably
the
putting
it
behind
a
flag.
F
L
A
G
A
A
P
P
P
P
Basically,
the
idea
is,
you
add,
a
module
dot
route
to
your
package.json
and
that
would
expand
your
package
from
say,
lodash
and
wherever
that
lives
on
the
file
system
to
lodash
/
dist
/
wherever
so
they
solved
it,
be
a
path.
Expansion
there's
been
various
ups
and
downs
on
social
media
about
it.
I
still
have
concerns
about
it,
but
it's
definitely
much
better
than
the
previous
ones.
I.
P
Don't
really
see
much
technical
advantage
to
it,
except
for
keeping
the
file
extension.
There
are
still
some
cons.
There
are
some
odd
remarks
in
the
proposal
itself
that
I
won't
go
into
detail,
but
I
was
just
trying
to
grade
it
on
technical
pros
and
cons.
There
are
some
claims
that
the
workflows
we're
concerned
about
are
not
important.
P
P
P
Mother,
pretty
much
all
the
use
cases
for
common
J's
modules
right
now,
all
the
basic
ways
to
create
a
common
J's
module
you'd
be
able
to
create
an
es
module.
There's
requires
a
package.json
and
it
isn't
always
feasible
with
existing
workflows
to
introduce
a
package.json.
We've
talked
before
about
how
some
asset
pipelines
like
rails
really
want
you
only
to
use
the
file
extension.
P
There
are
some
other
scripting,
tooling
issues
we've
talked
about
before,
and
just
some
existing
workflows
are
flat,
shared
directories
from
multiple
scripts
or
don't
have
package.json.
These
are
actually
aren't
the
biggest
use
case
for
node,
but
they're
well-known
use
cases
that
we've
seen
in
the
wild
there
is
some
collision
is
well
with
file
names
on
their
proposal
very
proposing
instead
of
index
JSA
module,
j/s
entry
point.
F
P
The
the
main
concern
is
things
like
firewalls
would
be
blocking
transmission.
However,
like
we've
looked
at
HTTP
rewrite
rules
can
jump
around
that
most
file
servers
and
things
won't
be
affected.
Browsers
are
unaffected.
So
it's
really
these
like
in
betweens.
That
are
the
concern
for
the
file
thing.
There's
also
a
lot
of
terms
being
passed
around
which
are
non
technical,
so
I've
avoided
them,
but
they're
things
like
defense,
beloved
I,
share.
A
I'd
like
us
to
dig
into
this
a
bit
more
because
I
know
because
I
keep
on
hearing
this
sentimentality
towards
the
file
extension.
What
are
what
are
the
objective
reasons
that
we
should
be
concerned
about
file
extensions?
That's
what
I'm
not
hearing
it
up
like
you
mentioned
some
stuff
about
fire
configure
and
conservative
figurations.
A
It's
not
like
we've
had
we
haven't
had
to
do
this
stuff
in
the
past,
but
and
can
you
help
us
try
and
empathize
with
some
of
these
concerns
without
file
extensions,
because
I
the
what
that's
the
main
difference
that
I'm
seeing
we're?
Just
not
we're
unable
to
be
sentimental
about
it
and
if
there's
objective
reasons
to
take
it
into
account,
then
let's
hear
them,
but.
P
There
are
concerns
mostly
around
people
directly
requiring
files
between
packages
so,
for
example,
some
people,
one
of
our
problems-
was
require
lodash,
/
core
and
some
people
are
requiring
directly
lodash
/
court
j/s.
Those
people
could
break
if
lodash
migrated
to
new
file.
Extensions
and
lodash
did
not
introduce
a
proxy
script
that
we
put
in
the
EPS,
so
there
there
is
some
concerns
there.
P
H
Expectation,
I
guess,
on
that
javascript
is
contained
in
a
jay
s
file
and
we've
kind
of
in
a
nice
place.
Now
that
all
JavaScript
goes
in
j
s
files
with
some
minor
exceptions.
This
would
be
a
rather
huge
exception
for
that.
So
it's
yet
another
concept
that
people
have
to
be
aware
of
when
they're
developing
in
this
ecosystem
for
typescript
it
kind
of
becomes
like
I
guess,
I
guess,
we'd
have
to
add
an
MTX
file
extension
and
a
dot
mts
X
file,
extension
and
them,
hopefully
that
combinatorial
thing
won't
explode
any
further.
A
A
You
do
understand
that
you
do
understand
that,
like
we,
we've
been
given
a
totally
new
file
loading
target
here,
like
we
were
expected
to
load
modules,
which
are
completely
different
to
scripts
within
the
one
format
here.
So
it's
not
like
I
mean
we
don't
want
to
have
to
injured,
introduce
a
new
file
extension,
but
the
modules
proposal
or
the
modules
spec
has
decided
to
introduce
totally
new
target.
So
I-
and
I
am
the
thing
I'm
struggling
with
here-
is
you
know
we
just
want
to
make
it
work
for
users
as
smoothly
as
possible?
H
H
H
F
In
mind,
though,
that
I
mean
other.
If
we
don't
go
to
the
file
extension
way,
then
people
are
going
to
have
to
track
extra
properties
detach
from
the
files
in
a
package.json
somewhere
else.
I
mean
the
only
way
it
can
be
in
the
same
files
as
if
we
have
like
a
pragma
at
the
top,
which
means
we
need
to
double
part
of
the
file
and
that's
already
been
rejected
for
performance
and
related
reasons.
F
H
A
C
A
P
H
So
so
that,
on
you
know,
many
of
these
scenarios
were
considered
as
part
of
the
initial
module
design,
so
I
wouldn't
I.
Don't
think
it's
correct
to
say
that
you
know
it
would
be
nice
if
this
kind
of
stuff
was
considered
like
we
have
been
thinking
about
this
kind
of
stuff
on
the
on
the
committee
side
for
for
a
number
of
years
now,
but
if
I
don't
think
spec
changes
are
off
the
table
here,
like
I.
Think
if
tc39
is,
you
know,
as
a
group
is
hesitant
to,
you
know,
have
a
separate
file
extension.
H
You
know,
for
whatever
reason,
I
think
that's.
You
know
a
good,
a
good
reason
to
ask.
Well,
can
we
can
we
make
some
spec
changes
to
to
address
those?
Those
pain
points,
we're
kind
of
saved
by
the
fact
that
modules
aren't
broadly
implemented
yet
in
browsers
or,
of
course,
in
node,
so
I
wouldn't
I,
wouldn't
say
that
anything
is
off
the
table
like.
If
you
have
ideas
for
how
the
spy
could
change
to
improve
matters,
then
we
should
bring
those
up,
and
you
know,
I'm,
not
promising
anything.
F
P
It's
the
forward
detection.
We
have
a
problem
with
which
is
in
part
due
to
separate
parsing
part
due
to
a
different
bootstrap.
We
need
to
include
instead
of
the
function
scope,
we
actually
are
going
to
be
using
an
import
statement
and,
in
part
due
to
runtime
semantics,
changing
but
not
causing
errors.
P
Basically,
as
long
as
implicit,
strict
mode
is
enforced,
we're
going
to
have
a
really
tough
time
without
forward
detection,
and
so
the
two
proposals
are
basically
just
different
ways
to
forward
detect
it.
We've
nobody's
really
argued
anything
about
basic
path.
Resolution
they've
not
argued
about
the
interop
wrappers
we
have
for
the
different
spec
objects
being
emulated.
P
It's
just
this
forward.
Detection
is
an
issue
because
it
sounds
like
from
everything
I've
read.
There
was
an
expectation
always
that
either
there
wouldn't
be
interrupt,
or
the
interrupt
would
be
done
in
a
way
that
was
configured
out-of-band,
so
not
in
the
source
file
or
related
metadata
to
the
source
file
itself,
but
that's
a
lot
of
hearsay.
A
lot
of
the
tc39
notes
are
a
little
bold.
A
So,
what's
the
path
forward
on
this
bradley,
my
reading
of
how
previous
discussions
on
this
is
that
how,
on
our
side,
there's
very
little
interest
in
investing
more
in
the
package.json
format
and
doing
double
file
loading
to
get
to
a
single
file,
and
so
this
there
was
a
strong
preference
for
the
file
extension
I.
Guess
you
know
if
compelling
reasons
were
presented,
then
people
would
change
their
minds.
But
what
do
you
see
is
the
path
forward
for
the
fervor
node
on
this
one?
Oh,
how
we
matter
we
make
progress,
I've.
P
P
There,
a
lot
of
things
about
people,
disliking
that
there
is
a
new
file
extension,
but
not
a
lot
of
yay
configurate
in
package.json,
also
so
its
convention
versus
configuration
at
this
point
and
I'm
just
kind
of
voting
on
convention
because
configuration
it's
a
little
error-prone
I.
Think
in
this
realm.
A
A
K
F
A
A
A
I'm
sorry,
Jeremiah,
okay,
there
anything
else
to
add
to
that
before
we
move
on
I
guess
the
status
is
that
we,
the
EP,
is
not
fine
is
not
does
not
mean
it's
absolutely
final.
Four,
four
node!
This
is
a
preferred
way
of
going
forward,
but
it
doesn't
mean
it
can't
change
if
there
are
compelling
reasons
to
do
so,
but
I
think
convincing
the
script
to
change
is
I.
G
Yeah
we
had
a
vision,
p
are
looking
to
add
a
new
constant
here
is
a
discussion
of
some
well,
you
know,
and
the
practice
constantly
aren't
even
documented,
very
well,
but
we
have
them.
We
mention
them
in
a
couple
places
in
the
docks,
but
don't
say
anything
about
the
stability
of
that
or
anything
else.
There'd
be
a
good
idea
to
just
go
ahead
and
get
him
documented,
but
it
took
a
first
run
at
it
guys
looking
at
how
the
constants
were
were
organized
and
so
I
really
doesn't
make
a
whole
lot
of
fence.
G
So,
let's
see
I
need
a
little
bit
better
structure
to
it.
What
speed
are
basically
does
is,
instead
of
having
a
required
constants
with
a
whole
bunch
of
things
just
kind
of
jammed
in
there.
It
kind
of
logically
separates
it
out
things
like
OS,
specific,
related
constants
for
like
error
messages
and
windows.
Pacific
error
messages
and
signals
are
dropped
off
of
os
constants
file
system.
Related
ones
are
alpha
at
SS,
constants
and
crypto.
Related
ones
are
crypto
Thompson
under
the
covers.
G
If
the
maid
of
my
level,
it's
just
a
slightly
different
structure
there
and
helping
them
all
on
one
object,
kind
of
structure
am
a
little
bit
aside
from
just
making
the
structure
a
little
more
logical.
It
actually
allows
us
to
make
a
couple
of
minor
improvements
in
the
code,
for
example
right
now
when
we
want
to
check
to
see
if
we
have
a
valid
signal
in
one
case
when
we're
in
child
process,
any
any
currently
defined
constant
could
actually
be
would
pass
the
check
as
a
valid
signal,
which
is
fun
and
then
another
place.
G
When
we
were
checking
signals,
they
were
a
little
hack
for
checking
to
make
sure
that
the
key
started
with
s
haiji
before
it
can
actually
do
to
look
up
this
kind
of
reorganizes
it.
So
the
signal
thing
you're
all
just
kind
of
grouped
together,
so
nothing
else
is
on
there.
It
makes
a
code
a
little
cleaner
and
it
actually
adds
proper
documentation
for
the
constants
that
still
in
the
process
of
going
to
learn
filling
out.
G
That's
kind
of
quite
a
bit
of
detail
to
kind
of
slog
through,
but
really
the
discussion
is
we're
going
to
go
ahead
and
do
this,
and
if
we
did,
but
this
actually
be
assembler
major,
because
we've
never
actually
documented
the
car
console,
and
this
change
doesn't
actually
stop
require
constant
from
continue.
The
worker
would
be
a
soft
deprecation
only
so
anybody
that's
currently
using
require
companies
would
not
see
any
change.
F
G
Yes,
it's
deprecating
in
the
sense
that
we're
providing
an
alternative,
but
since
we
have
nowhere
nowhere
to
indicate
that
it's
it's
difficult
ever
documented
in
the
first
place,
the
only
other
option
would
be
to
do
a
hard
deprecation
and
I
did
that
at
first
just
of
on
a
test
and
pretty
much
everything
blew
up
because
NPM
users
require
constant
and
obvious
went.
Users
require
constants
and
quite
a
few
others.
So
adding
a
warning
just
doesn't
make
any
sense.
Yeah.
G
F
G
B
A
G
A
G
Right,
so
it's
not
what
I've
done
it's
what
everyone
is
done.
The
changelog
grew
up
to
such
a
point
work.
It
was
no
longer
visible
or
usable
and
God.
You
had
to
drop
down
to
the
wall
mode
and
ordeal
to
see
it.
We
in
order
to
make
it
at
least
visible.
Last
week,
a
sort
of
release,
miles,
went
ahead
and
just
split
it
hop
into
an
archive
as
a
temporary
measure
about
what
this
does
is
actually
for
a
long-term
measure
reorganized
the
change
log
into
separate
files
by
version
with
an
index
between
them.
G
This
changes
to
an
index
here,
if
you
not,
this
version,
go
to
the
spa
closelink,
it
does
break
an
existing
links
because
mark
down
and
get
out
doesn't
give
us
a
way
of
redirecting.
Fortunately,
I
tried
a
number
of
different
ways
of
kind
of
hack
that
in
I'm
gonna
force.
So
at
this
point,
it's
pretty
much
manual.
There's
links
there
that
if
you
find
yourself
on
the
root
page,
you
can
click
on
it
to
follow
through.
We
hit
a
much
more
logical
structure
and
my
opinion.
G
G
G
E
A
So
the
just
one
note
on
the
archive
we've
put
the
ioj
of
stuff
in
there.
It
may
be
best
to
put
the
ioj
of
stuff
in
a
separate
file
simply
because
a
lot
of
the
stuff
between
V
0,
12
and
v4
easy
in
that
list.
So
if
you're
looking
for
changes,
then
you
can
find
a
lot
of
details
about
what
went
on
in
the
analogy.
Is
files.
A
Look
at
the
top
of
the
outlets.
Let's
move
on.
Let's
check
that
discussion
back
to
you.
If
there's
anyone
strongly
objecting,
please
do
so
and
you
have
otherwise
James
I
guess
you
can
roll
forward
yeah,
okay,
so
and
we'll
just
quickly
a
lot
of
time
for
a
public
QA
via
YouTube
IRC
or
the
github
issue
for
this
meeting,
please,
if
you
are
tuned
in
and
please
drop
a
note
there,
we
will
wait
a
minute
or
two
to
allow
people
to
catch
up
to
that
is.