►
From YouTube: C/C++ Compiler Options Best Practices (May 10, 2023)
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
B
A
B
A
A
Yeah,
for
some
reason,
I
thought
that
the
the
group
will
be
participating
in
the
open,
ssf
no
meeting
in
Canada.
Was
it
this
week,
I
thought
no.
B
So
I'm
not
sure
so
like
from
where
I
sit
so
I
I
sort
of
sit
more
like
in
this
part
of
the
organization
that
usually
doesn't
do
like
to
open
ssf.
But.
D
B
I
think
that
we,
at
least
at
Ericson,
have
other
people
who
are
more
active
in
actually
following
the
so
I'm
unfortunate
they're.
Not
the
right
person
to
not
the
right.
A
B
Yeah,
so
we
are,
we
are
in
the
same.
We
are
in
the
same
boat
as
well,
then,
but
we'll
try
to
we'll
try
to
manage
and
get
get
to
know
yeah,
but
I
I
did
notice
that
I
did
notice.
That
I
think
there
haven't
been
like
meeting
notes
for
a
lot
of
these.
So
if
there
is
some
physical
meeting
that
might
explain
why
they
don't
have
them,
but
I
think
as
far
as
I
concerned,
we
presume
with
regular
regular
programming.
So
that's.
A
B
I'm
going
to
put
the
meeting
notes,
meeting
notes
document
into
the
chat
again,
so
please
mark
your
presence
and
if
you
have
any
anything,
you
want
to
add
to
the
things
to
discuss.
Please
feel,
please
feel
free
to
go
ahead.
A
Yeah
see
that
says
his
mic
is
not
working
yet.
B
Yeah
I
think
that
it's
already
a
couple
of
minutes
past
and
okay,
that's
all
right.
I'll
I'll
have
the
chat
open.
So
if
you
want
to
chime
in
with
anything,
please
go
ahead
and
we'll
try
to
do
it
we'll
try
to
do
it.
This
way.
B
Okay,
so
I
think
if
we
are
just
going
to
be
the
couple
of
us
I
think
Qatar
can
try
to
make
some
make
some
notes.
Okay,
so
from
from
from
my
side,
I
actually
wanted
to
I
actually
wanted
to
talk
a
little
about
a
little
bit,
not
about
technical
things.
B
Today,
even
though
we
had
some
really
good
discussions,
I
think
in
the
past
past
couple
of
sessions,
but
I
wanted
to
a
little
bit
more
focus
on
sort
of
ways
of
working
and
having
like
this
meta
discussion
about
how
do
we
work
on
the
guide?
B
Also
considering
now
that
you
Gabriel
and
your
colleagues
are
working
on
this
on
the
Microsoft,
C
plus
plus
compiler
guidance,
so
I
think
it
would
be
a
good
good
sort
of
time
to
have
this
discussion
and
to
try
to
make
sure
that
we,
we
sort
of,
can
a
little
bit
like
harmonize
just
like
these
kind
of
things
that
we
know
have
been
like
an
issue
in
the
in
the
draft
and
as
well.
I
think
we
are
also
now
getting
more
and
more
contributions
from
outside
this
side.
B
A
little
little
group
as
I
think
it's
evidenced
by
like
these
open
issues
and
and
pull
requests
and
I
I
think
it
sort
of
would
be
would
be
good
to
could
a
little
bit
talk
about
this,
that
how
do
we
want
to
sort
of
organize
organize
our
ourselves.
A
D
B
I'll
pre-paste
repays
the
link
and
you
haven't-
you
haven't,
joined
this
little
group
from
before.
Have
you.
D
Basically,
on
the
repository
about
the
Microsoft
stock
missing
from
the
current
guide,
so
I'm
here
to
to
help
move
along
this
process
as
much
as
I
can.
B
I
think
that
would
be
like
a
great
addition,
great
addition
to
the
to
the
guideline
yeah,
but
but
I
think
that
if,
if
no
one
objects,
I
I
would
like
to
spend
a
little
bit
talking
about
this
sort
of
like
ways
of
working
ways
of
working
and
and
these
kind
of
this
kind
of
issues
today
and
I
I,
don't
know
like
maybe
sort
of
given
that
we
now
have
you
here
so
other
from
from,
like
the
point
of
view
of
like
this
Microsoft
guidance,
so
other
from
you,
your
point
of
view,
any
kind
of
things
that
you
see
that
we
could
sort
of
actually
spend
it
to
spend
time
on
today,
I
had
a
couple
of
things
on
the
agenda
agenda
too.
B
D
D
But
I
don't
want
to
take
too
much
time
from
me
from
me
from
your
agenda.
So
maybe
we
can
park
this
and
and
get
back
to
that
after
the
regular
agenda.
What
do
you
think.
B
Yeah
I
think
that
we
I
think
that
we
we
can
so
I
think
that
we
have
so
what
I,
what
I
usually
do
for
the
regular
agenda.
Is
that
sort
of
I
I
sort
of
try
to
prioritize?
Somehow,
and
usually,
we
have
like
a
lot
of
these
open
issues
to
discuss
anyway.
We
won't
get
through
because
we
won't
get
through
so
I
think
that
it's
okay
to
I
think
spend
time
on
I.
B
Think
these
collaborative
things
in
the
in
the
in
the
beginning
as
well,
so
I
think
that
so
I
I
think
that
maybe
the
like,
from
my
side,
sort
of
the
two
things
that
I
picked
up
from
our
past.
B
Our
past
meeting
notes
is
that
we
had
sort
of
one
suggestion
from
Randall
around
basically
having
like
some
kind
of
stable
versions
of
the
guide,
so
I
think
that
we
are
not
really
there
yet,
because
we
have
sort
of
still
sort
of
a
couple
of
different
documents.
We
have
been
merging.
B
We
have
sort
of
this
big
contributions
coming
in,
such
as
like
the
like,
the
the
Microsoft
Microsoft
guidance,
but
I
think
that
this
would
be
kind
of
like
a
good
I
think
to
I
think
discuss
already
now
that
we
do
we
do.
We
want
to
sort
of
aim
for
aim
for
some
kind
of
like
a
stable,
stable
release
or
or
do
we
look
more
of
this
as
kind
of
like
a
living
living
document
right,
so
I
think
that
it
will
sort
of
like
it's
part
of,
of
course,
it's
part
of
like
this.
B
Oh
this,
this
larger,
open,
ssf,
open,
ssf,
best
practices
repository.
So,
of
course,
there
are
a
lot
of
these
other
guide
being
developed
in
in
parallel
there.
But
I
think
that
we
could
sort
of
consider
this
that
if
we
want
to
actually
say
that
we
have
certain
stable
versions
but
I
think
then
it's
also
up
to
us
to
Define
what
those
milestones
we
want
to
meet
are,
and
maybe
also
think
about
that.
B
Can
we
actually
publicize
this
guide
at
some
of
these
some
of
these
Milestones,
because
I
think
that
I
think
that
sort
of
this
would
be
like
a
really
nice
piece
piece
of
guidance
to
actually
try
and
try
and
get
some
attention
to
attention
to
as
well.
A
I
see
merits
on
both
sides
like
having
a
a
living
document,
allows
us
to
accommodate
for
the
unsynchronized
release,
schedule
for
of
the
various
tool
set,
GCC,
Clank
and
nmsvc,
and
other
compilers,
you
know
so
having
living
document
allows
us
to
very
quickly
integrate
the
information
as
they
they
come
in.
A
On
the
other
hand,
you
make
a
very
good
point
that
if
we
had
like
Milestone-
let's
say
V1
or
V2,
we
can
very
much
Point
people
to
hey
look
at
this
version.
Maybe
something
in
between
is
that
we
always
strive
to.
A
B
Yeah
Randall
you
want
to
chime
in.
B
Yeah
yeah.
E
And
I
would
also
just
to
chime
in
and
to
agree
having
some
sort
of
release
would
help
us
to
also
I,
don't
know,
write
a
blog
post
or
something
obviously
about
it,
so
yeah.
It's
I
would
like
to
see
that
happening
too,
and.
F
I
don't
know,
maybe
you
should
actually
can
answer
this,
but
every
so
often
I
see
certain
flags
do
get
changed
around
or
merged
and
whatnot.
So.
C
F
C
C
That
that
would
probably
be
useful,
where
you
know
someone
could
very
conveniently
pointed
a
specific
version
saying
that
here
you're
using
gcc14,
so
these
flags
would
be
perfect
and
here's
the
you
know
the
the
tag
in
the
GitHub
wrapper,
which
right,
which
kind
of
gives
you
the
entire
details
of
all
of
the
flags
that
you
can
use.
B
It
that's
actually
like
an
interesting
point
that
I
hadn't
considered
so
you,
you
sort
of
want
to
tie
like
the
releases
of
the
guide,
with
specific,
specific
compiler
like
the
compiler
release
schedule,
but
I
I
think
maybe
like
one
point
that
I
would
maybe
sort
of
want
to
a
little
bit.
Discuss
more.
Is
that
like?
Currently,
we
are
saying
that
we
sort
of
also
want
to
com
want
to
cover
like
older
compiler
versions,
right
now,
right
and
so
just
to
clarify.
B
So
you
will
sort
of
like
suggesting
that
we
should
have
like
versioned
versioned
version
of
the
guide
for
different
compiler
versions.
Or
is
it
still
that
we
sort
of
have
one
guide
which
sort
of
have
has
everything,
but
then
you
can
sort
of
pinpoint
like
which
release?
Has
the
latest
information
about
the
company
yeah
yeah,
pretty.
C
Much
I
mean
I
did
not
suggest
having
like
a
different
version
for
different
compilers,
but
checkpoint
diet.
The
single
document
at
various
points
where
we
say
this
is
this-
is
true
up
to
GCC
14.
Something
else
might
be
true
at
Clank
18
somewhere
in
future,
so
we
don't
need
to
name
the
tags
based
on
the
compiler
versions.
We
could
probably
do
it
like
at
a
six
month,
Cadence
or
something
like
that
which,
which
kind
of
loosely
coincides
with
the
company
release,
dates
but
yeah.
C
Something
like
that
where,
where
we
have
snapshots
that
we
can
point
to
for
for
various
distributions
or.
F
B
Yeah
that
sort
of
fossil,
like
jumps
a
little
bit
together
with
so
we
had
like
this
open
issue
from
from
David
wanting
us
to
look
at
specifically
like
the
new
new
features
in
new
features
in
GCC.
I.
Think
that,
like
with
the
with
the
latest
GCSE
release
and
I,
think
that
having
this
sort
of
workflow
I
think
also
makes
sense
right.
So
like
eventually,
we
are
like
I.
B
Think
that
now
we
have
like
sort
of
like
a
big
backlog
of
things
to
look
through,
but
I
think
that
once
we
sort
of
be
able
to
work
work
through
that,
I
think
that's
sort
of
like
the
main
maintenance
of
the
guide.
I
think
kept
as
quite
naturally
like
this,
that
we
follow
what
are
the
new
features,
and
then
we
do
an
update
and
yeah.
B
Maybe
it
sort
of
naturally
comes
that
the
schedule
sort
of
follows
these
so
yeah
that
that
could
be
like
that
could
be
like
a
possible
possible
sort
of
driving
force
for,
like
our
internal
internal
Milestones,
yeah.
B
This
ways
can
spend
a
little
bit
more
timing
going
through
the
weave.
We
still
have
some
material
in
like
this
initial
draft
and
then,
of
course,
like
this,
this
Microsoft
work
as
well
and
potentially
the
the
Intel
work,
but
maybe
maybe
after
sort
of
these
big
packages
are
are
merged.
We
could
consider
consider
a
kind
of
like
a
more
regular,
regular
schedule.
A
Oh
yeah
I
was
going
to
ask
for
reminds
to
on
how
what
are
your
thoughts
about
how
to
drive
it?
Would
it
be,
driven
by,
let's
say,
feature
set,
meaning
that
a
a
fixed
set
of
topics
that
need
to
be
covered
and
when
we're
done
with
that
and
we
release
or
is
it
more
date
driven
which
is
I
know
by
September
30th?
A
Whatever
is
in
the
document
get
released?
You
know
any
thought
on
that.
B
A
I
mean
the
an
obvious
counts
of
having
future
said
is
that
it
might
take
a
while
before
we
get
it
completed
and
the
cost
of
having
deep
driven
is
we
most
likely
have
a
bunch
of
sections,
CBD
and
and
I,
don't
know
to
a
degree
that
reflects
on
the
quality
of
the
overall
documents
and
yourself.
But
you
know
there
are
also
Pros
on
both
sides.
I
mean
Pro,
for
which
you
said
is
well.
When
we
release
something
we
know
it
is
instantly
consistent
and
well
contained
and
and
for
a
a
dates
driven
release.
B
Yeah
so,
like
my,
my
sort
of
initial
thought
is
that
I
think
that
we
are
Maybe.
You
know
I
think
that
initially
it's
going
to
be
like
this
feature
like
this
feature
based
right,
so
we
still
have
we.
We
still
have
sort
of
like
the
a
couple
of
big
this,
this
sort
of
this
big
Sub
sub
categories,
some
categories
to
work
on
as
well
as
I
think
this
I
think
interesting.
B
You
know
interesting
discussions
around
how
we
sort
of
present
the
information
like
dividing,
dividing
into
different
into
different
tables
and
I
think
it
will
probably
take
like
a
little
while
before
we
before
we
before
we
get
there,
but
but
I
I
I
I
also
wanted
to
sort
of
start
this
discussion
exactly
for
this
reason
that
ER
pointed
out
right
so
I
think
that
they're
all
sort
of
like
opportunities
to
publicize
publicize
to
guide
I
think
we
should
I
think
we
should
actually
do
that
so
just
to
to
make
sure
that
the
the
the
work
we
put
into
it
is
actually
also
is
also
noticed.
B
That
sounds
good
to
me,
yeah,
but
but
I
think
that
maybe
maybe
sort
of
maybe
there
is
sort
of
like
this
obvious
obvious
packages
that
we
want
to
finish.
First
right
and
I
I
think
that
maybe
we
can
then
bring
this
bring
like
the
question
of
doing
like
first
release
up
when
we
are
sort
of
happy
with
with
sort
of
like
this
big.
A
B
B
You
know
complete,
but
I
think
that,
as
long
as
we
can
get
it
to
a
point
where
it's
consistent,
yeah,
that
I
think
would
be
I
think
a
sort
of
like
a
criteria
we
could
use
to
to
decide
to
go
forward
with
that
with
a
stable
version.
Yeah.
F
B
But
it
seems
that
this
sort
of
somehow
tying
it
together
with
the
compiler
releases
seems
to
have
some
support,
at
least
so
I
think
that
would
that
can
be
something
that
we
can.
We
we
can
sort
of
aim
towards
and
for
like
a
regular,
regular
schedule,
Okay.
So.
A
Given
the
amount,
the
value
number
of
compilers
Run,
is
it
up
to
us
or
is
it
up
to
the
contributors
from
those
compilers
to
kind
of
at
O'reilly,
blog
post
and
says
hey
for
this
version
for
this
version,
recompiler
Look
at
that
version
of
the
document.
B
B
I
I
think
that's
like
in
practice.
That's
an
open
question,
open
question
right,
so
I
think
that
we
don't
have.
We
don't
currently
have
representatives
of
every
compiler
right,
so
I
think
that
I
think
that,
from
my
side,
I
would
sort
of
leave
this
up
to
those
who
those
of
us
who
actually
represent
the
compiler,
whether
they
want
to
you,
know
publicize
this
guide
on
on
their
own.
But
I.
B
Don't
think
that
that
prevents
us
from
publicizing
this
through
open
ssf
or
you
do
you
have
a
I
think
a
gear
or
suggested
this
blog
post.
So
did
you
have
like
an
idea
for
for
where
to
where
to
publish
where
to
publish
yeah.
E
The
blog
post
is
probably
an
uncontroversial
thing,
but
at
the
end
of
the
day
we
want.
Ideally
we
we
want
developers
to
pick
this
up
right.
So
you
need
to
create
a
little
bit
of
awareness
for
the
document,
so
otherwise
we
would
just
do
it
for
ourselves.
The
compiler
folks,
of
course,
yeah
and
I,
would
try
to
publish
really
something
in
the
blog
and
then
make
both
the
developer
and
the
compiler
side.
E
Aware
of
this,
maybe
the
compiler
folks
first
to
give
them
a
heads
up,
saying:
hey:
we've:
we've
done
this
feel
free
to
take
a
look,
but
then
for
the
adoption
on
the
developer
side.
I
would
I,
don't
know,
yeah
really
create
a
blog
post
and
mention
it
elsewhere.
We
also
should
mention
that
in
the
maybe
we
should
bring
it
back
to
the
large
attack,
so
crop
obvious.
The
larger
group
group
obviously
is
aware
of
this.
E
Maybe
we
wanna
report
this
out
to
the
tech
at
some
point
in
time
as
part
of
what
the
what
we've
done-
yeah
I
think
that's
just
valuable
in
general,
pointing
people
to
this.
C
B
C
When
I,
when
I
made
the
suggestion
of
playing
it
with
compiler
releases,
I
was
I
was
thinking
of
basically
Cadence
like
like,
for
example,
GCC
has
a
release
roughly
every
April,
May
right
and
clang's
releases
are
I
think
at
least
for
the
last
five
years.
It's
been
March
April
and
September
October.
C
So
if
we
have
releases
that
are
like
document
releases
that
are
say,
December
June,
like
one
release
in
December
One
release
in
June,
then
it
gives
us
enough
time
to
kind
of
collect
new
features
and
whatever
compiler,
that
we
want
to
Target
and
update
that,
and
you
know
checkpoint,
that
as
like
this
new
thing
that
has
happened
in
the
in
the
compiler
Flags
space.
C
So
the
June
release
will
probably
be
busier
because
you'll
have
GCC
as
well
as
clang
to
cover,
but
then
the
December
one
will
probably
be
a
little
easy.
B
Yeah
I
I,
actually
I,
actually
like
that
idea.
I
guess
that
we
it's
sort
of
like
a
thing
where
we
have
to
maybe
get
a
little
bit
of
experience,
sort
of
how
how
this
work
will
go
in
practice
and
we
can
sort
of
then
so
the
schedule,
depending
on
how
much
time
we
feel
we
need,
after
each
after
each
compiler
compiler
release
but
but
I,
but
but
I
think
that
generally
there
isn't
like
there
isn't
like
a
huge
amount
of
sort
of
new
features
coming
coming
in.
B
So
I
think
that
so
the
I
think
that's
sort
of
like
the
most
important
thing
would
be
would
be
to
just
make
sure
that
we
are
sort
of
aware
of
what's
happening
and
I
think
I'm
sure
there
is
I'm
sure
there
is
volunteers
to
actually
then
write
the
material
when
the
new
features
come
in
and
help
in
helping
editing,
yeah.
D
Okay,
yeah.
D
So
I
think
that
a
right
there
is
a
the
life
cycle
that
is
sandbox,
incubation
and
graduation,
so
I
I
think
that
fits
best
with
finishing
the
the
release
and
then
going
back
to
the
working
group
and
then
the
developer
adoption
and
the
work
group
adoption
should
support
that
for
long
term
right.
So
maybe
this
group
is
not
the
long-term
support
of
of
these
guides,
but
the
working
group
is.
E
Technically
I
would
say
it's
not
a
project.
It's
it's!
A
working
group,
really
a
project.
There's
been
quite
some
discussion
some
time
ago
in
the
tag
trying
to
figure
out
these,
these
life
cycle
matters
and
the
project
would
really
more
fit
to
something
like
yeah
a
code
producing
activity
there
are
six
stories,
may
be
a
bad
example,
but
but
scorecards
or
something
like
that.
E
That's
that's
a
project,
whereas
many
of
the
the
documents
being
created
and
maintained
in
the
open,
ssf
are
more
like
working
groups
so
and-
and
the
working
group
has
a
much
simpler
life
cycle
in
the
sense
that
it
is,
it
is
active
and.
E
Yeah
and
then
it
may
or
may
not
shut
down
shut
down
at
some
point
in
time,
whereas
a
project
obviously
is
more
like
and
that's
also
inspired
by
the
cncf
life
cycle.
The
the
graduation
steps
are
defined
also
in
terms
of
adoption.
How
many
different
maintainers
do
you
have,
and
so
it
just
feels
to
be
yeah.
B
I
think
that
we've
characterized
ourselves
as
like
a
sub
initiative,
underneath
this
best
best
practices
for
working
group
but
I,
think
that
there's
also
like
sort
of
two
ways
to
look
at
it.
B
So
you
can
look
at
this
sort
of
like
likes
life
cycle
in
in,
in
terms
of
like
this
overall
open
ssf
open
if
it's
a
structure
but
I
think
that
that
doesn't
prevent
us
from
having
you
know
our
own
life
cycle
for,
like
the
actual
technical,
technical
work
and
at
least
that
I
have
a
at
least
I,
have
a
hope
that
that
we
can.
You
know,
make
sure
that
this
document
doesn't
become
obsolete
right,
because
that's
sort
of
like
one
of
these
challenges
right
so
I
like
now.
B
We
have
like
a
lot
of
I,
think
and
I'm
very
happy
to
say
that
we
have.
We
ever
had
a
had
a
lot
of
sort
of
momentum
with
this
right
and
very
much
exciting
contributions
coming
in.
But
I
also
want
to
make
sure
that
we
can
ensure
some
kind
of
longevity
to
this
as
well.
So.
B
Okay,
I
think
that
was
a
that
was
a
very
good,
very
good
discussion.
Are
there
any?
Are
there
any
other
comments
on
this
particular
on
the
on
the
on
on
the
versioning?
Anyone
wants
to
point
out
still,
or
can
we
move
on
a
little
bit.
B
Okay,
if,
if
not
so
you
so
you
had
some
something
you
wanted
to
discuss
in
terms
of
like
this
project,
like
the
sort
of
like
the
project
abhishe
right,
maybe
now
would
be
a
good
time
to
get
into
that
to
ensure
we
have
time
of
project
on
considerations.
D
D
We
need
to
prepare
more
the
content
that
we
currently
have
before.
It
is
ready
for
contribution
and
I
guess.
One
of
the
question
is.
D
D
So
maybe,
as
Gabriel
mentioned
feature
like
feature
set
different
features
set
which
are
released
on
a
on
some
Cadence
or
I?
Don't
know
what
what
do
you
think
Gabriel?
Do
you
think
that
there
is
anything
that
we
can
do
to
make
this.
A
Process
yeah
so
I
guess
yeah
the
you
know
2.0
the
I
guess
the
problem
is
so
the
issue
is
trying
to
figure
out.
Is
the
current
document
has
certain
structure
I
think
two
weeks
ago
we
agreed
to
have
sections
and
subsections
and
and
tables
and
and
linking
and
and
so
forth.
So
the
question
is
for
the
initial
contribution
from
Microsoft.
Are
you
expecting
that
all
those
sections
be
covered,
or
would
it
be
okay
to
incrementally
submit
or
request
that
go
section
by
sections.
B
B
Like
from
my
side,
I
would
think
you
should
proceed
in
a
way
that
makes
sense
to
how
you
organize
the
work
among
among
you,
I
I,
don't
see
like
personally
I,
don't
see
a
problem
with
going
iteratively
and
that
might
even
be
the
better
like
a
better
option,
because
I
think
that
then
we
can
I
think
probably
be
more
effective.
And
discussing
these
changes
in
these
meetings
right
so
generally,
we
will
like
run
out
of
time
in
these
meetings
right
then.
B
Hopefully,
we
can
take
part
of
the
discussion
offline,
but
but
of
course,
this
I
think
that
we
we
like
actually
make
some
really
good
progress
as
well
in
in
we're
as
well
in
these
meetings.
But
I
would
maybe
Sort
of
hope
that
you
know
we
have
like
this
certain,
like
the
certain
structure
we
have
is
basically
that
we
have
like
for
each
option.
We
have
sort
of
this,
this
sort
of
like
a
high
level,
high
level
summary
that
gets
reused
in
the
tables,
and
then
we
have
sort
of
like
a
like
the
synopsis.
B
Then
we
have
the
performance
considerations
and
then
we
have
this
when
not
to
use
or
Orca
beat
section
and
I
I
think
it
would
be
good
to
stick
to
that
kind
of
format.
I
I
think
that
we
have
we
sort
of.
We
are
open
to
like
changing
these
wordings,
I
think
so.
I
think
we've
already
had
like
a
little
bit
of
back
and
forth
this
that
maybe
it
should
be
called
cavits
and
not
been
not
when
to
use.
B
I
think
that
this
is
maybe
a
discussion
we
should
have,
but
I
would
prefer
that
to
have
like,
with
a
sort
of
like
a
larger,
larger,
larger
group
group
present.
So
we
can.
We
can
fine-tune
the
sort
of
the
wording,
but
I
think
this
overall
structure
would
be
good
to
aim
for,
because
these
are
the
sort
of
the
usual
concerns
for
these
options
right.
So
with
people
like
generally,
when
they
want
to
use
them,
they
will
be
first
considered
about
performance,
and
you
know
second,
when
I
shouldn't
I
shouldn't
use
this.
B
So,
in
that
sense,
I
would
think
it
would
be
good
to
address
address
them
in
in
in
in
in
in
that
sense,
if
it's,
if
it's
possible,
yeah
and.
A
B
And
and
I
I
think
that
the
conclusion
from
the
last
meeting
was
that
it
makes
sense
to
also
have
a
separate
table
for
this
right,
so
it's
sort
of
in
that
sense
also
a
little
bit
independent
from
the
overall
structure
and
I
think
that
we
are
also
going
to
going
to
rework
part
of
that
overall
structure.
Now,
based
on
discussion,
we
had
I
think
that's
on
my
to-do
list.
B
I,
unfortunately,
haven't
been
able
to
provide
a
concrete
proposal
yet,
but
I
hope
to
be
able
to
be
able
to
do
that
too,
from
the
for
the
overall
tables,
but
I
I
also
think
that
it
like
we
have
it
sort
of.
We
anyway,
have
like
this
clear
separation
right,
so
I
think
that
there's
nothing
nothing
that
prevents
pushing
a
new
section
separately
without
linking
it
to
any
of
the
tables
yet.
But
then
the
linking
also
like
in
a
separate
in
a
in
a
separate
in
a
separate
change.
A
Thanks,
that's
good
thanks
avici
for
bringing
that
up.
B
Yeah,
do
you
have
anything
anything
else
from
from
your
side.
D
No,
we
hope
to
to
have
more
updates
for
the
next,
for.
B
B
Yeah
then
I
wanted
to
also
bring
up
this
discussion
from
that.
We
started
during
a
past
meeting
on
like
the
style
and
implementation
of
citations,
because
I
think
this
is
also
relevant
for
these
new
contributions,
because,
presumably
you
are
also
going
to
cross-reference
a
lot
of
the
existing
Microsoft
documentation,
so
I
yeah,
so
I
think
that
now,
of
course,
we
don't
have
I,
don't
think
that
we
don't
have
necessarily
like
the
proponents
from
both
of
these
camps
since
crop
and
David,
unfortunately
couldn't
attend,
but
I
I
would
anyway.
B
I
would
still
want
to
sort
of
bring
this
up,
because
I
think
that
we
sort
of
have
like
this
sort
of
two
suggestions
and
if
I,
if
I,
try
and
try
and
and
sort
of
cover,
both
sides,
so
I
think
that
on
one
hand
we
had
this
suggestion
just
just
to
use
like
inline
linking
wherever
wherever
it
makes
sense.
And
then
you
know.
B
On
the
other
hand,
we
had
this
sort
of
style
that
I
think
that
we
are
a
little
bit
striving
towards
right
now,
but
not
necessarily
consistently,
which
is
this
using
the
internal
footnotes
and
having
like
a
short
description
of
what
the
link
points
to
together
together
with
the
URL
I.
Think
I'm
personally,
in
favor
of
the
of
the
latter.
One
and
I
think
that
sort
of
the
way
I
think
is
that
the
inline
linking
has
this
risk
that
those
links
will
stop
working
at
some
point.
B
And
then
you
don't
really
know
what
information
was
contained.
So
I
would
sort
of
you
know
I
I'm,
I'm
sort
of
like
a
fan
of
this
thinking
about
like
what
what
sort
of
like
the
practice
is
on
like
these
questions,
I
select,
stack
Overflow,
where
the
actual,
like
information
is
also
like
visible
in
the
place
where
you,
where
you
sort
of
link
to
it
and
I,
think
that
this
footnotes
may
not
be
like
the
ideal
solution.
B
But
this
is
maybe
the
best
we
can
do
with
markdown
without
sort
of
adding
a
lot
of
material
in
this
like
sort
of
without
adding
can
add,
like
this
supporting
information,
not
in
the
text
itself,
but
a
little
bit
a
little
bit
on
the
side,
so
I'm
I'm
more
in
I'm,
I'm.
Thinking
more
a
fan
of
this
of
this
latter
approach,
but
I
think
it
would
be
good
to
hear
if
there
are
what
people
here
think
currently
okay
and
see
that
posted
a
link
to
something
like
a
security
features.
Matrix.
A
B
E
B
No,
no
I
think
so.
So
what
I
mean
is
that
what
I
mean
is
like
when
you
link
to
external
sources?
But
you
just
do
like
the
basic
markdown
links
in
text
right,
so
you
will
attach
it
to
like
a
word
right
and
and
then
like
the
word,
becomes
a
link
and
then
that
just
takes
you
to
the
outside
so
that
it's
actually
part
of
the
text
as
opposed.
E
Does
the
Randall
does
the
the
super
linted
is
that
Lind
links
in
terms
of
checking
that
they're
valid?
E
B
Yeah
but
I
think
that
link
style
would
be
up
to
us
too
up
to
us
to
enforce,
right
and
I.
Think
it's
not
a
consideration
of
whether
the
link
is
working
currently
right,
so
I
think.
Naturally
we
should
ensure
that
the
other
links
are
working
currently,
but
then
links
will
Decay
right
so
like
these
sources
will
cease
to
be
available.
B
Yeah
so
I
guess
that
there
isn't
that
many
that
many
opinions
right
now.
So
we
have
this
suggestion
for
some
cedar
to
use
like
this
Matrix
like
things
to
organize
organize
information,
so
maybe
to
take
this
discussion
forward.
What
I
could
maybe
do
is
that
I
could
maybe
try
and
try,
and
so
we
have
like
this
readme
file
with
this
contributing
contributing
guidelines
which
is
currently
empty,
so
maybe
like
the
citation
style
could
be
something
that
I
could
try
to
write.
B
So
then,
there's
also
like
the
question
that
we
are
all
already
using
using
like
footnotes
pretty
extensively
right
now,
but
I
think
that
those
also
need
to
be
read
and
not
to
use
the
hard-coded
numbering,
but
an
actual
keyword
and
that's
on
that's
on
me.
So
I
made
this
initial
contribution
fairly
fairly
quickly,
so,
but
but
I
think
going
forward.
We
can
we
can
sort
of
move
to
using
like
these
keyboards
because
the
ordering
will
not
otherwise
be.
E
B
Yeah,
do
we
have
do
you
have
anything
else
on
the
ways
of
working,
so
one
one
thing
that
I
was
sort
of
thinking
thinking
about
is
that
how
do
we
deal
with?
How
do
we
deal
with
this
sort
of
reviews
of
of
pull
requests?
So
I
think
that
I
think
that
now,
we've
sort
of
done
this
a
little
bit
ad
hoc
and
I,
think
that
sometimes
we
are
doing
them
offline,
where
people
actually
mark
mark
their
support
on
mark
their
support
on
the
the
the
GitHub
issues.
Also,
sorry,
the
GitHub
pull
requests.
B
B
B
Just
to
make
sure
that
this,
this
sort
of
we
don't
end
up
with
a
lot
of
these
pull
requests
that
are
that
are
that
are
unmerged,
but
but
I
also
hope
to
continue
with
this
style
of
working
where
we
actually
try
and
discuss
at
least
the
major
ones,
major
major
ones
here,
because
I
think
that
I
I
think
that
we've
had
some
really
good
technical
discussions
around
this
I
think
it's
always
good
to
get
multiple
points
of
points
of
view
on
these
and
I.
B
I
hope
that
I
hope
that,
even
though
but
I
but
I
hope
but
I
hope
that,
even
though
you
know
we
might
proceed
a
little
bit
more
rapidly
with
some
of
these.
B
So
if
you
have
any
sort
of
current
concerns,
please
you
know
feel
always
free
to
bring
up
issues
on
on
this
content,
because
I
think
that
I
I
think
that
this
is
sort
of
the
more
the
more
viewpoints
we
can
get
on
this,
the
better
better,
simply
I,
don't
know
if
anyone
else
have
sort
of
Reflections
on
on
these
kind
of
on
this
kind
of
Concepts
how
we
have
been
working
with
the
pull
requests.
So
so
far.
B
Then
I
think
we
then
I
think
we
actually
have
then
some
something
from
this
recently
opened
open
issues
that
would
be
maybe
worthwhile
discussing.
So
we
already
discussed
this
the
the
Microsoft
editions,
and
then
we
had
two
different
issues
opened
by
Robert
C
C
chord
on
this
Linker
option,
specifically
so
I
think
that
he
mentioned
this
Dash
said
no
dump
and
dash
said
no
exec.
B
So,
and
there
was
this
sort
of
like
an
open
questions
that
should
we
be
recommending
options
that
aren't
very
portable.
At
the
very
least,
we
should
be
more
specific
about
where
this
is
supported.
Being
util
isn't
very
helpful
since
50
of
the
linkers
don't
work.
C
Is
more
or
less
deprecated,
it's
been
unmaintained,
so
using
gold
as
a
reference
is
probably
pointless,
but
the
question
about
lld
is
probably
valid.
So
when,
when
I
think,
when
we
say
benutters,
it
should
implicitly
mean
BFD,
LD
and
not
gold,
in
fact
that
there
have
been
conversations
of
gold
of
removing
gold
as
in
dropping
it
from
the
Source
base
altogether.
C
But
there
are
still
some
kind
of
residual
users
that
that
will
hopefully
move
on
either
to
lld
or
to
BFD
in
future,
so
that
we
can
drop
it.
B
F
C
I
think
that
that
argument,
I
think,
is
not
valid,
because
gold
is
essentially
deprecated.
It's.
B
So
so
do
you
think
that
do
you
think
that
this
could
be
like
addressed
by
just
making
sure
that
we
would
be
sort
of
explicit
about
the
state
of
support
in
the
silang
C
language
llvm
link
right,
so
I
think
that
now
we
only
state
B,
New
Deals,
and
there
is
some
I
think
also
some
discrepancy
between
the
versions,
because
we
are
not
giving
the
versions
for
all
of
this
all
of
these
features,
but
I
think
that
can
be
fixed,
but
should
we
also
have
then
like
a
line
that
this
is
not
supported
in
llvm.
C
Maybe
so,
maybe
what's
what's
more
appropriate
would
be
a
comment
stating
that
gold
is
not
actively
maintained
and
when
we
refer
to
the
neutrals
I
recommend
that
you
refer
only
to
the
BFD
Linker
in
the
neutrals
and
not
gold.
I
could
probably
draft
something
to
that
effect.
C
Ahead,
sorry,
just
to
finish,
it
won't
address
the
entirety
of
Robert's
comments,
which
is
the
portability
bit,
but
at
least
it
will
clarify
the
status
of
gold
and
and
the
fact
that
we
we're
not
going
to
refer
to
gold
when
it
comes
to
this
hardening
guide.
And
when
we
talk
about
Ben
neutrals.
We
only
talk
about
the
Benito's
bft
Linker.
Another.
A
Oh
yeah
I
was
just
going
to
mention
that,
so
this
issue
has
several
sub
issues.
One
is:
should
we
even
be
talking
about
gold
and
I?
Think
sireesh
argument
is
Goldie
is
on
maintained
and
for
the
purpose
of
issuing
guidelines
for
hardening.
We
shouldn't
be
talking
about
it,
but
that's
to
leaves
unanswered
the
question
of
versioning,
so
the
issue
is
here
saying:
lld.
Now
something
is
not
supported
in
nld.
A
B
Yeah
I
would
I,
wouldn't
I
would
say
that
that
is
accurate
right,
so
so
I
think
it's
sort
of
like
yeah,
as
you
said
that
there's
like
several
aspects
to
this,
so
on
one
hand
we
can
improve
in
preciseness,
right
and,
and
then
the
other
is
I.
Think
this
open
question
that
should
should
the
guide
include
at
all
things
that
are
not
portable
but
but
I
think
that
we've
we've
already
sort
of
made
this
decision
that
we
are
happy
with,
including
things
that
are
not
portable.
B
As
long
as
it's
sort
of
clear
to
what
tools,
those
those
guides,
those
recommendations,
sort
of
pertain
to
so
I
think
that
we
should
sort
of
apply
that
principle
consistently
to
also
the
also
the
link
or.
But
then
it
I
think
it's
sort
of
like
valid
feedback
right
that
we
need
to
be
more
more
more
precise.
B
Yeah
but
I
think
it
would
be.
It
would
be
great
if,
if
sidesh
can
do
a
proposal
to
to
address
this
and-
and
we
can
then
apply
that
same
principle,
so
something
like
be
specific-
whether
we
are
talking
about
whether
we
are
talking
about
the
bin,
utils,
LD
or
any
other
Linker
and
possibly
then
add
the
note
on
this-
that
we
don't
really
provide
recommendations
for
gold
currently,
because
that
is
that
is,
that
is
deprecated.
B
So
there
was
some
references
of
issues
with
the
Sleep
CP
assert,
which
I'm
not
exactly
sure
what
he
what
he
refers
to,
but
then
there
was
also
this
suggestion
to
include
the
debug
and
I
think
that
I
think
that
now,
overall,
we've
sort
of
come
to
the
conclusion
that
we
should
somehow
restructure
the
options
to
talk
about
both
sort
of
options
that
are
relevant
for
operations
and
like
production,
use
and
operations
that
are
relevant
for
test
use.
So
I
think
that
this
suggestion
sort
of
falls
into
this
testing
testing
use.
B
Potentially
you
want
to
comment
on
that.
Gabriel
yeah.
A
I
was
want
to
say
that
I
think
this
points
back
to
the
nation
that
we
kind
of
surfaced
last
time,
which
is
that
we
have
a
topic
and
then
how
that
gets
implemented
by
each
compiler.
So,
for
example,
here
originally
we're
talking
about
G,
lib,
C,
plus
plus
assertions
right
and
the
purpose
of
that
is
bound
checking.
You
know
access
to
containers,
for
example,
so
we
could
have.
We
have
the
topic
bound,
checking
containers
access
to
Containers,
then,
how
is
that
implemented
for
GCC,
for
example?
A
That
would
be
what
we
already
have
in
a
document
and
then
for
client.
It
will
be
lib,
CPP
assert,
but
the
feedback
we're
getting
here
is
that
the
quality
isn't
there
yet
so
because
the
quality
wasn't
there
yet
their
next
approximation
was
debug,
which
is
not
which
really
doesn't
fit
in
the
original
topic.
A
Like
you,
you
correctly
pointed
out
that
we
have
testing
debug
and
then
release
so
the
what
the
document
has
today
is
about
release
mode
like
when
you're
building
your
software
for
release,
then
you
should
have
that
defined
and
then
I
think
the
third
bullets
in
the
way
I'm
reading
the
issue
is,
is
that
the
llvm,
sorry
clientele,
VM
documentation
itself
doesn't
actually
say
much
about
how
to
use
these
this.
A
This
macro
that
it
doesn't
like
it.
It's
not
as
clear
as
the
GCC
documentation
is
my
subscription
is
it's
follows
the
same
pattern,
and
that
raises
the
question
so
when
we
make
recommendations
like
this,
like
how
do
we
know
that
those
recommendations
have
been
tested
right?
You
know,
because
the
issue
here
is:
oh
I,
try
and
didn't
work,
and
so
is
how
do
we
handle
that
aspect
of
it
now.
B
Yeah,
that's
that's
really,
that's
a
really
good
point
right,
so
I
think
like
in
this
particular
particular
case.
There
were
like
some
indications
in
these
talks
provided
by
the
11
people
that
this
is
sort
of
like
a
feature
that
they,
actually
you
know
it's
not
in
the
documentation
that
they,
but
they
covered
this
in
talks,
and
there
is
also
these
like
active
work
on
this
in
that
that
you
can
find
from
the
from
like
the
version
history
that
they,
they
sort
of,
have
audit
audited
it
that
that
they
conform
to
some
properties.
B
But
yeah
I.
Think
that,
without
having
like
clear
statement
from
the
developers
themselves
on
the
maturity,
yeah
I
I
think
that
maybe,
like
one
option,
is
that
we
can
sort
of
try
and
actually
then
engage
those
developers
and
get
sort
of
like
clear,
clear
guidance
but
yeah.
B
That's
like
a
really
interesting
discussion
right
that
would,
it
also
like
actually
makes
sense
to
you
know
work
towards
some
kind
of
like
some
kind
of
like
a
you
know,
test
harness
for
these
features
features
right
so,
but
but
I
think
that
sort
of
yeah
I
think
that
we,
we
I,
think
that
this
is
also
a
sort
of
like
a
broader
discussion
that
we
should
probably
like
return
to
so
now.
B
It
seems
that
we're
also
out
of
out
of
time,
but
I
think
that
we
can
definitely
return
to
this
right
because
yeah
it
it
could
be
really
great
to
have
like.
Actually
it's
like
some
kind
of
like
code.
That
applies
these
options
and
provides
you
like
some
kind
of
like
a
way
to
you
know
actually
actually
test
them,
but
whether
or
not
that
is
sort
of
like
within
within
the
scope
of
this
this
activities,
then
then
another
question
but
yeah
how
about
we?
B
A
B
Yeah
yeah,
but
we
are
up
to
the
up
to
the
hour
but
yeah.
Thank
you.
Everyone
for
a
very
good
discussion,
this
time
as
well,
and
let's
try
and
see
that
we
can
push
these
issues
forward
offline
and
let's
return
to
the
open
ones
next
time,
but
I
hope
you
all
a
very
good
continuation
of
your
day
and
I'll
talk
to
you
all
in
two
weeks.
I
hope,
yeah
right,
bye,.