►
From YouTube: SunPy Developer Discussion
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
All
right,
hello,
okay,
so
I
honestly
have
been
paying
no
attention
to
means
on
Phi
repo
over
the
last
week,
I've
been
too
busy
kind
of
right.
This
right
this
right.
This
workshop
this
week
and
my
talk
that
I
gave
on
Friday
its
regular
try
to
go
into
a
nice
meeting
in
London
and
ask
people
to
start
up
and
talk
about
some
pie,
how
many
people
just
used
python
and
about
half
the
room
for
the
handle,
which
made
me
very
happy.
That's.
A
C
D
Talk
hey,
they
were
a
carpet
that
they
were
Larissa
from
Walter
homes
and
leanings.
It
was
a
topic
were
local.
They
were
another.
Why
Oliver
a
little
searching
over
for
that?
A
yes
see
down
from
for
that
and
for
a
talk
at
msl.
Thank
you
and
I
think
unless,
if
she
was
also
ready
with
it,
because
a
the
Rio
Sonia
said
mostly
read
in
a
hundred
years
of
the
woman,
oh
woman.
F
B
G
Found
not
much
to
be
honest,
Debbie.
A
H
A
D
A
D
A
D
So,
while
doing
that,
I
found
out
that
we've
been
testing
wrongly
over
the
last
son,
how
many
time,
how
much
time
and
then
with
the
numpy
versions.
So
at
the
moment,
all
the
tests
as
it
is
now
is
always
tested
with
the
latest
version
of
nampa,
and
we
were
thinking
that
we
were
doing
with
the
1.9
and
one
point.
D
I
D
Jackets
I
a
sec
that
so
when
I
managed
to
run
in
with
the
version
that
we
were
thinking.
The
hash
is
the
same
one
that
it
was
that
used
to
be
on
the
on
the
last
time
we
change
the
hash
and
I,
and
so
that
those
hashes
are
you
the
same,
and
that
is
roughly
when
they
dump
I
version,
like
maybe
a
week
after
the
vampire
version
was
released.
So
it's
probably
the
time
that
Conda
updated
their
repository.
J
G
D
A
D
D
A
G
D
H
H
A
A
hash
you
know
your
hash
in
the
committal
dash.
Well,
no
in
the
test
runner
for
the
figure
test.
You
know,
we've
got
the
mark
the
figure
2
yeah,
so
that
if
you
run
the
figure
tests,
it
validates
your
version
of
number,
I'm
a
poor
look
and
if
you
don't
have
the
right
version
on
by
Michael,
look
at
skips
a
bigger
tests.
Today,.
G
A
D
G
Expecting
polygamy
have
advocated
us
before,
but
wouldn't
it
be
a
little
bit
easier
instead
of
getting
stuck
with
these
figure
hashes
all
the
time
to
turn
them
into
wines
as
opposed
to
failures
and
then
make
it
really
in
the
warning
provide
a
really
easy
way
to
compare
to
look
at
the
old
figure
and
the
new
figure.
Something
like
that.
Yeah.
A
There
are
we
this
yes,
the
current
figure
tests
that
we
have
is
far
from
optimum
if
the
dancers
that
I
mean
week.
The
easy
solution
is
that
we
just
stick
the
actual
PNG
files
in
the
repo
and
use
the
podcast
extension
that
Tom
Robitaille
wrote,
because
that
has
a
small
tolerance
for
small
changes
to
the
figure
and
break
that
often,
and
then
we'll
have
version
figures,
which
means
we
could
do
the
comparison
birth.
Well,
we
didn't.
G
G
Mean
you
can
have
that.
Why
not
have
the
figure
hash
have
the
figure
do?
What
we're
doing
now?
Have
the
figure
hash
where
we
have
them
issue
a
warning?
And
then,
if
you
want
to
compare
it
goes
out,
and
you
can
just
gives
you
the
link
to
the
figure
on
github
the
old
figure,
and
it
creates
the
new
figure
for
you.
And
then
you
can
look
at
the
two
yeah.
A
H
D
D
D
This
post
posting,
where
Utley,
because
our
Travis
was
wrong.
We
were
using
the
different
numba
version
that
we
were
thinking
we
were
using,
so
we
were
not
doing
the
proper
testing
as
we
thought
and
that,
okay,
that
failure
is
worth
clean,
ask
that
they
were
something
wrong.
We
were
against
updating
the
hashes
and
we
were
not
saying
why
anyone
so
now
that
we
have
found
that
as
a
problem
that
nullify
Burcham.
C
G
A
C
A
A
H
H
G
K
D
Yeah
we
should.
We
should
have
this
figure
safe,
somewhere
and
yeah,
be
able
to
check
and
different
hashes
body
changed
a
probably
we
could
have
a
github
repository
and
each
version
or
each
time
this
has
change.
We
create
a
new
version
and
then
get
help.
You
can
compare
the
different
versions
with
different
github,
saying
that
they
have
a
movie
that
the
only
owner
there
yeah.
G
D
H
D
G
I
still
feel
like
this
is
this
is
sunk
cost
you
guys
sticking
with
the
system
because
we
spent
all
this
time
developing
it,
but
it's
actually
not
providing
very
much
useful
functionality
right
now,
so
it
is,
it
is
still
doing
its
job.
If
the
figures
were
to
change,
it
would
break
maybe
one
out
of
ten
times
it
actually
told
us
something
useful
yeah.
A
A
All
the
time
we've
been
bonding
it
we
also
haven't
modified
any
of
the
plotting
code
or
anything
that
would
change
the
plotting
code.
That's
not.
I
smoke
sure
what
all
right
you
have,
but
they've
actually
done
it
Scott.
You
know
it
hasn't,
found
an
error
because
it
has
been
an
error.
It's
not
like
it
isn't
working.
D
Yeah,
but
is
no
it's
not
because
they
called
or
the
technique
is
just
because
the
testing
we
were
running
they
brought.
We
were
thinking
that
we
were
in
a
different
num
paya
version
when
he
was
not
so
our
it
probably
was
on
Travis
when
we
were
selling
their
versions
to
run
the
code
and
numpy
was
in
all
them,
updated
to
delete
this
one
and.
G
A
A
This
is
what
this
has
been
an
outstanding
issue,
since
we
implemented
this
the
first
time
we
will
massively
underpowered
here
and
it's
one
of
the
many
things
that
someone
needs
to
get
around
to
doing
right,
but
I
still
think
we
should
rule
converted
to
warnings
because
it's
causing
our.
But
let's
see
what
happens
when
we
change
the
not
fixed
where
we
do
a
fixed
version
built
and
add
the
package
burgeoning
like.
A
A
G
A
D
I'm
stupid,
you
need
to
fix
us
up.
I
helpless
as
well,
yes,
oh,
that
will
be
so.
Hopefully
the
fix
of
that
will
make
it
more
robust,
yeah,
so
I
don't
know
the
people
looked
and
at
the
same
time
I'm
using
a
this
thing,
which
is
call
so
I'm
trying
to
change
the
way
that
we
are
using
promised
by
a
package
in
master
file,
which
is
called
a
strop
I,
see
I
helpers,
which
is
kind
of
a
simple
way
to
to
configure
all
the
travelin
and.
A
The
different
runs
and
stuff.
It
also
means
in
the
way
that
the
CI
helpers
works
is
that
it's
cloned
at
the
beginning
of
the
Travis
helper
build
the
Travis
build,
and
then
it
uses
that
so
none
of
the
configure
scripts
for
any
of
the
continuous
integration
lives
in
some
pie
anymore.
It's
in
a
separate
a
trip,
I
repo,
so
it
can
be
updated
independently
of
the
source
code.
C
D
A
I
I
A
B
G
A
A
A
No,
you
only
wanted
to
you
only
want
it
to
its.
If
that,
if
you
do
only
the
only
time
you
wanted
to
write,
the
PNG's
to
the
directory
is
when
it
updates
the
figure
tests
per
click,
a
hatches
file,
so
the
only
time
it
writes
a
new
figure
hashes
file
is
if
the
figure
hashes
file
has
been
removed.
So
you
just
want
to
modify
the
bit
where
it
writes
a
new
figure
hashes
far,
but
also
save
the
PNG's
to
a
given
by
rectory.
A
A
C
K
Well,
it
was
going
to
be
a
additional
module,
wasn't
a
senpai
affiliated
package
and
there's
really
just
I
need
to
do
some
sanding
around
the
edges
to
get
it
to
be
sort
of
to
the
standard
that
I
think
that
we
need,
but
it's
it's
there,
it's
functional,
as
is
I,
need
to
potentially
also
do
a
bit
more
documentation.
I've
done
some
examples,
but,
to
be
honest,
I
haven't
really
worked
a
lot
over
the
last
month
or
so
couple
months.
B
C
K
F
H
G
C
G
D
D
I
F
J
D
A
A
Like
where
you
calculate
the
power
where
the
power
is
calculated
it,
unlike
line
1358
on
the
pull
request,
yeah
you,
you
calculate
the
rounded
down
exponent
and
then
do
ifs
on
that,
but
you
could
do
it
with.
If
Don
you
could
do
it
using
actual
full
bottom
quantities,
and
then
you
wouldn't
have
to
do
the
conversions
and
the
value
extraction
and
stuff
you
could
just
do
comparisons
on
quantities.
G
G
That's
the
last
opening,
like
you
know,
we
don't
in
this
pull
request,
I'd
like
to
remove
the
other
functionality.
We
had
this
functionality
twice
in
some
pie.
Already,
it's
not
like
I
added
a
second
one.
I
was
just
fixing
the
second
one
yeah
it
should
remove
the
other
one
as
well.
I
just
haven't
had
a
chance
to
do
that.
A
A
A
A
C
G
C
E
E
D
A
D
C
D
C
A
A
A
A
There's
no
way,
there's
no
way
to
remake
the
hashes
for
just
one
thing:
at
a
time
you
just
remake
all
the
hashes.
Ok,
I
will
do
that.
So
do
that
and
push
it
and
we'll
see
what
happens
and
if
it
doesn't
work
then
we'll
fix
it
in
master
first
and
then
you
can
deal
with
it.
Awkward
I
would
make
sure
that
you,
when
you
create
the
hashes,
you
create
the
hatchet
with
the
latest
version
it
not
by
enough
of
it.