►
From YouTube: 2022 08 31 Git Cache Maintenance
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
Welcome
it's
the
first
of
september
2022.
This
is
get
cash
maintenance,
google,
summer
of
code,
so
fushakesh.
The
first
topic,
I'm
aware
of,
is
we
need
to
be
sure,
everybody's
okay,
with
extending
extending
the
end
date
a
little
bit
so
was
it?
Did
you
agree
with
john
mark
on
an
amount
and
tell
us
more
about
it.
B
First,
I
thought
of
discussing
like
what
are
the
pending
stuff,
like
what
are
the
things
which
are
still
you
know,
pending
things
we
have
to
work
on,
so
few
of
the
things
are
first
like
the
prefetch
command,
isn't
working
and
like
there
is
an
issue
for
private
repositories,
so
that
has
to
be
fixed,
and
then
I
need
to
write
few
tests.
There
are
some
pending
tests
in
for
you.
You
know
for
functionalities
for
some
of
the
functionalities
and
then
the
javadoc,
the
readme
is
pending
and
a
blog
post.
A
A
Thinking
about
the
layering
of
things,
you
would
have
to
somehow
remember
the
credential
id
that
was
used
to
access
the
cache
and-
and
that
may
actually
not
be
what
the
administrator
wants
is
they
may
say:
hey
I
don't
want
to
grant.
I
don't
want
to
have
something
that
is
remembering
credentials
of
the
cash.
B
Go
ahead.
Okay,
we
have
we.
I
was
thinking
of
skipping
the
caches,
for
you
know
not
like
non-prior
like
for
private
repositories,
but
then
I
don't
know
how
do
we?
How
do
I
know
it's
a
private
repository
or
not
so.
A
Yeah
and
I
think
the
the
answer
there
is
it's
pretty
simple:
if
it,
if
the
repository
uses
ssh
protocol,
it
must
be
private
all
right.
So
if
if
the
repository
protocol
is
ssh,
you
can't
access
it
without
an
ssh
key
now
is
that
rishabh
help
me?
Is
that
correct?
Is
that
true?
Can
I,
if
I
attempt
to
access
a
public
github
repository
with
a
here?
Let
me
do
a
quick
check
just
to
be
sure,
but
I
think
that's
accurate.
C
B
Protocol-
and
it
says
it's
kind
of
expired
on
github-
they
removed
that
you
know
a
password-based
authentication
was.
A
Somewhat
pro
there
is,
there
were
periods
where
github
said
it
was
their
preferred
protocol,
because
it's
there
there
are
elements
of
https
that
are
much
more
scalable
than
ssh
and
so
for
an
organization
like
github
that
wants
to
do
massive
scale.
Delivery
has
much
more
known
quantities
in
terms
of
how
do
you
scale
https
connections
and
how
do
you
scale,
https
processing,
but
the
specific
use
of
the
past
username
password
pair
using
your
literal
password
from
github?
Yes,
they
correctly
have
deprecated
that
and
said
stop
doing
that
what
you
do
is.
B
A
And
if,
if
people
come
back
and
say
hey,
I
want
I
desperately
want
prefetch
for
for
private
repositories.
It'll
take
some
more
design
work
because
I
don't
we
would
have
to
have
some
way
of
capturing
which
credential
was
used
to
create
the
cache.
Remember
that
and
then
reuse
that
remembered
information.
B
So
how
do
we
plan
on
skipping
it
like?
Is
there
any
way
or
mechanism
where
I
could
you
know
any
command.
B
B
Those
on
I
tried
testing
them,
but
I
wasn't
able
to
test
that
so
those
are
the
pending
stuff.
Last
week,
I've
done
a
lot
of
work
on
improvising,
the
ui
the
table.
So
I
spent
a
lot
of
time
and
effort
thinking
which
data
structure
to
go.
So
I
think
we
can
have
a
look
at
that
if
you
want.
I
know,
okay,.
A
C
B
Last
time,
if
you
remember,
we
were
scheduling
a
maintenance
task,
every
minute
and
every
two
minutes-
and
there
were
you-
know,
one
of
the
maintenance
tasks
wasn't
running
right
idea.
It
was
a
bug,
and
that
was
the
same
bug
which
I
faced
during
the
presentation
so
that
so
actually
I
created
a
global.
B
You
know
the
calendar
object
and
then
I
have
to
I
had
to
check
the
date
and
time
every
time
right
whenever
I
read
the
crowns
and
taxes,
but
then,
whenever
I
hated
the
object,
that
was
the
only
time
I
you
know
created
that
calendar
object.
So
it
was
not
created
every
single
time.
So
that
was
the
reason.
A
B
Also,
the
ui
needs
to
be
a
little
more
friendly
like
little
help
tags.
What
is
gc?
What
is
prefetch
little?
You
know
those
things
are
kind
of
missing
right.
A
B
B
Oh,
that's!
Fine!
That's
fine!
Just
refresh
this
page,
oh
just
refresh
it
okay!
So
now
we
have
no
data
as
well
in
this
table.
Right
now
and
now
we
can,
you
click
on
execute
so.
B
B
Yes,
so
this
in
the
next
in
the
next
maintenance
task,
you
would
see
those
as
well.
Okay,.
B
And
you
can
click
on
the
expand
as
well
you'll
find
that
common
craft
data
and
incremental.
So
here,
basically
I'm
storing
five.
You
know
sorry
yeah
five
recalls
for
each
maintenance
task,
so
there
would
be
around
six.
How
much
five
is
25.
I
guess
25
records
for
each
cash.
B
So
this
is
what
I
was
working
on.
This
took
a
lot
of
time.
You
have
to
you
know,
figure
out
how
to
get
the
data
and
all
the.
B
Be
updated
because
this
is
this
is
the
internal
hash
which
is
being
used.
B
We
we
could
provide
the
remote.
I
wasn't
sure
what
do
I
do
with
that.
So
I
didn't
proceed
with
that,
but
if
you
want
like
we
can
go
ahead
with
you
know,
putting
a
remote
url.
A
C
So
expanding
this
feature,
so
if
I,
if
a
particular
task
is
executed
on
a
on
a
cache
and
if
I
expand,
I
can
see
multiple
tasks
associated
to
that
cache
yeah.
So
if,
if
each,
if
each
cache
is
having
its
associated
history
of
all
the
tasks
that
are
executed,
then
why
do
we?
Why
are
we
showing?
Why
do
we
have
this
column
called
tasks
there.
B
C
B
C
So
then,
there
is
no
entry
for
this
cache,
which
has
a
task
called
incremental
repack.
B
Yeah
that
that
is
being
in
the
sub
table
right
now,
so
if
you
click
on
the
expand
right
so
for
that
cache.
So
if
you
click
on
that,
thus
so
the
incremental
read
would
be
stored
here.
C
B
Yeah
yeah
yeah
when
it
ran
it,
would
make
an
entry
there
and
when
a
new
maintenance
task
has
been
executed
on
that
cache,
it
would
be
replaced,
and
this
would
be
added
into
the
sub
table.
A
A
B
So
if
you
observe
right
now,
if
you
refresh,
I
think
we
would
have
one
more
like
can
you
you
get
incremental
repack
here?
Okay,
because
this
is
the
latest
maintenance
stars
run
on
all
the
caches,
and
that
has
also
been
added
into
the
soft
table.
B
A
A
B
I
wasn't
I'm
not
sure
about
that,
because
when
I
tried
it
I
didn't
have
any
tmp
directories.
So,
okay,
I.
B
Empty
get
repositories,
I'm
not
sure.
C
Why
why
are
these
directories
created?
These
temporary
directories.
A
A
C
C
So
rushikesh,
if,
as
a
user,
I
want
to
understand
what
is
the
kind
of
optimization
that
the
maintenance
plugin
has
performed
here,
I
would
have
to
go
to
the
caches
repository
and
just
like
check
the
size
of
the
repository
right
like
do
a
before
yeah.
B
Okay,
wait
so
basically,
this
is
a
string.
I
am
not
sure.
How
is
this.
I
think
this
is
sorting
it
based
on
the
first.
A
A
A
A
A
Sorting
it's
not
expanding.
If
I
sort,
then
I
can
no
longer
expand.
If
I
sort
twice
now
time
for
more
testing
mark
be
a
good
tester.
Okay,
I
can
expand
here
without
sorting.
Now,
if
I
sort
by
repo
size,
I
can
still
expand
if
I
invert,
the
sort
of
by
repo
size
it
doesn't
expand,
is.
Is
that
because
so
is
that,
because
it's
non-zero
no,
this
one
is
25
meg
and
this
one
is
160
meg,
so
I'm
definitely
able
to
expand
them,
and
here
we
see
a
garbage
collection
that
is
non-zero.
A
B
B
It
has
the
thing,
but
then
you
have
to
insert
your
own
html
code
into
it,
so
I
have
inserted
my
own
table
html
into
it
and
that's
how
I'm
displaying
it.
I
see.
Okay,.
A
So
if
I
sort
by
repo
size
and
now
go
here,
let's
go
to
50
entries
sorted
by
repo
size
and
so
we're
going
to
choose
one
that
has
754
megabytes
and
it
won't
expand.
So,
let's
sort
by
repo
name
and
it
doesn't
expand.
Let's
try
refreshing
the
page
and
now
give
me
a
hundred
entries
and
there's
adopt
open,
jdk,
oh,
but
it
still
doesn't
expand.
Maybe
that
means
that
there's
nothing
to
expand.
A
I
did
not
think
that
I
had
sorted
it,
but
let's
try
this.
So
what
I've
done
is
I've.
I
just
refreshed
the
page,
so
that
should
have
stopped
all
sorting
right.
Then
I'm
going
to
increase
the
number
of
entries
to
100
and
that
will
bring
it
onto
the
page
somewhere
there.
It
is
and
when
I
click,
expand,
no
action.
A
B
B
B
Here
we
can
search
for
that
file
so
that
cache.
A
A
B
So
I
think
it's
not
working
for
page,
you
know
pages
after
10.
So
if
you
go
into
that
yeah.
C
A
B
A
Okay,
now
back
to
that,
how
is
it
be?
Oh,
it
won't
help
us
because
the
large
repositories
are
not
in
the
first
ten.
Are
they
okay,
yeah
we've
got
754
megabytes,
which
is
actually
that
is
a
pretty
large
repository,
but
its
number
is
at
number
11
in
the
list
right
now,
although
wait
a
sec
that
doesn't
help
us
right,
because
this
is
actually
not
a
cache
managed
by
the
caching
system.
A
All
right
so
krushakesh
in
terms
of
the
tasks
yet
to
be
performed.
I
think
we
also
need
to
get
a.
We
need
to
release
a
version
of
the
get
client
plug-in
with
your
with
your
api
in
it
yeah,
and
in
order
to
do
that,
it
would
be
good
to
have
javadoc
for
those
api
entry
points.
If
you
can,
if
you
could
take
the
time
to
create
the
javadoc.
Let's
let
me
look
to
see
what
it
looks
like
now,
or
is
it
already
docked?
B
I
didn't
test
the
legacy,
get
maintenance
because
I
was
you
know
for
a
few
version,
so
basically
the
gc
and
the
legacy
get
maintenance
works
based
on
you
know
the
when
it's
required.
So
even
if
I
trigger
a
gc,
it's
not
a
compulsory
that
it
would
run
a
gc
so
basically
based
on
how
optimized.
B
So,
basically,
if
the
cache
isn't
optimized,
then
only
it
would
run
on
gc.
So
I
have
no
control
over
when
I
can
run
on
gc.
So
that's
the
reason
why
I
wasn't
able
to
test
a
gc.
The
same
thing
for
comic
graphs
as
well.
So,
basically
in
previous
versions
of
get
a
comment
graph
is
run
only
if
it's
only
after
gcse
there
is
a
configuration
for
you
have
to
set
a
config
file
in
the
get
a
global
config
file
to
you
know
enable
the
commit
graph.
A
Okay,
so
so
then
do
you
you
attempt
to
launch
the
commit
graph
command
and
if
it
fails,
just
don't
worry
about
it.
Oh.
B
It
would
drop
so
basically
what
would
happen
when
you
run
a
comment
graph,
the
thing
command
runs
or
the
would
you
tell
underlying
git
software
runs
it,
but
it
doesn't
performs
any
action.
It
will
just
exit
like
without
throwing
an
error.
Oh
okay.
So
all
right,
so
I
won't
even
know
that
if
the
comment
graph
has
been
run
or
not,
okay,.
A
B
Because
we
have
disclosed
it,
oh
and
the
you
know
community
bonding
phase,
and
then
I
asked
the
question
as
well
like
do
we?
Do
we
do
a
gc
with
an
auto
or
a
forced
tc?
You
were
like,
we
would
prefer
going
with
an
auto
because
it
would
be
safe
and
we
wouldn't
disrupt
the
you
know
caches
on.
Oh
right,.
B
B
So
here
I've
written
the
tests
as
well
for
all
of
them,
except
prefetch,
except
prefetch.
A
Very
good,
okay!
Well,
so
so
the
calling
the
legacy
I'm
I'm
not
sure
it.
I'm
not
clear
on
the
difficulty
of
calling
the
get
the
legacy
maintenance,
it
seems
like
you
certainly
can
call
it.
You
may
have
excuse
me,
you
may
have
difficulty
deciding
if
it
was
executed
or
not,
but
even
there
there's
this
logger.
So
you
could
certainly
check
to
see
that
the
message
was
logged.
A
A
A
A
A
A
C
Mark
wasn't
there
a
way
for
us
to
get
the
builder,
I
mean
the
job
logs
from
the
jenkins
instance.
Yes,.
C
A
A
A
Okay,
let's
check
it
in
the
git
client.
Maybe
that's
where
it
is.
A
So
I
think
this
listener.getlogger
will
let
you
then
read
the
contents
of
that
we're
logged
to
the
listener.
I
think
okay,
so
for
your
cash
at
least
it's
worth
your
exploring
okay.
B
B
There's
a
object
right
in.
A
It
would
help
if
I
didn't
have
these
regular
expressions
properly.
Okay,.
B
A
A
A
A
So
this
one-
maybe
that's
what
I
was
thinking
of
was,
but
in
the
worst
case
you
could
do
that
as
well.
For
your
cash,
add
log
statements
and
assert
the
log
statements.
Okay,
you
know
say
hey
at
fine
level
logging.
I
expect
to
get
this
and
then
during
normal
operation,
you're
not
doing
fine
level
logging,
and
so
it's
okay.
A
Okay,
so
back
to
the
back
to
the
tasks
or
I
guess
we
had
started
this
this
conversation
with,
do
we
do
you
need
extra
time
to
be
able
to
successfully
complete
and,
and
can
we
give
the
extra
time
john
mark's
answer
to
me
was
yes,
we
can
give
it
within
a
limited
amount.
What's
your
sense,
do
you?
Will
the
extra
time
benefit
you
do
you
need
it?
Oh.
B
I
am
not
sure
about
it
because
we
have
few
pending
tasks
like
our
javadocs.
Are
there
and
then
the
prefetch
is
left
and
then
writing
these
tests
are
left
and
then
yeah
every
week
a
bug
keeps
popping
up
so
I'll
have
to
so
I'm
not
sure
about
it,
but
I
think
it
would
be
fine
if
they
go
with
another
extra
week
like
I
think
that
would
be.
You
know
sufficient
is
what
I'm
thinking.
C
C
And
the
the
presentations
that
we
give
within
the
jenkins
community.
A
So
fushakesh
would
do
the
presentation,
potentially
during
during
the
that
that
period
and
that's
okay.
A
A
C
B
C
A
And
I
I
think
that
I
think
that
is
okay,
because
I've
got
increasing
pressure
in
about
that
same
time
as
that's
approaching
the
date
for
devops
world,
and
I've
got
major
things
that
I'm
delivering
a
devops
world
so
extending
by
one
week
is
workable.
Extending
anything
more
than
one
week
for
me
is
not
workable,
but
I
think
we
could
do
one
week
and
and
we'll
just
go
from
there.
I
don't
feel
nervous
about
what
we've
got
like.
B
Because,
right
now
we
are,
I
think
we
are
in
a
good
state
like
things
are
working
as
expected,
but
if
we
have
that
extra
week,
I
think
we
could
finish
off
better.
You
know
try
to
wrap
up
stuff,
you
know
even
better
way.
Try
to
figure
all
about.
You
know
fix
all
the
bugs
in
right.
You
know.
A
Yeah,
so
I
mean
we're
at
a
point
now
where
it
would
be
good
for
us
to
get
a
release
of
the
get
plug-in
of
the
get
client
plug-in
and
get
plug-in
out
with
the
changes
in
it,
so
that
we
can
say
hey
we're
set.
That
would
be.
That
would
be
a
for
me
next
week
is
probably
too
fast
for
that,
because
there
are
still
some
diagnose
diagnostics
that
russia
cash
you've
got
to
do.
The
following
week,
though,
would
be
really
good.
B
Is
there
a
way
to
write?
You
know,
java
dogs
like
easy
away,
you
know,
is
there
a
standard
or
something
which
I
need
to
follow
like
to
write?
It.
A
C
B
So
I
I'll
get
this
detail
I'll
get
the
javadoc
for
the
gitline
plugin
ready
by
this
week.
I'll
also
try
to
integrate
the
prefetch,
the
prefetches
and
I
tried
looking
into
it,
but
I
was
facing
some
difficulties
but
I'll
try
yeah
you!
Last
week
you
recommend
using
the
get
ls
remote
command.
If
I'm
not
wrong,
do
you
know
check
whether
it's
a
private
repository
or
not.
A
B
B
This
week,
we
I
think
by
this
weekend
the
git
line
me
plugin
would
be
ready
and
then
I'd
also
try
to
checking
at
looking
at
the
errors
in
the
ui,
like
the
error,
which
you
have
seen
right,
the
bugs
or
other
than
that
other
than
that
any
ui
changes
which
needs
to
be
made
like
the
ui.
A
I
did
not,
I
did
not
see
any,
but
the
the
total
experimenting
I
did
was
with
the
with
what
we've
got
now.
So,
let's,
let's
do
some
well
and
I
apologize
I'm
running
out
of
capacity
I've
I
flew
in
from
alaska
this
morning
on
an
airplane
and
I
haven't
slept
yet
so
I'm
now
approaching
20
hours
awake
so
so
I
have
to
get
a
hard
stop
here
pretty
quickly.
A
C
As
soon
as
it's
as
long
as
it's
the
first
10
increase,
it
works.
A
B
A
B
I
I
can
about
that.
I
can
send
a
message
right
if
we
need
to
meet
like
I'll.
A
A
So
so,
for
my
for
my
clarity
for
me,
as
far
as
I
can
tell
the
get
client
plug-in
changes
that
you've
made
have
been
quite
stable
right.
They
they,
you
haven't,
had
to
make
any
significant
increment
increases
to
the
functionality
there.
So
it's
it.
I
think
it's
safe
for
you
to
market
that
it's
no
longer
in
draft
and
add
add
a
few,
a
javadoc
comment
or
two,
and
then
then
we're
probably
ready
to
release
that,
and
I
need
to
do
a
release
of
the
get
client
plug-in
anyway,
because
I
need
to
release
the
here.
A
A
C
A
C
C
A
C
B
Oh
okay,
also,
can
you
upload
this
the
recording
for
the
session
so
that
yeah.