►
From YouTube: Elsa Workflows Community Meeting 24 ( 2022-11-03)
Description
Demos
- Asynchronous activities
- Workflow driven UI by Frans
Topics
- Elsa 2.9.x
○ 2.9.2 to be released ASAP
Next time:
- Code spaces by Mohammed?
- Workflow driven UI for V3? ;)
Feedback
- How to deal with running jobs and join activities?
Master branch currently broken if you don't have .NET installed
A
A
A
A
B
Good
evening,
hey
class,
hey
yeah
I
had
once
more
question
actually
and
the
extent
of
what
we're
using
for
the
UI
driven
Elsa
support.
Basically,
we
noticed
that
when
you
have
a
mixture
of
a
huge
tasks
and
background
tasks,
basically
there's
an
Gap
in
between
the
presentation
that
you
can
do
in
the
uise
when
the
user
task
isn't
running
and
the
background
task
is
running,
pretty
I've
query,
for
instance,
all
the
user
tasks
to
get
a
to-do
list.
B
But
when
it's
running
a
background
task,
it's
not
waiting
for
an
user
task,
so
I
it
pops
in
and
out
of
the
list
and
to
mitigate
that
problem.
I
saw
there
was
a
metadata
which
I
can
use
then
to
say,
for
instance,
the
current
state
and
so
on,
and
the
current
user
that
is
working
on
it
and
so
on.
But
I
was
wondering
what
is
the?
Are
there
any
best
practices
in
using
metadata
versus
variables,
for
instance,
and
if
there
is
any
guidance
on
that
metadata.
A
Versus
workflow
variables,
yeah
all
right,
all
right,
yeah,
let's
see
if
we
can
talk
about
that
and
get
an
understanding
of.
What's
what
and
what
would
be
better
for
for
your
scenario:
So
Great
Migration
from
V2
to
V3.
Let's
start
with
that,
and
the
question
more
elaborately
typed
in
here
was
based
on
where
else
of
version
3.
Is
that
currently
how
much
work
would
it
be
to
migrate
from
version
two
to
version
three?
A
The
this?
The
short
answer
is,
basically,
you
will
have
to
re-create
your
workflows.
The
model
of
V3
is
completely
different
from
V2.
That
being
said,
we
have
been
talking
about.
You
know,
potentially
adding
a
let's
say,
an
LSAT
2
activity
type,
because
in
answer
three,
you
can
create
custom
activities
that
represent
different
kinds
of
workflows
or
diagrams.
So
technically
that
should
be
possible,
but
it's
also
a
bit
time
consuming.
A
A
A
Close
to
you
know,
one-on-one,
because
of
the
the
bulk
of
the
logic
is
applique
would
be
application
specific
or
specific
to
that
particular
activity,
and
it
the
the
main
differences,
will
just
be
the
syntax
around
it
to
to
follow
the
also
three
syntax
that
there
will
be
some
work
involved
there,
but
that
would
be
mostly
copy
and
paste
in
terms
of
existing
workflow
instances.
They
would
not
Port
over
automatically.
You
would
have
to
implement
some
some
data
transformation.
Some
some
mapping
or
another
strategy
could
be
to
let's.
A
But
if,
if
you
could,
if
you
need
something
somewhere
in
the
beginning
of
next
year,
then
elsa3
might
be
a
good
option
to
start
looking
at
at
the
beginning
of
next
year.
I
wouldn't
recommend
looking
at
into
it
right
now,
because
there's
there's
still
broken
aspects
that
that
are
real
blockers
to
for
reality
to
be
properly
used
without
any
guidance
or
documentation.
A
B
C
A
Yeah,
that's
that
makes
sense.
If
you
don't
have
too
many
workflows
or
just
a
few,
and
you
need
some
custom
activities,
they
will
be
easily
transported
or
ported
to
elsa3,
and
then,
if
you
don't
have
too
many
workflows
or
they're,
not
too
complex,
then
it's
easy
to
so
supposedly
to
be.
It
should
be
easy
to
recreate
those
workflows
and
also
three
once
it's
ready.
A
A
Which
is
the
new
documentation
place
for
version
two?
So
let
me
explain
a
little
bit
about
what's
going
on
here,
so
we
have
documentation
for
Elsa
2.
Today
it's
hosted
on
Elsa,
workflows,
GitHub,
IO
and
and
that's
fine
that
that
there's
no
no
problem
with
the
host.
The
problem
is
with
the
underlying
npm
packets
that
generates
this
website
based
on
markdown
and
it's
using
docus
hours,
which
is
an
open
source
project
created
originally
by
Facebook
I
believe.
A
A
A
It's
host
on
versal,
which
is
a
pretty
cool
hosting
environment
and
and
it
has
cicd
building
so
every
time
you
put
something
to
this,
repo
person
will
build
we'll,
take
that
and
and
create
a
new
build
and
deploy
it
automatically
to
the
to
this
host
and
all
of
the
content
is
there,
it
has
a.
It
looks
a
little
bit
different,
but
the
structure
is
the
same.
A
A
So
that's
all
that's
automatically
solved
here,
which
is
nice,
although
it's
not
yet
implemented
so
I
need
to
sign
up
for
algolia,
which
I
did
just
now,
and
they
are
going
to
review
and
make
sure
that
this
is
indeed
an
open
source
project.
You
know
some
commercial
or
project
so
I'm
waiting
for
their
approval.
Once
that's
done,
we
I
can
start
hooking
in
hooking
up
this
search
UI
into
the
Argo
area.
Backend
and
very
important
is
the
dark
mode
feature
which
we
don't
have
with
the
current
website.
B
A
And
they
have
a
new
sex
and
compare
plates
and
it's
it's
looks
pretty
amazing,
so
they
use
just
Tailwind
CSS
to
create
all
of
these
beautiful
templates,
so
they
have
Salient
or
a
SAS
marketing
template
something
for
your
personal
website,
transmit
Etc
and
then
the
documentation
uses
the
syntax
template
yeah
and
it
looks
it
looks
like
this
as
you
can
see.
So
it's
it's
pretty.
B
A
B
B
A
Yeah,
okay
yeah,
it's
just
to
just
to
include
with
the
notes
it
is
here.
It's
also
documentation.
It's
this
one.
So
with
the
V2
documentation
goes
into
the
V2
folder
V3
goes
there
yeah
the
link
right
and
sort
of
fee
three
documentation.
This
is
all
it.
This
is
not
the
the
real
text.
Well,
this
is
just
a
tryout,
but
the
menu
Etc,
it's
still
just
from
the
from
the
syntax
template.
A
So
it's
a
I
I
have
I
really
have
to
just
still
get
started
with
this
documentation,
but
I
want
to
get
the
version
2
documentation
out
of
the
way
first.
So
next
up
for
me,
is
to
get
get
started
with
the
other
three
documentation,
turning
all
of
the
packages
or
the
projects
into
nougat
packages,
so
that
you
know
they
can
be
actually
be
downloaded
from
my
gets
or
GitHub.
A
C
A
C
C
C
Free
for
public,
open
source
projects,
I
want
to
say,
I,
don't
know,
I'm,
not
sure
what
the
like
storage
limit
is
on
it,
but
you
there's
also
built-in
there's
some
actions
from
GitHub
to
clean
up
old
packages.
So
you
can
go
clean
up
all
your
old
previews
on
a
release.
So
if
there's
a
PR
to
master
closed,
you
can
go
clean
up
all
the
old
previews
that
aren't
valid
anymore,
with,
like
five.
B
A
That's
great,
then
it's
then
everything's
in
one
place.
Even
the
doctor
registry,
it
seems
like
it's
forces
replaced
by
container
at
the
Sri
Lanka,
very
cool
yeah.
So,
needless
to
say,
I
cannot
wait
to
get
started
with
this,
and
once
once
we
have
some
some,
you
know
quick
starts
and
installation
guides
with
elsa3.
Then
then
it's
then
that's
also
the
moment
where
people
can
of
course
start
to
try
it
out
and
and
be
a
assured,
hopefully
that
it
won't
change
too
much.
A
There's
still
a
lot
of
missing
features,
little
quirks
and
and
little
usability
issues
in
the
designer,
but
but
part
of
the
goal
for
me
to
go
through
documentation
and
creating
these
quick
searches.
To
is
for
me
to
go
through
all
of
these
steps
and
then
and
run
into
those
issues
so
that
I
can
you
know,
identify
them
and
create
issues
for
them
and
fix
them,
because
there's
a
lot
still
there's
there's
a
ton
of
work
left,
but
not
everything
is
equally
important.
A
C
A
Yeah,
that's
good,
but
probably
we
should
add
some
some
Banner
that
this
is.
This
is
work
in
progress
and
it's
it's
it's
not
actually
to
be
released.
I
I
pushed
it
just
to
try
out
the
cicd
to
try
reversal
Etc
but
yeah.
We
should
make
clear
that
this
is
there's
absolutely
not
the
release
and
it's
it's
all
very
early
three
Alpha.
Even
let
me
make
a
note
of
that.
Yeah
like
this
yeah
I,
didn't
expect
that
to
be
for
people
to
just
find
it.
C
A
C
A
It
still
is,
it
still
is
just
place
voltex,
except
for
the
home
page,
it's
a
little
tryout,
but
it's
already
out
of
date,
because
you
don't
have
to
manually
protect
your
endpoints
anymore
out
of
the
box.
The
endpoints
will
be
protected
so
and
it's
I'm
not
even
sure
I'm
gonna
go
with
open
edict
out
of
the
box
or
open
ID
text.
I,
don't
I,
don't
know
how
to
pronounce
pronounce
this
open,
I,
detect
or
open
edict.
A
Yeah
I
think
so
too,
and
then
then
I
had
a
meeting
with
somebody
and
he
pronounced
this
open,
ID
dick
and
it
would
make
sense
it's
open,
ID
and
index,
but
to
normal
humans.
I
would
imagine
you
read
it
as
open
edict
in
any
case,
so
everything
else
still
is
template
text.
So
please
don't
read
all
of
this.
A
One
one
thing
I
still
haven't
figured
out:
I
guess
is
the
how
to
deal
with
different
versions.
So,
for
example,
first
year
else
at
two
it
it
gained
a
lot
of
features
com.
If
you
compare
version
2.0
with
2.9
and
I'm
wondering
how
to
best
structure
this
in
the
documentation.
So
with
the
current
version
of
the
documentation
for
version
two,
it's
it
works
like
this.
A
Docusarus
has
versioning
built
in,
so
you
can
go
to
the
1.0
documentation
or
to
2.1,
as
you
can
see
here
or
2.2,
which
is
the
current
stable
version
according
to
the
documentation.
Of
course,
this
is
not
the
case.
We
are
at
version,
two
point
9
and
then
there's
privileged
versions,
which
is
supposed
to
be
the
the
next
version.
So
this
is
a
nice
system
of
Docker
service.
That's
not!
We
don't
get
that
with
the
syntax
template,
but
one
idea
I
have
is
that
maybe
this
should
be.
A
A
B
This
is
why
I
created
that
option
to
have
different
menus
on
the
left
and
basically,
in
the
pages
of
the
syntax,
you
could
just
copy
it
to
a
specific
version
and
change
the
menu
and
in
the
menu
either
on
the
left,
like
I
mentioned
or
in
the
menu
itself.
You
can
switch
them
to
different
versions,
and
it
basically
goes
to
the
old
site
by
just
copying
it
to
a
subfolder.
A
A
A
B
Yeah
Hugo
was
asking
if,
because
he's
working
on
some
interesting
projects,
that
I
told
you
about
that,
he
was
seeing
how
you
can
migrate
old
or
existing
running
workflow
instances
into
the
new
version
of
it.
And
that's
he's
writing
a
paper
on
that
one,
and
he
ran
into
that
problem
that
the
instances
that
were
idle
were
weren't
starting.
A
A
B
A
B
C
B
A
B
A
Yeah,
because
the
The
Bookmark
was
resumed
right
yeah,
so
the
user
performed
their
action
which
resumes
The,
Bookmark
and
the
bookmark
being
resumed,
it's
basically
being
consumed,
which
means
it's
being
deleted
and
therefore
there
will
not
be
any
tasks
in
for
to
be
displayed
to
the
user
and
if
there's
some
background
worker
processing,
the
workflow,
which
could
take
a
while
and
for
a
certain
amount
of
time.
There's
no
actions
were
used
to
be
taken.
Yeah.
A
C
B
B
I
see
when
I
get
the
workflow
instance
details
I
also
get
the
metadata
back.
It
basically
is
a
dictionary
of
key
value
and
I
can
project
some
major
things
into
that
metadata,
indicating
it's
this
case
number
it's
this
status,
that
it
has-
and
this
is
the
log
for
instance,
so
the
UI
could
still
show
some
details
about
the
workflow
and
the
state
of
it,
while
not
being
actively
running
on
a
user
task.
B
Now
I
was
a
bit
doubting
like
okay.
Is
that
the
best
practice
Yeah
and
I?
It
was
kind
of
funny
because
they
implemented
it
and
they
moved
now.
Everything
into
the
metadata
all
variables
are
gone
and
everything
is
in
the
metadata.
I
thought.
Well,
it's
an
option.
I
have
to
say
that,
but
if
it's
a
good
option,
I,
don't
know
and
I
wanted
to
hear
about
the
perspective
of
it,
and
maybe
the
things
that
you
encountered
already
by
using
metadata.
A
Yeah,
so
let
me
see
I
I,
don't
recall
using
metadata
myself,
but
it
is.
It
is
there
to
store
to
give
activities
and
an
option
to
store
some
some
data,
or
maybe
some
application.
I
I,
don't
recall,
but
let
me
take
a
look
at
some
of
the
usings.
If
there
are
any
and
looks
like
in
the
Elsa
solution
itself
nobody's
really
using
it,
I
mean
it's
persistent
and
it's
being
loaded,
that's
basically
it,
but
from
from
a
data
structure,
point
of
view
I
think
it's
perfectly
fine.
A
The
variables
are
there,
you
know
just
to
just
as
a
more
explicit
structure
for
or
construct
for,
workflow
implementer,
so
process
design
is
to
to
use
you
know
to
to
this
to
set
variables
and
get
variables,
but
I
don't
see
any
any
issue
if,
if,
if
your
application
is
usually
metadata
directly
because
at
the
end
of
the
day,
both
kinds
of
stores
get
persisted
as
part
of
the
workflow
instance-
and
this
is
just
a
dictionary
but
variables
you
know
are
like
this.
This
object
here
is.
A
A
A
Dictionary
under
the
cover,
so
from
that
point
of
view,
it's
fine,
it's
more
semantically,
it's
different
than
I
I
would
say:
variables
are
intended
to
be
used
by
process.
Designers
and
metadata
is
supposed
it's
more
intended
would
be
more
intended
to
use
to
be
used
by
activities
or
application
code
just
to
store
whatever
you
want
to
associate
application.
Specific
information
with
a
workflow
instance.
A
It's
possible,
but
I,
don't
know
if
there
are
any
expressions
for
that
available
today,
but
you
it
would
be
easy
to
implement
either
from
your
application
itself
or
by
adding
it
to
one
of
the
just
the
core
Library
itself.
So,
for
example,
go
to
scripting
and,
let's
say
JavaScript,
it
would
be
good
to
to
add
it
to.
A
Let's
see
handlers,
don't
forget
JavaScript
engine
to
add
a
function
somewhere
here
to
get
metadata.
It
would
make
sense
and
I
just
also
found
that
actually
there
is
code
using
set
metadata
and
get
metadata,
for
example,
the
compensation
service
that
was
added
relatively
recently,
and
it's
as
you
can
see
it's
storing
some
internal
State.
A
B
B
Okay,
cool,
okay
and
yeah.
I
will
go
like
this
because
I
like
the
way
that
just
the
way
to
get
the
workflow
instances
and
just
yielded
the
metadata
automatically
as
well.
So
I
could
leverage
quite
easy
on
that
to
extend
the
UI
and
show
some
more
details
about
the
workflow
instance,
while
not
running
a
user
task.
So
yeah
great
feature,
though.
A
Yeah
makes
sense,
one
thing
that
it
doesn't
provide
you
with
is
the
ability
to
search
on
any
of
those
custom
fields,
and
for
that
reason
we
need
to
implement
at
least
I
think
what
we
should
do
is
Implement
and
some
search
indexer
like
Lucine
or
elasticsearchs,
so
that,
potentially
you
know
whatever
your
application
Associates
with
your
workflow
instance,
you
should
be
able
to
search
on
any
of
those
fields.
Let's
say
you
have
workflows
dealing
with
customers
or
orders.
B
A
A
A
We
are
looking
at
the
alpha
3
project
board
here,
and
this
is
an
attempt
to
to
visualize
everything
that
is
still
going
on.
It
doesn't
communicate
what
we
need
to
do
for
the
release.
Candidates
for
that
I
actually
should
set
up
a
milestone,
but
at
least
this
is
some
known
work
that
well
that's
going
on
in
that
that
probably
most
of
these
will
have
to
be
done
before
the
release
candidate,
especially
the
little
bugs
protoactor
activities,
absolutely
not
a
must-have.
This
is
a
nice
to
have,
but
there's
little
things
like
automatic
layout.
A
That's
very
important.
It's
not
a
complicated
feature
but
important,
because
it
allows
the
designer
to
display
workflows
that
are
created
from
codes
and,
as
you
may
know,
the
lsa3
designer
uses
position.
No
information
like
X
and
Y
coordinates
because
it
Sports
drag
and
drop,
but
for
coded
workflows
there
is
no
such
information
available.
So
we
need
some
automatic
layout
feature
to
lay
out
the
workflow.
So
this
is
important
for
the
release
candidate
other
than
that
I
think
it's
mostly
things
like
the
ability
to
add
a
workflow
variable,
there's
no
button
to
add
it.
A
There
will
be
some
some
categories
of
activities
that
are
lacking
at
the
moment
or
in
some
of
these
activities
are
there,
but
they
lack
certain
features
so
stuff
like
that
and
as
I
mentioned,
while
I
go
through
documentation
and
imagine
using
else
of
query
from
from
the
perspective
of
a
new
user,
that's
just
just
starting
out
going
through
a
journey.
I
will
identify
the
most
important
issue,
so
it's
gonna,
probably
gonna,
take
a
few
weeks.
A
I
would
imagine
and
then
there's
cool
new
features,
so
Elsa
priority
has
many
new
cool
features
like
asynchronous
activity,
execution
to
name
one
thing.
Another
thing
is,
of
course,
composite
activities
which
is
not
the
same
as
Elsa
tools,
concept
of
composition.
It's
different.
It's
more
similar
to
work,
Windows,
workflow
Foundation,
as
we've
covered
in
past
meetings,
but
also
things
like
generator
activities,
which
it's
experimental,
but
it
would
allow
you
to
put
an
activity
on
a
workflow
somewhere
in
the
middle,
doesn't
have
to
be
starting.
A
Nonetheless,
so
yeah
timeline
wise
still
looking
at
q1
of
the
next
year,
so
hopefully
January
there
will
be
at
least
some
packages
to
try
it
out
so
I'm
I'm,
not
so
convinced
it
will
be
release
candidate
quality,
but
it's
also
not
impossible.
So
I'm
not
gonna,
say
it's
not
going
to
happen.
I
also
not
I.
Also
can't
say
it
will
happen.
We
will
see
it's
getting
closer,
that's
for
sure
all
right,
that's
all
I
have.