00:00:04.350
okay so everyone's warmed up okay so
00:00:12.389
after nary one so my name is waiting I'm
00:00:16.090
actually top engineer Ministry of
00:00:18.370
Education Singapore so we do that for
00:00:20.290
schools to help them teach better so
00:00:23.020
today I'm going to talk a bit about meta
00:00:24.430
programming so first I think I want to
00:00:27.190
talk a bit about why I want to talk
00:00:29.170
about this right so so I wasn't a
00:00:33.040
computer science student and years ago
00:00:35.020
actually I got a job in software
00:00:36.339
engineer so of course the first thing
00:00:38.199
you do is to write code and this was my
00:00:40.329
face when I was reading real bills Xcode
00:00:43.780
right so I don't I don't really get
00:00:45.460
what's going on the next best thing you
00:00:47.499
do is just copy what the existing
00:00:49.420
modifier bit and then submit a pull
00:00:51.760
request and three correct so over the
00:00:56.289
last three years there's been some very
00:00:58.809
nice things that they have gone my way
00:01:00.100
very helpful team one of the other
00:01:02.679
things that happen also so that the team
00:01:05.080
that worked on actually circulated a
00:01:06.370
book so metaprogramming Ruby right and
00:01:09.540
so at that part time it wasn't a
00:01:12.189
priority for me I took one year to
00:01:14.860
finish this book right so that's my
00:01:17.049
place where my parties are but in fact
00:01:20.110
over this one year I think after reading
00:01:21.729
through and rethinking about things I
00:01:23.469
learned a bit more about will be right
00:01:25.509
it's quite similar to how urgently be
00:01:27.850
true that would be sorta the C source
00:01:29.560
code path for me reading through the
00:01:31.360
concepts you know in a story like
00:01:32.799
manners more palatable right and six
00:01:36.189
months ago actually attended Ruby HG
00:01:39.369
Mesa and dragos lost out there trying to
00:01:41.619
give a talk about submitting copper
00:01:44.140
proposed ourselves sitting out there
00:01:45.909
thinking about whether I should try for
00:01:47.740
something like that in six months
00:01:50.320
utopian right so this is my first top
00:01:52.689
four in a row because they're gonna
00:01:55.450
thank you for your assistance
00:01:57.130
yeah okay so I think the other thing
00:02:00.789
that I realized is that for meta
00:02:03.369
programming is for me if I don't do it
00:02:05.950
do it deliberately
00:02:06.969
it comes with bits and pieces as
00:02:08.380
intuition so what I want to do is to try
00:02:10.600
to actually identify certain key tips or
00:02:13.990
set and hope that both the junior
00:02:17.440
developers and a senior developers can
00:02:19.030
also pick up some of this and at least
00:02:21.300
benefit and keep the team
00:02:23.200
SM when you came so today I'm going to
00:02:25.810
talk about why we do meta programming
00:02:27.370
some kids and how not to get yourself
00:02:31.090
killed or do meta programming
00:02:33.480
okay so let's first start with the
00:02:36.640
definition of meta programming so Google
00:02:39.790
reviews many many kinds of definitions
00:02:42.780
the one I like I really like this code I
00:02:45.610
write small code but the more accurate
00:02:48.940
version of this is actually the writing
00:02:50.740
of programs there are multiplayer blah
00:02:53.380
blah so what it means is that if you su
00:02:56.550
as of now we write will be methods which
00:02:59.620
are set data and we give an output today
00:03:02.080
so meta programming means that you are
00:03:04.390
set the program that you have written as
00:03:06.430
a piece of data or an object and you
00:03:08.410
modify it so what you are doing is
00:03:10.510
you're inherently modifying the
00:03:11.860
application behavior at runtime everyone
00:03:16.240
got that ok so the next question said oh
00:03:20.140
this has a complex process so why do you
00:03:21.820
have a metaphor game in the first place
00:03:23.740
right so what I would try to do here is
00:03:26.860
to convince you that it is sometimes
00:03:29.680
better to make a program and I will try
00:03:32.350
to show some examples so we start with a
00:03:34.630
very easy one suppose I need I have an
00:03:37.510
object called a and I need to define the
00:03:41.230
standard getters and setters so the most
00:03:44.230
the easiest way to do this is just to
00:03:46.120
define a Ruby method as follows so just
00:03:47.890
as it goes get and set correct so this
00:03:52.090
is quite a few lines and actually what
00:03:55.090
happens is that you can actually rewrite
00:03:57.640
this in a you know nice length in only
00:04:00.850
by using attribute assessor right so I
00:04:04.390
think everyone's less familiar with
00:04:05.410
attribute Assessor
00:04:06.360
and if you want a very naive
00:04:09.760
implementation of this this is how it's
00:04:12.100
going to look like right but overall
00:04:14.020
what you have is that you have actually
00:04:15.970
extracted some of the logical ways right
00:04:18.220
so developed when developers will be
00:04:20.109
class ad instantly understand what foo
00:04:23.440
and bar actually means to the class
00:04:25.950
right so one of one of the benefits of
00:04:28.830
Mehta program is actually that you get
00:04:30.450
slightly more readable and concise code
00:04:32.160
in the class members in a class
00:04:35.220
definition okay so um another example
00:04:42.750
would be on my to Mehta program this
00:04:45.270
step in certain cases we have we might
00:04:48.720
want to extend this attribute access
00:04:50.910
level further so I guess how many of
00:04:53.340
y'all newsreels here many of you so if
00:04:56.760
you actually include active record base
00:04:59.160
in this manner you realize that rails
00:05:01.470
actually includes all the attributes for
00:05:03.480
you you do not even need to define them
00:05:05.820
and and as you edit the into the code
00:05:08.190
you realize that oh my reels does is
00:05:10.230
that it actually digging through this
00:05:11.610
the schema the database schema he
00:05:14.400
identifies this the set of attributes
00:05:17.780
that this model has and then it
00:05:20.760
populates that right and and so what
00:05:23.220
this does is that it gives you a very
00:05:25.890
awesome attic behavior and when you
00:05:27.750
define a typical classes these are
00:05:29.670
things that would help you along the way
00:05:31.380
so you don't have to spend a lot of time
00:05:32.880
writing it you don't really need to read
00:05:35.130
it because all these columns actually
00:05:38.310
listed in the schema schema and actually
00:05:40.830
all the developers are obey office okay
00:05:44.940
so another another value of
00:05:47.880
metaprogramming the following so suppose
00:05:49.560
we have a rapport model and and this
00:05:52.590
report requires three sticks so you have
00:05:56.880
a job are we doing in a published state
00:05:58.440
and then you want to transition between
00:05:59.850
the different states so what you have
00:06:02.820
this you want code that looks like that
00:06:04.770
on the left however I can see so you
00:06:09.720
have a state you create a new report it
00:06:12.270
is in draft scale what I can do is I can
00:06:14.670
try to submit it motion to release the
00:06:16.500
answer and sort of socialized a machine
00:06:17.940
alright so again a nice implementation
00:06:20.850
of this would be to just cut out the
00:06:23.850
methods as we can right but what happens
00:06:27.810
is that as business requirements change
00:06:29.940
you sub introduced a lot more things so
00:06:32.730
for example the boss comes student is no
00:06:34.620
I think three states is not enough we
00:06:37.170
might want to resize this
00:06:38.990
you know so what you can then do is to
00:06:41.880
go back to the class redefine right even
00:06:45.120
more methods but if you look at it
00:06:47.490
carefully
00:06:49.410
okay so again and so what can also
00:06:53.190
happen is that the boss could also ask
00:06:55.110
you for things like you know I want pre
00:06:57.210
and post transition hoops so for example
00:06:59.100
if I if I were to submit a report can I
00:07:02.130
actually send an email out to everyone
00:07:04.110
to just notify them of this and so and
00:07:06.750
so forth right you want to raise errors
00:07:09.060
so some of these are very common with
00:07:11.460
Pete about logic so what you can then do
00:07:14.310
especially meta program this right for
00:07:17.280
example if you can actually define the
00:07:19.340
state and the transition programming
00:07:22.680
program exactly right then what happens
00:07:26.280
is that you can actually write code to
00:07:28.170
actually generate these methods for you
00:07:29.790
automatically and the folks in there
00:07:31.980
will be a lot easier so what I'm going
00:07:34.290
to do here actually I've actually copied
00:07:36.200
what a gem called workflow has done so
00:07:39.750
what flow does is that it wraps up all
00:07:41.700
this logic into a very nice DSL as you
00:07:45.270
can see so what you have is that you
00:07:47.640
define the initial state then you try
00:07:50.400
the laser pointer money
00:07:52.380
okay memory so what happens that you
00:07:54.960
define state draft reviewing and publish
00:07:58.380
and then you define transitions between
00:08:00.780
them what you get here is that go a bit
00:08:05.040
that this code is more readable for
00:08:06.300
everyone so I think it's beautiful you
00:08:10.830
know because when currently we only
00:08:13.200
existed if you have ten states and penny
00:08:15.900
transitions between the ten states this
00:08:17.520
will look a lot more readable
00:08:19.110
then then having probably about 20 tons
00:08:22.950
about 100 200 methods in there right so
00:08:26.960
what you then do with metaprogramming is
00:08:29.520
then with this so what you do is that
00:08:32.370
you define domain domain-specific
00:08:34.260
languages and then you can hide the
00:08:36.060
logic elsewhere right so this is
00:08:38.070
definitely visible concise and of course
00:08:40.800
not reusable so now you can take this
00:08:42.719
workflow DSL and you can actually bring
00:08:45.450
it to other models and reuse it so just
00:08:49.170
to sum up I think these are just some of
00:08:51.360
the very
00:08:52.500
examples of why metaprogramming will
00:08:54.120
help you so there are more dynamic
00:08:57.210
they provide dynamic behaviors readable
00:08:59.790
concise you don't really repeat yourself
00:09:01.500
but I think all in all it provides code
00:09:06.570
that's a lot more readable I'm sure all
00:09:09.510
of you know that reading code that's not
00:09:11.550
done by you is actually quite painful
00:09:13.230
right so the more visible it is the less
00:09:17.040
pain you feel okay so I'm going to give
00:09:22.800
treated to start method opening so the
00:09:25.590
first is what GPS cover so when I saw
00:09:28.650
toggles correctly but it's actually what
00:09:31.050
you need to do is because for
00:09:33.660
metaprogramming what you are trying to
00:09:34.950
do is to modify the behavior in essence
00:09:38.100
you are trying to modify the behavior of
00:09:39.480
the object or the class so it is very
00:09:42.150
important for you to get your mental
00:09:43.560
model right so if you are to define
00:09:45.960
classes include classes subclass and and
00:09:48.180
so and so forth create objects right you
00:09:50.940
must be very clear on what you are
00:09:53.100
creating and what you're observing so
00:09:55.260
this is really ugly inside you get idea
00:09:59.330
so the first important thing is actually
00:10:02.040
for for people wanting more meta
00:10:03.750
programming to get this mental model
00:10:04.980
right if you don't get this right it's
00:10:06.630
very hard for you to implement proper
00:10:09.030
code in a meta programming sense okay
00:10:13.260
the second step to to actually drink
00:10:15.750
meta programming is to learn how to
00:10:17.220
modify the behavior a sub a sub point of
00:10:21.420
this is actually to understand how to be
00:10:22.980
evaluated so things like how are methods
00:10:26.310
being caught how do you how the messages
00:10:29.430
pass between the objects right what is
00:10:32.060
how do they catch missing message and
00:10:35.100
things like that
00:10:35.790
all right so a very good step for me I
00:10:38.430
assume is that you should read the code
00:10:41.580
will be dogs I would like to point to
00:10:45.180
two to set of functions that I think are
00:10:47.760
quite useful for everyone so if you do
00:10:50.880
need to understand a class or you need
00:10:53.640
to understand class you can use methods
00:10:56.250
like instance variable getter methods
00:10:57.990
one instance methods singleton methods
00:11:00.210
but a single class and if you need and
00:11:03.540
once you have if you need to understand
00:11:06.150
and you need to modify the behavior of
00:11:07.800
it then you can go on to the method on
00:11:10.650
the right so these are things like
00:11:12.000
defining method so you can program it
00:11:14.460
tickly define new methods you can add
00:11:18.510
methods or you can even and define
00:11:21.510
methods right so just a trivial for
00:11:24.330
those who are quite experienced these
00:11:27.450
methods are defined in very different
00:11:29.430
parts of movie so you need to do some
00:11:31.170
hunting if you want to find them okay
00:11:34.070
the last thing is is that it for you to
00:11:37.740
do whenever Graeme is to actually figure
00:11:39.900
out what's the DSL that you want so I
00:11:43.160
can't really give much advice that
00:11:45.840
because it's that context show depends
00:11:47.400
on the team but wise learn is that the
00:11:50.100
first copy is never the best you can
00:11:53.160
always integrate to get slightly better
00:11:54.750
DSL more readable code and what I would
00:12:00.930
suggest is actually that you copy for
00:12:03.360
mothers so I think wills has very good
00:12:06.060
implementations you can look through is
00:12:07.770
very very complex even for the attribute
00:12:10.950
accessor definitions but you can look
00:12:14.520
into some of these libraries or give you
00:12:15.990
an idea of how the dss are done and how
00:12:18.810
they are implemented okay so the problem
00:12:25.500
with meta programming is that if you can
00:12:27.300
modify behavior you can knowingly you
00:12:31.860
can also modify behavior unknowingly
00:12:33.630
right and that's where it gets very
00:12:35.550
dangerous right so simple things like
00:12:38.850
that can happen I have been caught many
00:12:40.800
times where my my teammates have with
00:12:44.580
vision code hidden in in a very small
00:12:47.250
folder somewhere and it which is often
00:12:50.940
very implicit behavior which I don't
00:12:52.290
recognize so it takes me a few hours to
00:12:54.450
debug which is kind of key right so I'm
00:12:58.740
gonna go through three big things so
00:13:00.630
some of the gotchas
00:13:01.350
so I think the first thing is su-metal
00:13:04.800
program try to keep the meta programming
00:13:07.890
code independent so if you do modify a
00:13:11.700
class or modify object please try to
00:13:13.830
keep it towards the object and not and
00:13:15.780
not introduce more dependencies
00:13:19.620
right the second thing is don't matter
00:13:22.380
program meta programming code so what
00:13:24.960
why can happen is that you can add a set
00:13:27.990
of defined methods they look very
00:13:29.190
similar I go oh why not let me just move
00:13:32.220
them together and I can do meta meta
00:13:34.050
programming right there if you I have if
00:13:37.650
this is that if you remember reading the
00:13:40.050
meta programming code it is actually
00:13:42.600
very hard to be meta programming code in
00:13:44.130
itself that's actually an extra loop
00:13:45.900
when you're hit to try to figure out
00:13:47.820
what's going on what you are creating
00:13:49.410
while you're modifying so try not to
00:13:52.200
make it someone readable and that's
00:13:54.420
necessarily yeah the 13 out the slaves
00:13:57.540
exceed that you know more is not always
00:13:59.580
good after this talk I don't think it's
00:14:02.670
right to just go and say let's go make
00:14:04.380
meth above them every day right so
00:14:06.830
imagine the following case where you
00:14:09.000
have a class and you have 20 includes
00:14:12.589
under in front of after just after you
00:14:15.450
define a class so if you if you have an
00:14:18.480
out encounter that is very hard for you
00:14:20.760
to identify which piece of code actually
00:14:23.060
defines this behavior or modify his
00:14:25.440
behavior here the second thing on the
00:14:28.950
talk about is readability so I have kind
00:14:32.190
of shown you ways in which meta
00:14:34.650
programming can help in vwt but there
00:14:37.709
are also cases where it might not have
00:14:39.150
so much so let me give an example I
00:14:41.550
actually found this code online from
00:14:44.459
this gem called restclient so we have on
00:14:47.820
the left a method programming kind of
00:14:50.910
example and we have on the right a more
00:14:53.010
explicit example so who who prefers the
00:14:56.880
meta programming and who prefers the
00:15:01.170
right side the Mike's this example okay
00:15:03.810
so from those arrays that has actually
00:15:06.000
so this is the range testing question I
00:15:07.800
presented this to my team I also got
00:15:09.600
kind of a half ass at the end of the day
00:15:11.880
I think it boils down to the team's
00:15:14.880
composition and it is about what the
00:15:17.339
skin feels right certain certain people
00:15:19.860
if you have a team that has various
00:15:22.529
developers who are in this while working
00:15:24.180
on the Signet for a long time it is okay
00:15:26.550
to have more meta programming code
00:15:28.579
versus the team which has more turnover
00:15:30.810
or prototype that may be explicit is
00:15:32.730
much
00:15:33.240
it's a lot easier to get up to speed so
00:15:38.569
so yeah I want to say is that shot it's
00:15:41.699
not way to be double right and sometimes
00:15:45.540
at some point you have the choice
00:15:47.040
between whether you want more implicit
00:15:48.420
or my explicit behavior okay the last
00:15:51.029
thing to note is always remember to
00:15:53.730
document and test because
00:15:55.679
metaprogramming squad will be it's very
00:15:58.050
important to actually explain what you
00:16:00.029
are trying to do right so I yeah I think
00:16:05.069
we we understand this with you have done
00:16:07.619
a lot of bills you should know that you
00:16:09.480
are you are using a lot of meta
00:16:10.920
bergamini in there before I excellent
00:16:12.990
objections all the callbacks and things
00:16:14.759
like that so the documentation helps a
00:16:16.439
lot in identifying what you actually
00:16:17.939
want to do the other thing is also to
00:16:20.879
write tests so when you introduce some
00:16:24.059
meta programming code right it is when
00:16:27.540
when you read a test it also got you
00:16:28.980
know Monica to get your behavior so you
00:16:30.839
can community confidence it's also
00:16:32.970
important for you to write it so that
00:16:36.209
you kind of tell readers what you are
00:16:39.089
intending to do with this piece of metal
00:16:40.769
programming code so if you see the test
00:16:44.129
below here you you can actually tell if
00:16:47.629
you can see tell like the behavior they
00:16:52.170
are trying to to bring out in this
00:16:54.240
particular model right and so this is
00:16:57.360
for people who don't really like the
00:16:58.769
documentation and if they prefer to read
00:17:00.420
test this is a much better platform for
00:17:02.279
them okay so I think oh I know I just
00:17:11.159
wanna say that actually meta program is
00:17:12.510
your - so it's not really a silver
00:17:14.370
bullet for anybody but in certain cases
00:17:16.380
it does work wonders and is quite
00:17:18.809
magical and the code when the code you
00:17:21.209
have refactor this is a lot more
00:17:22.860
readable right so I just want to share
00:17:26.399
that also be aware of some of the risks
00:17:29.429
that this might have right so when you
00:17:32.039
have when you have a lot of meta
00:17:34.020
programming there's a lot of implicit
00:17:35.159
big it's hard to debug in fact it also
00:17:37.799
takes a long time to actually design it
00:17:39.600
and to do this okay so it's everyone has
00:17:43.740
enjoyed this
00:17:45.140
I hope you did but and if you want to
00:17:47.240
look more into it I can introduce you to
00:17:49.250
this book in the book they would they
00:17:51.860
will go more detail into some of the
00:17:53.480
possible spells or design patterns that
00:17:55.490
you can use and I hope out of this
00:17:59.090
answers since violence so so we get more
00:18:02.510
readable code okay so if that's all then
00:18:07.610
I think thank everybody for listening
00:18:09.980
and not falling asleep yeah and any
00:18:13.309
questions is it possible
00:18:14.840
oops so yes
00:18:23.450
so we take questions now anyone oh yeah
00:18:33.429
yeah thanks for a cameo um basically my
00:18:37.850
thermometer screaming is that you say
00:18:39.650
makes more readable code but anyway it
00:18:41.510
seems like you're just shifting the
00:18:42.650
readability to it makes it readable for
00:18:45.440
the use of the API but you are that you
00:18:47.809
have additional complexity and where the
00:18:49.910
course being written which is why Rios
00:18:51.230
is so easy for beginners to pick up but
00:18:53.480
is so magical so how do you actually
00:18:55.970
integrate the meta programming part of
00:18:57.799
the application into your project you
00:18:59.510
give managers is separate no librarian
00:19:05.230
okay so how do we do it okay so okay
00:19:13.040
actually I have two sliced in China
00:19:19.760
okay I think yeah I was actually working
00:19:24.540
on this example to just try to show some
00:19:26.070
examples of this so for example in you
00:19:28.890
know that actually we have many many
00:19:30.630
models and suppose in one of these
00:19:33.660
models you want to attach okay so the
00:19:35.700
models have South my report said baby up
00:19:38.610
there pictures right and you have
00:19:40.200
article solutions many pictures and
00:19:42.480
these things at attachment site so one
00:19:44.730
way of doing it is the following where
00:19:46.290
you can define it so it's still part of
00:19:49.740
the application code right by hidden off
00:19:52.860
in a text editor
00:19:55.230
I think we keep it we keep it separate
00:20:00.480
okay in the app but separate from the
00:20:02.760
rest of the model and then we just test
00:20:04.650
it yeah so I reviewed that yes it does
00:20:09.090
it does hide some of the complexity away
00:20:12.930
but so that's why I think if you have
00:20:15.900
tested that properly I think is fine so
00:20:18.900
yeah I think I did a video the easy way
00:20:22.590
to unfurl it here alright thanks
00:20:34.340
thank you so much for your talk
00:20:36.990
I was curious to know you mentioned that
00:20:39.620
depending on certain teams
00:20:41.309
metaprogramming might be a wise choice
00:20:43.080
and sometimes not so I was curious to
00:20:45.990
know in your own team how do you
00:20:48.990
determine whether something is a good
00:20:52.409
candidate to be met a programmed and
00:20:54.059
when it should not be okay
00:20:58.610
it's all for us what happens is that
00:21:04.399
when the best kind of candidates are
00:21:08.490
those where you find yourself repeating
00:21:09.690
the same the same things over and over
00:21:12.360
again so if I for this example of
00:21:17.159
attachments right we actually had this
00:21:18.750
we have to repeat this probably about 20
00:21:22.409
or 30 models and I think that was when
00:21:24.720
we made more sense to actually do meta
00:21:27.450
programming on this right for think they
00:21:30.840
are very well defined and the behavior
00:21:32.520
is understood really you should do this
00:21:34.760
for things that they aren't clear
00:21:36.630
because the design the design portion
00:21:39.120
takes quite a long time to do right to
00:21:41.190
come out the DSL and to implement it so
00:21:43.110
for unclear requirements it is better
00:21:45.120
not to involve the much of meta
00:21:46.830
programming right so yeah cool thank you
00:21:56.810
I was wondering what your approach to
00:22:02.150
method missing because all of these
00:22:03.710
examples that is shown here or like
00:22:05.240
defined methods so this is pretty safe
00:22:06.920
meta programming as you're you're
00:22:08.930
defining some sub method and you know
00:22:10.640
that this method will be there
00:22:11.960
method missing is like in all the double
00:22:14.180
explore that there is you you call it
00:22:16.700
whenever the method doesn't exist so I
00:22:18.650
was wondering if you think that it is it
00:22:21.290
is recommended by you to use this
00:22:22.580
approach or you should be extremely
00:22:23.810
careful and use it only when you have no
00:22:26.120
other choice or as the exam wearily yeah
00:22:28.670
okay so I I didn't want to talk about
00:22:30.890
meta meta missing there was anti-pattern
00:22:34.340
on it online but I think I think going
00:22:38.090
to bash in the details but for me if you
00:22:40.340
ask me if you can avoid meta missing you
00:22:43.580
should avoid method missing at all cost
00:22:45.580
right because method missing is if I
00:22:49.850
will see as a global catch-all for all
00:22:51.860
our missing methods but to implement it
00:22:55.040
correctly you need to also implement
00:22:57.700
respond to missing if I remember
00:22:59.960
correctly because that will identify
00:23:01.730
whether method is actually missing or
00:23:03.680
not so you can implement you can catch
00:23:06.410
some of these missing methods but Ruby
00:23:08.840
doesn't know that this method is missing
00:23:11.030
or not and so that's why I think you
00:23:13.670
have to understand object model little
00:23:15.170
bomber two and and see how the message
00:23:17.360
passes yeah so if you can you should
00:23:20.300
avoid it but always good to understand
00:23:23.180
and know how meta missing is implemented
00:23:24.920
one more question do you have any
00:23:26.720
favorite DSL out there in the open
00:23:28.430
source some that you particularly like
00:23:30.410
how readable it is so so I think I
00:23:35.110
really like workflow because I think we
00:23:37.520
have about we do is about four you know
00:23:43.190
okay so there but I what one is actually
00:23:45.530
a like a assessment net so we have
00:23:48.560
answers and wishes for four or five
00:23:50.540
sticks and actually they can rotate with
00:23:52.760
each other so I should've up flow DSL is
00:23:54.680
a very nice example there was one more
00:23:57.170
recently that I came in contact with
00:23:59.750
which I really liked
00:24:01.130
which was
00:24:05.630
it had to do with some passing I forget
00:24:08.820
I I can't like I I can't answer then how
00:24:12.390
much just do it later when you away okay
00:24:14.880
awesome thanks hi anymore okay
00:24:36.100
hi thanks for the talk for the example
00:24:40.010
of the HTTP actions and metaprogramming
00:24:45.440
yeah what would your approach to testing
00:24:48.950
the methods defined in the public API
00:24:52.460
would you just do separate tests for you
00:24:55.730
defining the explicit method like have a
00:24:58.580
test forget put post and greet or test
00:25:01.760
one test one of the possible verbs if
00:25:05.450
they exist what what do you do usually
00:25:07.910
when you met a program like public
00:25:10.700
methods in the objects okay so um okay
00:25:15.230
that different like examples um so okay
00:25:17.300
for this case what I might do is
00:25:20.960
actually to sort out the possible versus
00:25:24.350
just one but let the let the set of four
00:25:28.790
run and see are executed RBC's the
00:25:32.000
method call for something I did for the
00:25:35.300
workflow one wears a bit slightly
00:25:37.340
different now probably have a concrete
00:25:39.040
so I have a dummy class include some
00:25:41.600
workflows defined a methods defined
00:25:44.030
about fluid States and transitions there
00:25:45.560
now just has a test that thanks
00:25:49.250
yep Izumo can take one more question
00:25:55.070
before move on
00:26:01.550
okay thank you very much reaching let's
00:26:04.800
give them our laws
