Player is loading...
Why you need to be reviewing open source code
Gris Cuevas Zambrano & Holden Karau, Google Cloud;
Wednesday, June 12, 2019 · 5:46 p.m. · 52m 22s
If you want to understand a new project, or grow your skills faster in a project your already working in, doing code reviews can be one of the fastest ways to improve your skills. Documentation is often suggested as a good starting point into a new code base, but I think for many folks reviews can provide a smoother on-ramp to a new open source project. Doing code reviews provides the opportunity to interact with people already familiar with the code base, ask questions, and learn. Not only is this good for your personal growth, the community needs more reviewers to remain healthy.
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
um so thanks for coming early uh i have a survey which or restrain trick people into filling out um
00:00:08
it's not related to this talk particularly but that should not stop you from
00:00:11
filling out this survey you any thoughts about testing spark a one of these days
00:00:16
be really estimate estes park nodded trash fire uh i can i do that with your help
00:00:22
um but this talk is about why you need to be reviewing open
00:00:26
source code um i'm home either co presenter but unfortunately she ended up in
00:00:31
london uh so she's not with me uh but to talk about that a
00:00:34
little bit um if anyone wants the slides this is a bit the link
00:00:39
which has at least a seventy five percent chance of taking
00:00:42
you to the blog post where the slide will end up
00:00:47
eventually after they cut off the alcohol tonight at the party um
00:00:53
chris couldn't make it she did some of the analytic saying a show here i burn a
00:00:57
rather version of us at the bus with a slightly different take on some of the same data
00:01:01
uh i'm home i prefer pronounce if you're or i'm a developer advocated google
00:01:06
i'm on the sparking m. c. i. contribute to a lot of other projects um and in
00:01:11
doing so i have a lot of opinions about what we need to review open source code
00:01:16
um but that being said you know i'm not speaking as a member of the sparkly m. c. these are my own first audience
00:01:22
um i work it will but i'm at a whole bunch of other places they all pay
00:01:26
me money and were very nice people um i'm a co author of two books on sparc
00:01:32
uh if you don't own it the second one high performance park i got a much better world video on that one
00:01:38
uh and you should buy several copies of it uh learning spark whatever i don't care about that
00:01:45
uh informing and what or if you're the kind of people who are interested in
00:01:48
code reviews i have heard of you light streams we're gonna talk about those little bit
00:01:52
and i also collect feedback on my talks if you have a sort of
00:01:57
opinions about what i could do to make this talk better or other talks they do better in the future uh please you
00:02:02
know when you know the only thing i really care for his comments on my style i 'cause that's probably not gonna change
00:02:09
a cool in addition to who i am professionally i'm trends clear
00:02:14
canadian uh in americana with visa expires april uh and part of the leather maybe
00:02:21
and this is i mean not directly related to open source uh although certainly being from canada him
00:02:27
might see some more sort of communal aspects and maybe even reason america uh but i think it's it's important
00:02:34
for us in open source communities to try and build diverse communities amount of things that we need to do that
00:02:39
is talk about we're all from other backgrounds are and if only talk about our backgrounds we realise
00:02:44
that we're all from the same background it's maybe a sign that we need to work on improving art
00:02:50
um and also heather has this wonderful
00:02:54
collection of us go up fried logo is
00:02:58
um and you can buy t. shirts of these uh they are quite lovely i
00:03:03
own one i think actually to um and all were on the morrow and this
00:03:08
is the website if anyone wants to buy one of those will be t. shirts
00:03:13
um and actually the dress i'm wearing is not that but it is about would spark road uh it computes pie
00:03:19
unnecessarily on uh distribute yeah very very lovely okay i don't talk about what you can get
00:03:26
from participating open source code uh views why did you need your help reading open source curve
00:03:32
how to go about doing this um and then i'll probably try and putting you really quick
00:03:39
uh apache sparkled review at the very last minute uh will
00:03:43
review a p. r. together uh and maybe you'll find some
00:03:48
features that we cannot the original author no about together or maybe you know
00:03:52
will just be able say looks good to me and that's pretty fun too
00:03:56
okay so i'm hoping are nice people um you're not laughing at
00:04:00
my mediocre jokes though so i don't know um maybe you wanted
00:04:05
or your free or maybe you just wanna make open source less
00:04:07
terrible that's why you're here um is anyone here for another reason
00:04:14
now okay zero people how many people are here because they wanna make open source slightly more sustainable
00:04:21
okay so not everyone is that's lovely good good uh okay so this
00:04:25
is a picture of a cat um there will be more of these
00:04:30
okay uh okay this is a picture of a mermaid school
00:04:34
um and this is definitely for sure are related to open source
00:04:38
um because i expense mermaid school uh on my corporate credit card and my boss is like holden
00:04:46
does better have a good reason has like yes do not worry boss mermaid school is totally related to my job
00:04:53
i open source code reviews are like a mermaid school to help you grow your skills
00:04:58
meeting ways that you weren't necessarily expecting right i didn't expect to become a slightly better swimmer
00:05:03
by going the mermaid school you know that it's kind of a little bit of a surprise
00:05:07
on the building your existing skills so we can take our existing
00:05:10
knowledge swimming or existing knowledge of style and apply them in different ways
00:05:15
are you get better with time but but you you just need to start at some
00:05:18
point right if you're not just gonna put on a mermaid tail and get in the pool
00:05:22
you just never gonna become a mermaid right and and the same thing is true for open source countries
00:05:27
um and sometimes people will give you push back and uh not want to um
00:05:32
pay your expense or for for mermaid school or will be like
00:05:35
a a a what's with this like it did open source code reviews
00:05:39
on friday thing like that that doesn't sound like you do anything for
00:05:42
me and both of those situations um coffee makes it better and the
00:05:48
telling your boss too much information is a sed never gone well for
00:05:52
me um but but shudder to my boss for paying sixty six percent
00:05:56
of mermaid school uh and paying for my open source code preceded hit
00:06:00
take that uh expense report not not a hundred percent but you know
00:06:04
thank you uh i really hope you never watches my talks okay
00:06:10
so why does that meaning you um so a lot of projects
00:06:14
suffer from maintainer burnout uh if we look at the spur project there's
00:06:18
all kinds of people who can commit code in in theory and
00:06:22
in product is from this like a list of about like sixty people
00:06:26
we see maybe half of them remain active in a lot of them are just like oh it
00:06:31
i'm gonna go and do something else yeah you know i've i've worked on this
00:06:34
problem for so long it's not fun for me anymore um and the other one is
00:06:39
are reviewing code is less time and so we we have this thing where we like especially promote people
00:06:45
and you like open source projects and and just like promoting people in real life we tend to take
00:06:50
other skills you know reading code and then asked them to do something
00:06:54
different and just assume the skills are gonna translate and not provide any support
00:06:59
that goes as well as promoting engineers to managers um and it's it's often tends to be less fun right
00:07:05
we got into this 'cause we wanted to write code and now my job is we're viewing a lot of courage
00:07:11
right well it's up in the morning news like i'm gonna check out all the open pull requests on my favourite open source project
00:07:19
oh okay i wanna freak the three of you when becoming might open source projects um but that's that's lovely
00:07:26
um and some projects are limited by reviewers so even in situations or we don't
00:07:31
have burned out right we can get in the situation where projects become really really popular
00:07:36
all kinds of people are coming to train hopeless but then we don't have the time
00:07:40
to like figure out which of these things are good and which of them about writers just
00:07:44
too many changes being requested at the same time um another one is representing the user um
00:07:52
i didn't know a whole lot of sort of discussions on give
00:07:57
up on one could say where there's like really great technical insights
00:08:01
really good but then you know someone would actually coming to like so what what how are they gonna use this i i don't
00:08:08
although this looks like a really interesting solution but what was the problem
00:08:13
um and i'm sure we'd never have that happen in this colour community but you know
00:08:18
i think having people that can represent the users of the libraries present and participating
00:08:24
the review can help ensure that the things that we bill are actually useful to you
00:08:29
um some situations in in one of the projects their work on a we literally build in if you are an a.
00:08:34
p. i. that was designed to be called from java that could not be called from java but it was really beautiful
00:08:42
uh_huh so that took hundred for something but if we don't get a user they probably wouldn't like i can't
00:08:47
i can't pass you a class to a wide now with with with this i don't know anyway it's okay
00:08:54
um so you hear a bunch of open source projects are
00:08:58
without labels um and we're only gonna label the healthy ones
00:09:02
so the top people in the unhealthy ones don't yell at me um so please don't recreate this analysis so we
00:09:08
can say like hey people make pull requests um and then
00:09:12
they they open he orders and sometimes these players get merge
00:09:16
i'd say that the project is perhaps not super healthy if it's making a lot of p. r. s.
00:09:22
that are just getting close right that might be not design of the obvious to me i'm not saying
00:09:28
that we have to accept every poll request that comes our way i'm just saying
00:09:31
the ratio starts to look a little bit like that blue dot uh or that
00:09:36
well there's two blue dots okay the dark blue dot the the one over there yeah okay cool it might be set
00:09:42
i hear some happy project happy project happy project happy project happy project
00:09:49
i'm very happy project no oh no not so happy
00:09:53
project okay uh never mind was get that one okay cool
00:09:57
uh so how do we get into this situation so i think a lot of the time to get into the situation
00:10:02
'cause the rate of inbound pull requests and the rate of review comments looks
00:10:08
like this does anyone wanna guess which line is which oh a label to graph
00:10:13
just might not be such targets um but the blue line which is bigger is the
00:10:18
number of paul requests were receiving uh and this is not this part this is like
00:10:23
all projects and give hope for the data that i'm actually get my
00:10:26
little hands on um and then we see the the number of you know
00:10:30
were you comments they use a lot a lot smaller right and they can be kind of disappointing right 'cause you you know
00:10:36
maybe you're really excited you if you make a poll request to a project
00:10:41
you're like i fix my bo and hope to save the world or at the very least tell
00:10:45
people click on had more efficiently i live in san francisco supposed to wear the same thing um
00:10:52
and then no one no one comments on your poll request and who will take on those ads for you
00:10:59
that's true we should be set so okay um and then we could
00:11:03
we could look at stylish project this is kind of hard he uh uh
00:11:07
extraction of what's our projects are um discoveries made this morning
00:11:13
um but we can see we have sort of similar patterns
00:11:17
and uh we can we can still see the sadness that happens i'm there's probably some interesting
00:11:22
correlations but there's still more p. r.'s uh then there are reviews five that's needy little sound
00:11:29
um and then if we look at this think user accounts uh we can see that it looks pretty similar um and i
00:11:35
was honestly hoping for a slightly different result here i was hoping
00:11:39
that the red line would be even lower than it is here
00:11:42
uh but i figured i should show you it 'cause it disagrees with what i thought it would be on the spur project
00:11:47
the red line is really really low um but in the others cover projects it's not nearly as unhealthy as sparks so quickly
00:11:55
okay but still not healthy right we want these to be
00:11:58
more bounce right uh thirty to one ratio is kind and happy
00:12:05
okay so that's that's why that means you what benefit can you
00:12:10
get from doing these things so you can grow your skills and some
00:12:13
ways that will look at you can see the we're all the open source
00:12:18
it's uh yeah and if you do it enough you can see the real
00:12:22
world to be let us act um arguably you can get faster recognition um so
00:12:30
and and to anyone has contributed could spark i'm really sorry uh i have
00:12:33
no idea you are right uh but for the review words i i know them
00:12:38
because i see them again and again and it's a lot easier to learn if reviewers
00:12:44
judgement is like really really south summit yeah i would make the same call was that person
00:12:49
or i had that same question um and there can be picking review so many
00:12:53
more p. r. is then they could create um both good p. arsenal don't create
00:12:58
two hundred back here is another project just waiting and that's that's not gonna help
00:13:02
and and you get this deeper integration into the community right you tend to be interacting
00:13:07
the other people who are also involved in deciding the
00:13:11
direction of the curve right to where the project is going
00:13:14
um and you can also a time box sort of how much uh time you're gonna spend
00:13:20
on your open sort activity how many people here have like forty hours a week dismantle opens
00:13:27
yeah cool so uh not a lot of students ah that's not good
00:13:32
um and so i think one of the things is like like this honestly don't have enough time to me a large
00:13:38
number of p. r.'s anymore right like i i can make the p. r. is for the things that they really care about
00:13:43
but one of the things that i've done my life is being able to set
00:13:46
aside you know small amounts of time and allocate and one proving open source projects
00:13:52
so how do we see more of the world so starter issues often only touch a narrow parts
00:13:57
of the code for doing a code review though you can cover a lot more of the code
00:14:02
what is your to dig in 'cause we see what the proposed changes are along with descriptions of what these changes are
00:14:08
um and there's there's only so many hours in the day
00:14:11
that we could spend meeting changes to the code compared to spending
00:14:15
the same hours figuring out what's going on while looking at
00:14:18
someone else's proposal um and it gives you a better sort of
00:14:23
holistic picture of where the product is going and you know people will still yelling on the internet
00:14:28
but you know at the very least they want your annual but you're gonna be yelling
00:14:32
you about your decisions about someone else's kids that feel slightly less detached that's good okay
00:14:38
um so me he will get faster recognition from the community of viewers really stand out
00:14:44
in projects with large numbers of contributors um and it allows you to work with more people
00:14:50
um and it's easier to control your time right uh how many people here have tried to contribute upon request we're
00:14:56
really large project okay keep your hand up if that
00:15:00
folk um who less or equal time as you expect it
00:15:07
oh okay yeah um and so this this is this is okay it um but there was like one
00:15:14
person who was like maybe on the fence about that ah and the thing with meeting upon request is
00:15:19
wonderful reviewers will come to me with all of the suggestions of how to
00:15:22
improve my code nebula god damn it he to go to my day job now
00:15:27
i guess i'll get back to this and next tuesday um
00:15:31
but then i feel obliged to come back couldn't keep updating it
00:15:34
and if i don't do it it's gonna style it's not gonna move forward but if i come in
00:15:39
they can provide some us and you know feedback on how other co can move forward more quickly you
00:15:44
know i don't have to close that loop all the time i i can lead the suggestion and then
00:15:48
it's up to them whether or not they wanna listen to it or not ah yeah and you know
00:15:54
tend to be less frequently encounter in the one more bad situation we're like
00:15:58
i coat almost works i'll just drink another cup of coffee and everything all the
00:16:03
fine and it's it's a hammer like fungus off doing that with the cover view
00:16:08
stop it's nothing you know don't okay um and yeah it's not
00:16:13
it's not just like random projects uh it tends to be the larger
00:16:16
successful project which i need help uh with reviewers and we can see here
00:16:22
uh the apache euro project which is not our but it's it's another one out
00:16:26
was making me is probably the creator of uh we could say co creator or whatever
00:16:32
um and we can see that it's rapidly becoming a critical dependency blah blah blah or uh and we are
00:16:38
in quite serious knee of more project maintainer is um this is this is a really common thing right like
00:16:44
people are really excited about this project but no one stepping up to review the code 'cause it's not fun
00:16:50
so we can treat ourselves and tele sales there's
00:16:53
gonna be fun for all those reasons yea okey um
00:16:58
how many of you have experienced to wean code reviews of any sort
00:17:03
yeah yeah that's lovely keep your hands up if you
00:17:06
have experience with open source of requests reviews as a reviewer
00:17:11
okay that's a good number but not all of you so if you if you don't
00:17:14
pull request reviews internally or coders internally that's great uh if you're gonna start doing open source
00:17:20
code reviews there is some things to keep in mind if you probably don't know these
00:17:25
people um you may have less shared assumptions
00:17:28
and if you see things other accidentally mean
00:17:32
it's a lot harder to go to someone in the like i'm really sorry what i meant to say is that generator
00:17:38
looks slow not you're a terrible person can i buy you
00:17:41
a cup of coffee right um because it will be like
00:17:46
no i'm i'm in europe and you're in san francisco oh also
00:17:50
you're strange and from the internet i don't wanna work with you um right okay uh sometimes strangers
00:17:56
plus the internet become ass holes so it's it's true you might not be the person who's like
00:18:01
leaving you know for for full request a few common specs and maybe you know you
00:18:05
put your heart and soul into those out you you've really thoughtful really well thought out comments
00:18:10
and then you get back something that they don't interfere keyed on stage uh but it rhymes with
00:18:17
i don't know what rhymes with fuck um
00:18:21
okay well it rhymes with a yak yeah there that's not gonna get
00:18:26
it it is it oh whatever my boss as much michael it's okay um
00:18:31
and so hopefully the people that you work with your company are going to be also as
00:18:37
as much as strangers on the internet and if they are you should find a new job um
00:18:43
but you know this is one of the things you're gonna have to prepare yourself for is that unfortunately you might need some of
00:18:48
these wonderful people i am and alternatively you might just have very
00:18:52
different goals and visions for the project right um and that's totally okay
00:18:59
if you're new you're probably not gonna be able to steer it as much as you want but you can probably stuart
00:19:04
more by being the person who's making suggestions on people's proposed
00:19:08
changes uh then proposing the changes that are these that's my experience
00:19:12
okay if you want to watch videos of me sitting underneath
00:19:16
the bed in san francisco i because office space is really expensive
00:19:21
i'm doing open source for request reviews law right
00:19:26
and swearing a lot uh 'cause no one watches part sorry not know what are my boss definitely does
00:19:32
not watch those um you can go to this fine we're out here and you can watch some of those
00:19:38
how to perform okay right and if you're in the mood to reviewing
00:19:43
you can bring so much special value to our project um you can oh
00:19:49
uh that was not the special value of i was hoping for and you know just avoid touching my computer okay
00:19:55
and you can see the weird things that i've just been we've always done it this way or
00:19:59
like of course we start the python and then the j. v. m. and connect them with unix socket
00:20:04
what do you mean that's not normal right like you can be like whoa i um
00:20:10
and and you know maybe the i mean really does need to be
00:20:13
there sometimes we really do need to connect python to the j. v. m.
00:20:17
because my life is full of sadness up but sometimes we don't need to connect python to the
00:20:22
j. v. m. and you can be the person who can like point this out to someone and
00:20:27
just remove that blind spot on and this is this is really awesome and you can also do
00:20:32
a much better job of like being like a a uh this thing which you think is completely intuitive
00:20:38
i it's now i i have no idea what you're doing here right
00:20:43
and and nintendo as a newcomer to can also be much less threatening right
00:20:48
if you're pointing this out to someone else was also you
00:20:50
contributing it's like your hair her like barbies but like sometimes
00:20:56
uh if you like are working some has been our project for really really long time you're like oh my god
00:21:01
for so scary or so scary to hit my code i'm terrible but if
00:21:05
you have like another new person showing up and helping with you it's like
00:21:08
oh no it's cool we're we're in this together and you can you can
00:21:11
take advantage of this and be be less intimidating encourage you please be less intimidating
00:21:16
um okay yeah to do oh right and oh yeah you don't have the same assumption
00:21:21
uh sometimes uh it this is how we find out security box uh someone comes in to take a look at the
00:21:27
thing and they're like i'm i'm pretty sure you base sixty four in code that i wrote a letter to a file
00:21:33
pretty sure that was called password i. t. x. t. oh i have some questions and you'll be like uh
00:21:39
e. yeah yeah um and so you'll you'll look at a library and you can like actually spend the time
00:21:44
to configure their because it's the right thing to do but everyone else is just blind to that 'cause it's you know what they've always done
00:21:51
you don't you should take a project get familiar with the tools can find a p. r. review
00:21:57
i'm still taking a project works a lot better if it's in the same language that you already understand
00:22:03
i'm trying to learn a project by doing trying to bring evil programming language by doing code reviews
00:22:09
now language does not work well by my experience um and i think this is both someone who's been
00:22:17
doing that and on the receiving end that neither of those experiences good to try trying pick a project we're
00:22:23
you know the programming language at least a little bit that said morris teleconference so there is several options for you
00:22:30
um ideally one which needs review up right so if you're looking at a project
00:22:34
and it's got you know three pull requests and they're all actively commented on that
00:22:39
maybe they got snow pick another one um oh yeah and i'll pick something with uh let us goals
00:22:45
um right um i've checked the mailing list see if there's a code of conduct uh see if anyone's heady
00:22:52
unexpected gardening leave 'em or yeah that that that job did not translate um that
00:22:58
refers to when someone is put on suspension and it takes me from a project to
00:23:04
it's essentially the alternative to being fired um but yet so paper project that looks happy in in front of people
00:23:13
um this is a tool for mermaid school
00:23:17
and uh it's really important um if you're
00:23:21
not so good at not reading through your no so you can put this on your nose
00:23:25
so that you don't get a bunch of water up your nose when you're flipping around in circles and just like a mermaid school
00:23:32
uh yeah you need to get familiar with the projects what's right so like if i just showed you this much like okay cool
00:23:38
i go have fun emirates going might be via my school something different
00:23:43
uh right but yeah okay so we we need to like look at some of the other pull first don't just jump
00:23:48
in india like me what's up where is where is uh with this yeah system like put some time investigating what's going on
00:23:55
saying things you would do as if you were gonna start making poll request to it and but keep playing
00:24:01
get wall over all of those things are really learning
00:24:03
unless to make sure they're not hassles really good on figuring
00:24:07
out how to build upon request from the branch yeah c. i. systems are lovely the like actually being able to
00:24:13
verify that something works as intended locally as opposed to
00:24:17
just depending on the c. i. is is a good thing
00:24:20
um get familiar with the communication style or and yet for example i wouldn't take the
00:24:25
linux kernel as a place to try those um or some other projects um i think smaller
00:24:32
players are better if you can find something that you care about that's even better on
00:24:37
because then you can be like hey what's up this looks really close to solving my problem
00:24:42
well it's maybe we could do x. right um and so like one of the
00:24:46
things i care about is like price on packaging for java projects because i chose
00:24:53
i mean some interesting life choices early in my career and uh now every
00:24:57
six days i'm a screen virtual environment uh if you don't get that job that
00:25:02
that's good um and it's often easier to be one of the really reviewers so
00:25:06
if you look at a poll request it's got like six reviewers maybe they got it
00:25:10
maybe not you know definitely feel free to take a local um but i would probably pick something with less people
00:25:17
i'm doing your first review there are all kinds of like things
00:25:22
and i really encourage you like soften your language um which
00:25:27
is the thing that i still protect us um and this is this is very useful when you're working with people who don't know you yet
00:25:34
um i also when you are legitimately uncertain of what's going on
00:25:37
right ah sometimes you know when we're new rear projector like i
00:25:41
think this is what you wanna do is this really what you wanna do and that's actually really useful because they can come back and
00:25:47
be like no that's not what i wanna do we should clarify
00:25:50
that's um you can also read the style going to style checking
00:25:55
or you can also check to make sure that it works on your platform other someone out there who really cares about i'm sixty four
00:26:01
ah and then you show up a lot to tell us that things don't work
00:26:04
on 'em sixty four and that that provide value 'cause i'm not gonna do that
00:26:08
had that i just i don't i i haven't intel computer it's it's it's okay i
00:26:14
well no not is actually who knows um but it's it's not so bad right
00:26:18
um and also if you if you have an i. d. i
00:26:21
find uh i tend to working emacs because once again pour like choices
00:26:26
uh but when i work with other people who are new
00:26:29
uh to a project it and use gentle definition really fancy things
00:26:33
unlike fiasco anything do a lot better with that uh compared to
00:26:37
my like a job to fly all you get three um yeah
00:26:43
okay please communicate carefully i'm telling people the code socks it might be true
00:26:50
but you can do so much better right we can tell them hey i really appreciate the thought
00:26:57
you put into this it's a really great first
00:27:00
step i think the performance has some opportunities for improvement
00:27:05
right and that sounds like what my boss told me in our last performance review
00:27:09
cycle and it didn't saw can tell like i thought it through and i was like
00:27:13
she's telling impediment job right home but you know people won't think it
00:27:18
always looks good um you don't know how much time someone's but it might
00:27:21
like especially if you do viewing purpose from other new people maybe this
00:27:26
is been a week's worth of working you're gonna tell them that's all bad
00:27:29
so to speak and um and also it's okay for this
00:27:32
to be scary um i remember starting doing poll request reviews
00:27:37
and being like oh my god these people are making all these changes they must know so much more than i do
00:27:43
but i don't think i can use the software is like this but
00:27:46
i don't know this is really terrifying and it's okay to be scared um
00:27:52
yeah yeah whatever uh i drink um whatever you
00:27:55
do to cope with that uh is fine i recommend
00:28:00
if you do write your pupil request reviews about drinking us to make them the following day well so
00:28:07
uh we get have has really wonderful features for this um if you're using review board mm hum or whatever um
00:28:15
and also remember it's okay for people to be wrong on
00:28:18
the internet and it's okay for projects to choose suboptimal things
00:28:22
you don't have to fix everything if someone's too painful for you to work with just
00:28:27
move on and let someone else deal with them i'm cool yeah phrasing emmys matters what um
00:28:35
okay so i was gonna trying to eli apaches bar code
00:28:39
review because that seemed like a good way to kill i think
00:28:43
time yes um and so does anyone want me to do it in a project other than apache spark
00:28:51
thank god okay cool um 'cause i'd but like as as it's it's a lot easier okay
00:28:57
i'm also i'm going to do this live live it's also going to go on the internet
00:29:03
which is filled with cheese okay um to to to to to do
00:29:11
still good everything's fine okay cool so hopefully this works and we're
00:29:16
gonna do a life full request review um okay cool and uh
00:29:23
that is i'm gonna go to my good friend sparked
00:29:26
dash t. r. start ups quite dot com um which
00:29:33
it is really high quality python code uh which will one day have to be
00:29:38
updated to python free um which is very fitting for the theme of this conference
00:29:44
uh i think realistically actually this will just get stuck behind increasingly destructive web proxy
00:29:49
uh and then we'll just never update any of the security settings on it ah
00:29:53
goatee anyways um so this gives us a view of the
00:29:57
pole requests coming in the spark a i've sorted it by um
00:30:01
whose commented on usable requests 'cause i'm looking for things where people aren't involved
00:30:06
uh but we could also go ahead in like say like hey what's
00:30:09
up i'm interested in seeing paul requests that are like relatively up to date
00:30:14
um and then we could see someone's just propose one
00:30:18
thousand four hundred line change to do adaptive executions parks equal
00:30:24
we'll probably not gonna get through that one in ten minutes but
00:30:27
it sounds really exciting um although okay this one this one here
00:30:32
new no commenters and convert file past your ryan binary
00:30:38
file data source i mean and if you type bar who runs this
00:30:42
doesn't look so bad maybe let's let's put this one um i actually did
00:30:47
if we look at some other people requests so if i have nothing to say about this one i got
00:30:51
some other ones and i can pretend i except for the fact i told you my plan um but yeah okay
00:30:57
so take a look at this okay um so here's here's a proposal on
00:31:01
um okay and i was tested with unit tests that's that's nice oh someone
00:31:07
did leave a comment on at seven minutes ago um well i mean it's
00:31:11
okay so uh yeah this this magic pile python code of predates get hard
00:31:18
uh having a pushy p. i. n. instead does pauling so it is a uh
00:31:25
sometimes like the other day but it's it's not it gives us gives us an egg
00:31:28
okay oh and uh ignore that thing about my job
00:31:33
that nine okay um so okay we we
00:31:38
change the tapes that we've asked them work here
00:31:43
where you want it to be just change to
00:31:45
whom are okay ah okay files can r. d. d.
00:31:53
okay so this this is still stream or not actually taking a job
00:31:56
for your i object uh i'm converting it inside of here that works
00:32:05
like that thing it's not necessarily that okay sure um a super characters and found
00:32:11
in the require your encoding okay this tells us why they were actually doing let's
00:32:16
um so that's that's cool and uh we now we can load files with
00:32:20
spaces very exciting um and frequently does not work and spark 'cause software sorry okay
00:32:27
ah so let's take a look at this unit test so
00:32:33
there are gonna make a fire oh they're gonna write one two three into a very exciting
00:32:38
um and then we're gonna re a ah we don't like close the file
00:32:45
and can you wonder if there's a risk your your book of
00:32:50
java file right i have to uh because that's
00:32:58
mm mm or is that like oracle and they don't want to depend on stack overflow it's
00:33:06
uh have a seven hats and the best java okay um
00:33:14
and to do okay this
00:33:19
weekend's half do i have to
00:33:22
i specify all organising 'cause i can't i don't see it
00:33:32
wow okay ah so we we we need if
00:33:42
it does
00:33:47
nice okay cool so it's great so that's fine okay that's good um
00:33:53
so that's so now we're gonna go ahead and use or read format
00:33:59
binary file load geared i get ah okay and so
00:34:04
or deer is um with tender and that's worked uh in
00:34:11
yeah okay so but it looks fine actually is so it's like it's a pretty fine what's ah but
00:34:19
ah yeah so i guess mm the assertion that so
00:34:24
it's a pentium uh we could assert that we actually get
00:34:27
these bites back as a as a totally that would be a good thing that we could do to improve that's ah
00:34:38
i mean i think um
00:34:43
uh
00:34:45
men yeah maybe it would be good to assert that
00:34:50
we get the bites back a a just in case
00:34:55
let's see here yeah uh it just in case it's um it's getting files with
00:35:03
and also like this isn't actually gonna oh is this can we
00:35:10
no this isn't actually gonna test what how i don't think this is gonna test
00:35:15
eh i'll okay cool j. okay i'm excited sorry i i'm having ah okay
00:35:22
cool cool so um the thing that we're we're testing is nominally that uh and
00:35:28
the the target passes it's it's fine but um
00:35:33
we we only have a space in the file name right um so let's go take a look at the uh
00:35:39
interface that we change which one did we change here at this
00:35:46
ah okay suspicion fire ah
00:35:51
okay yeah um so
00:35:54
here's what we do here ah on is it okay that uh we only
00:36:02
have the space of file name uh or or do we need it in half
00:36:09
or providing a two
00:36:14
trigger spark too big for um and so that's just like a
00:36:19
union i'm not sure and i could spend like a lot of time
00:36:24
tracing lacquered figure it out but i don't really want to you so i'm just gonna ask question
00:36:30
ah um and then i can clarify the like yeah no it's fine or it'll be like oh
00:36:35
yeah whoops all all either space um to with ten year which is would actually be really annoying
00:36:41
will have to be factor uh i you chills function here there to allow us to specify um
00:36:48
chapters uh to have spaces inside of it but okay yeah cool uh
00:36:52
does anyone else have anything else to one i suggest to you uh_huh
00:37:13
yeah that's that's a good point um yeah
00:37:20
oh sure okay so yeah sorry the point is um but
00:37:23
this is touching more than just a binary file format um maybe
00:37:29
just has to belong somewhere else i think would would be uh how i would phrase it um so
00:37:38
um yeah change seems to
00:37:43
uh impact not just a binary
00:37:47
file format need be visible on this e.
00:37:51
one uh or a route data source tests
00:37:59
ah what do you think as you can see very we easily um which is just
00:38:06
like how i could be wrong but uh okay uh anyone anyone else one uh uh_huh
00:38:15
ah okay at ah this is the downside of taking the freshest p. r. is uh well let's hit refresh and see what
00:38:23
happened so that the the comment was that they've already made a commit ah what did they change oh the i didn't assertion dammit
00:38:31
oh
00:38:34
still close so close okay so at this maybe illustrates that yeah maybe we
00:38:38
should wait for paul request which is like and how we're all so they're not
00:38:42
actively pushing changes to it right now uh
00:38:46
but we can use this part here are comments still a a reasonable and the
00:38:51
comment about like where the uh to should live is is also probably so reasonable
00:38:56
um and oh actually but this vision yeah well let's let's see here
00:39:04
um yeah well yeah i mean
00:39:09
we could maybe make this better by 'cause
00:39:14
we call first and we could instead assert that we're getting one entry back first uh and i would be
00:39:20
a little bit clearer if it fails but like yeah that would be kind of pedantic so well with it
00:39:30
oh probably um
00:39:34
yeah we could test for your eye
00:39:36
constructor use exceptions are bubbling now um
00:39:43
but i think it's it's
00:39:45
off
00:39:50
yeah that's good take a look at what uh what it says it's going to
00:39:53
do when it fails here that's a that's a good point um yeah okay um
00:40:02
did you dudes maybe come and it's uh here that we see
00:40:11
for past uh in violet and your high um optional
00:40:19
mm actually i'm gonna go ahead and then and say
00:40:23
it and that's just that's the like hastens like
00:40:28
for for me when i we've done ads trying to say like if you um
00:40:33
this is like a small change that we could make and if we didn't have any other changes to
00:40:36
make a whatever just market but since we probably should consider some of the other things here as well
00:40:42
maybe you could make this change if you felt like it um which is you know
00:40:48
once again very very wobbly but for purpose that's happening reviews happening in like a hundred
00:40:54
i wanna be as waffle yes i can't um or if i was working on a
00:40:57
project i would also wanna be pretty wobbly uh okay so i'm gonna go ahead and uh
00:41:05
finish finish this week to uh to do thanks for me that's
00:41:14
whatever and thanks for concerts a proposal um
00:41:20
good to support space it's uh
00:41:26
oh oh uh do we want to maybe is support for u. t. f. a.
00:41:32
it's um hand or have tests for supporting
00:41:37
a fancy things like me with a snowman
00:41:42
if snowman mode he uh if you have fun doing your job
00:41:50
why bother showing up um okay to to to us
00:41:59
okay
00:42:04
uh at least people from the stem and entertaining um uh okay and yeah oh it
00:42:11
sounds sounds good and we can see actually it laughter outdated overpass common already that's weird um
00:42:19
we can pretend that we were indirectly send five to the
00:42:23
universe to cause them to add an assertion so that's good um
00:42:31
cool i think i'm in your time right okay cool uh so
00:42:35
i would be remiss if i did not trying to you to buy more
00:42:38
things that you don't need a 'cause i am a currently living in america um
00:42:45
right so you two channel remember to like unsubscribe something something over here
00:42:50
so much a blog posts about good things you can do well you're doing
00:42:54
a poll request reviews on most of them are focused on how to do
00:42:57
good poll request reviews inside of your company on and so occasionally suggest things like
00:43:03
walking over and talking to someone and that might be challenging to do depending on your travel budget
00:43:10
and or the comparability of other people doing poll request reviews an airport lounge um
00:43:17
hi hey we would be remiss affected not trendy you buy my book again please buy this it is
00:43:23
unrelated but that should not stop you um and if you're interested in learning more about how to teach children
00:43:31
distributed computing site you should go to distribute computing for
00:43:34
kids dot com um this is actually not a joke
00:43:41
unlike everything else um and you can sign up for the mailing list your
00:43:45
email address and i will not spam you because you all are in europe
00:43:49
um and you have no well as um uh okay uh so i think
00:43:56
it's time for like food and drinking how do we have time for question
00:44:03
oh okay okay good before the question is there is a very important and that's it like we had three that's who
00:44:10
are can mean where uh going from the to the question is it does not the deer which is very important right
00:44:18
so that it is at six forty five sets has has said then sixty after is you know we
00:44:28
can take which if you use local so she calls it i was using it for the international people it's
00:44:36
called the olympic museum and if you don't shame on you know i i i just want to see
00:44:44
but yeah uh lately is of your hands uh uh to count how many people are going to these dinner
00:44:51
this is perfect and how many of you actually registered for dinner
00:44:55
so we're going to have problems with the food you comes so you can go there
00:45:01
but uh yeah we order probably least little this last then as many people as except the first
00:45:08
bus is going in fifteen and so if you want to have some questions you have time okay
00:45:12
it was a quite your rent to you
00:45:19
it's or here and i i
00:45:25
do do you need to register as a contributor somehow get in the community you
00:45:29
really start to nothing to but uh yeah uh that's a that's a great question um
00:45:34
so spark just does get up a request reviews so you you could just do and have a
00:45:40
there are some projects where they use separate review tools and you have to
00:45:44
register to use them i think go is the one which comes to mind um
00:45:50
it it doesn't hurt to email they'd like to have list and say like high without
00:45:55
a first just so that like people know you're round um by e. certainly don't have to
00:46:03
you try to you but you discovered show a review in which
00:46:08
one one that what is it about contributing as a way to work
00:46:12
i think it's okay i've certainly learned many programming languages well
00:46:15
contributing code uh if you wanna see really bats colour code
00:46:19
uh you can go look at my early spark a stellar code actually wanna see moderately bats colour
00:46:24
code you can look at my current sparks delicate um i i think i think contributing is a
00:46:30
great way to learn language doing is is probably a very good with one language i think i
00:46:36
think reviewing is harder to learn a new language though um but i mean everyone's learning habits are different
00:46:45
um and the show when it comes to think one called the deal it decide
00:46:49
that that's all on this uh need to use it should be tolerated me and what
00:46:53
do you do then that's that's a really lovely question i 'cause that does happen
00:46:57
from time to time um sometimes you'll be like uh well we should we write this
00:47:02
and we should we write the same whoa maybe maybe you would just be
00:47:05
faster if i get that's and encouraged not to give in to that um
00:47:09
have it um it indeed might be faster to just we write it all
00:47:13
yourself um and if you need to go really sell or it's a security issue
00:47:19
or something that's very very important okay use use your judgement um
00:47:23
but i i think the art of of working with someone else you can you can treat it like uh you know we like working with an intern
00:47:30
if you just re write all the code that in in turn gives you um you're
00:47:34
not going to learn how to teach people and also the intern is not going to learn
00:47:38
how to do this to us not that random strangers on an outer interns but i i
00:47:42
find i'm having that mindset helps me like we're we're both there to learn from each other
00:47:48
um but at some point yeah you just go like oh
00:47:52
thank you for signing this issue and providing a proof of concept
00:47:56
how do you feel it may be we take it from here and that's that those are in situations where it's like we really
00:48:02
need this in this release or or like i really need this
00:48:06
for what i'm doing and i can't wait for for the other person
00:48:11
uh but i i try not to i'm starting to do curved in the other could this is where you're
00:48:19
not c. d. that's in the fishman to the yeah
00:48:22
the the concert and maternity did not because we're all the
00:48:27
someone external uh would start doing 'cause that's that's a really great question so i
00:48:32
personally have not dealt with that um i definitely done like with using apache airflow
00:48:37
and random projects where it's just like i it's mated one for requests and was fixing like doc start m. d.
00:48:44
um and people were like hey what's up welcome i'm part of
00:48:48
that is trying to pick places that are friendly and welcoming to newcomers
00:48:52
as you find push back from like trying to help people deserve used it's not worth
00:48:56
it just you know there are many many open source project that get up contacts something else
00:49:03
if you don't think that uh should be uh um question of cool
00:49:09
as a um don't you think there should be some kind of mentoring um
00:49:15
well at least some kind of a fair and people that you should be as as it
00:49:22
we view uh uh well eighty two in order to i mean most of the projects
00:49:29
uh uh well i started reviewing code uh
00:49:32
one point one the one the shoe use stock
00:49:37
yeah and and as a review uh it's it's
00:49:39
frustrating as well because you you think it's good that
00:49:43
you don't have the power to do anything else so most of the time i try to find like the
00:49:48
mm most usual people that in these parts
00:49:52
of code he's not h. knowledgeable enough to merit
00:49:56
should nasa quest but they have never seen a uh an open source project well these kind of
00:50:02
less possibilities where precisely the yeah
00:50:08
scene four focal commuters i i agree um
00:50:12
in so in apache uh uh suffer foundation there's the guide on mentoring and the guide is to do
00:50:19
um maybe someday someone will write it uh but i i think you one of the
00:50:25
benefits that you you can provide as someone who's reviewing code and and doing a first pass
00:50:29
used to reap menu credibility to this other person if you know if you really think it's
00:50:33
good then you go and you find the people who commit things and you can do this with
00:50:38
get blamed get logon looking like get up to see who was coming to uh and you you ping them and you try and help
00:50:45
get this on stock um and it that is a lot
00:50:49
of work unfortunately um for one of one of the things that
00:50:54
in in this part project and it's not it's certainly not perfect park has a lot of um
00:51:00
opportunities for improvement in our community management uh
00:51:05
mm thank you yes uh but one of the things that that that i find useful
00:51:09
is that we can see here are uh there's these little tiny pictures of the people who
00:51:13
have reviewed the hours uh and not all of those people are commuters um but as myself
00:51:19
as a commander there are people who are crossed and when i see them as a commenter
00:51:23
with the little green blocks surrounding name which means they said yeah yeah uh i go when i click on
00:51:29
it and because i know that it's not gonna be as much work for me to come after code right like
00:51:35
it is gonna be fine like i'll i'll do a second pass of course because
00:51:39
when they want to be able to nominate this other person to also become a commuter
00:51:43
uh so i can just say yeah just commit everything is an me but it's it's
00:51:47
uh informal structure i would say in the sparc project and i think in many other projects
00:51:52
uh there's informal structures exist but certainly um a
00:51:56
humour structures formal could be greatly beneficial but i
00:51:59
don't know anyone has bothered uh 'cause uh uh people are a lot harder than soft for i. e.
00:52:07
okay cool so no the question right now it's time to go give it
Conference Program
A Tour of Scala 3
Martin Odersky, Professor EPFL, Co-founder Lightbend
June 11, 2019 · 5:15 p.m.
8337 views
A story of unification: from Apache Spark to MLflow
Reynold Xin, Databricks
June 12, 2019 · 9:15 a.m.
1267 views
In Types We Trust
Bill Venners, Artima, Inc
June 12, 2019 · 10:15 a.m.
1569 views
Creating Native iOS and Android Apps in Scala without tears
Zahari Dichev, Bullet.io
June 12, 2019 · 10:16 a.m.
2232 views
Techniques for Teaching Scala
Noel Welsh, Inner Product and Underscore
June 12, 2019 · 10:17 a.m.
1296 views
Future-proofing Scala: the TASTY intermediate representation
Guillaume Martres, student at EPFL
June 12, 2019 · 10:18 a.m.
1157 views
Metals: rich code editing for Scala in VS Code, Vim, Emacs and beyond
Ólafur Páll Geirsson, Scala Center
June 12, 2019 · 11:15 a.m.
4695 views
Akka Streams to the Extreme
Heiko Seeberger, independent consultant
June 12, 2019 · 11:16 a.m.
1552 views
Scala First: Lessons from 3 student generations
Bjorn Regnell, Lund Univ., Sweden.
June 12, 2019 · 11:17 a.m.
577 views
Cellular Automata: How to become an artist with a few lines
Maciej Gorywoda, Wire, Berlin
June 12, 2019 · 11:18 a.m.
386 views
Why Netflix ❤'s Scala for Machine Learning
Jeremy Smith & Aish, Netflix
June 12, 2019 · 12:15 p.m.
5026 views
Massively Parallel Distributed Scala Compilation... And You!
Stu Hood, Twitter
June 12, 2019 · 12:16 p.m.
958 views
Polymorphism in Scala
Petra Bierleutgeb
June 12, 2019 · 12:17 p.m.
1113 views
sbt core concepts
Eugene Yokota, Scala Team at Lightbend
June 12, 2019 · 12:18 p.m.
1656 views
Double your performance: Scala's missing optimizing compiler
Li Haoyi, author Ammonite, Mill, FastParse, uPickle, and many more.
June 12, 2019 · 2:30 p.m.
837 views
Making Our Future Better
Viktor Klang, Lightbend
June 12, 2019 · 2:31 p.m.
1682 views
Testing in the postapocalyptic future
Daniel Westheide, INNOQ
June 12, 2019 · 2:32 p.m.
498 views
Context Buddy: the tool that knows your code better than you
Krzysztof Romanowski, sphere.it conference
June 12, 2019 · 2:33 p.m.
394 views
The Shape(less) of Type Class Derivation in Scala 3
Miles Sabin, Underscore Consulting
June 12, 2019 · 3:30 p.m.
2321 views
Refactor all the things!
Daniela Sfregola, organizer of the London Scala User Group meetup
June 12, 2019 · 3:31 p.m.
514 views
Integrating Developer Experiences - Build Server Protocol
Justin Kaeser, IntelliJ Scala
June 12, 2019 · 3:32 p.m.
551 views
Managing an Akka Cluster on Kubernetes
Markus Jura, MOIA
June 12, 2019 · 3:33 p.m.
735 views
Serverless Scala - Functions as SuperDuperMicroServices
Josh Suereth, Donna Malayeri & James Ward, Author of Scala In Depth; Google ; Google
June 12, 2019 · 4:45 p.m.
936 views
How are we going to migrate to Scala 3.0, aka Dotty?
Lukas Rytz, Lightbend
June 12, 2019 · 4:46 p.m.
709 views
Concurrent programming in 2019: Akka, Monix or ZIO?
Adam Warski, co-founders of SoftwareMill
June 12, 2019 · 4:47 p.m.
1974 views
ScalaJS and Typescript: an unlikely romance
Jeremy Hughes, Lightbend
June 12, 2019 · 4:48 p.m.
1377 views
Pure Functional Database Programming‚ without JDBC
Rob Norris
June 12, 2019 · 5:45 p.m.
6374 views
Why you need to be reviewing open source code
Gris Cuevas Zambrano & Holden Karau, Google Cloud;
June 12, 2019 · 5:46 p.m.
484 views
Develop seamless web services with Mu
Oli Makhasoeva, 47 Degrees
June 12, 2019 · 5:47 p.m.
785 views
Implementing the Scala 2.13 collections
Stefan Zeiger, Lightbend
June 12, 2019 · 5:48 p.m.
811 views
Introduction to day 2
June 13, 2019 · 9:10 a.m.
250 views
Sustaining open source digital infrastructure
Bogdan Vasilescu, Assistant Professor at Carnegie Mellon University's School of Computer Science, USA
June 13, 2019 · 9:16 a.m.
375 views
Building a Better Scala Community
Kelley Robinson, Developer Evangelist at Twilio
June 13, 2019 · 10:15 a.m.
245 views
Run Scala Faster with GraalVM on any Platform
Vojin Jovanovic, Oracle
June 13, 2019 · 10:16 a.m.
1342 views
ScalaClean - full program static analysis at scale
Rory Graves
June 13, 2019 · 10:17 a.m.
463 views
Flare & Lantern: Accelerators for Spark and Deep Learning
Tiark Rompf, Assistant Professor at Purdue University
June 13, 2019 · 10:18 a.m.
380 views
Metaprogramming in Dotty
Nicolas Stucki, Ph.D. student at LAMP
June 13, 2019 · 11:15 a.m.
1250 views
Fast, Simple Concurrency with Scala Native
Richard Whaling, data engineer based in Chicago
June 13, 2019 · 11:16 a.m.
624 views
Pick your number type with Spire
Denis Rosset, postdoctoral researcher at Perimeter Institute
June 13, 2019 · 11:17 a.m.
245 views
Scala.js and WebAssembly, a tale of the dangers of the sea
Sébastien Doeraene, Executive director of the Scala Center
June 13, 2019 · 11:18 a.m.
661 views
Performance tuning Twitter services with Graal and ML
Chris Thalinger, Twitter
June 13, 2019 · 12:15 p.m.
2003 views
Supporting the Scala Ecosystem: Stories from the Line
Justin Pihony, Lightbend
June 13, 2019 · 12:16 p.m.
163 views
Compiling to preserve our privacy
Manohar Jonnalagedda and Jakob Odersky, Inpher
June 13, 2019 · 12:17 p.m.
301 views
Building Scala with Bazel
Natan Silnitsky, wix.com
June 13, 2019 · 12:18 p.m.
565 views
Diversity and Inclusion Panel: Bring the Thunder!
Panel
June 13, 2019 · 2:30 p.m.
244 views
Asynchronous streams in direct style with and without macros
Philipp Haller, KTH Royal Institute of Technology in Stockholm
June 13, 2019 · 3:45 p.m.
304 views
Interactive Computing with Jupyter and Almond
Sören Brunk, USU Software AG
June 13, 2019 · 3:46 p.m.
681 views
Scala best practices I wish someone'd told me about
Nicolas Rinaudo, CTO of Besedo
June 13, 2019 · 3:47 p.m.
2706 views
High performance Privacy By Design using Matryoshka & Spark
Wiem Zine El Abidine and Olivier Girardot, Scala Backend Developer at MOIA / co-founder of Lateral Thoughts
June 13, 2019 · 3:48 p.m.
754 views
Immutable Sequential Maps – Keeping order while hashed
Odd Möller
June 13, 2019 · 4:45 p.m.
277 views
All the fancy things flexible dependency management can do
Alexandre Archambault, engineer at the Scala Center
June 13, 2019 · 4:46 p.m.
389 views
ScalaWebTest - integration testing made easy
Dani Rey, Unic AG
June 13, 2019 · 4:47 p.m.
468 views
Mellite: An Integrated Development Environment for Sound
Hanns Holger Rutz, Institute of Electronic Music and Acoustics (IEM), Graz
June 13, 2019 · 4:48 p.m.
213 views
Closing panel
Panel
June 13, 2019 · 5:54 p.m.
400 views
