Player is loading...
Integrating Developer Experiences - Build Server Protocol
Justin Kaeser, IntelliJ Scala
Wednesday, June 12, 2019 · 3:32 p.m. · 43m 05s
IDEs - Integrated Development Environments - traditionally provide out of the box support for many of the tasks that go into making working software out of source code. But increasingly, developers expect to be able to use any one of a variety of special-purpose tools for each task. This shifts the focus of the IDE from "Integrated" to "Integrating" external tools into a coherent experience. I have been focusing on one aspect of this - build tools - and will talk about challenges integrating tools and how the Build Server Protocol makes it possible for IntelliJ to interface with any build tool.
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
invoke compilers by hand we would manually run tar
00:00:05
and a bunch of compilation products just to package done
00:00:09
and maybe more sophisticated engineers word and where they're
00:00:14
dependencies and bill grass into a make file maybe
00:00:19
some of us to do that but personally i find that to be still too much manual labour
00:00:26
so along came the i. d. e. e. uh to offer an
00:00:31
interface for the whole software development life cycle from editing formatting re factoring
00:00:39
and giving a structured to your whole code base too structured into projects and modules
00:00:46
and on those projects somebody owes you could build and run and test
00:00:51
and debug and package and deploy all at the click of a button
00:00:56
uh so buttons are great but uh some people pray for key presses
00:01:06
but anyway so what does that have to do with me ah and
00:01:10
my talk and just and i work on the uh on an i. d.
00:01:16
uh the intelligence colour plugin
00:01:19
and i actually most of my team is here at this conference you
00:01:25
can meet us at our at the checkpoints booth and talk to us there
00:01:31
ah i already know most people probably use intelligent a for
00:01:36
this colour editing uh did you also know that we are
00:01:41
open source mostly and you can find some get up just please raise
00:01:44
your hand if you did okay that's it's a growing number to actually
00:01:50
new this net ah so the purpose of this talk is to
00:01:57
ah and give you another approach to integration so integrating another i. d.
00:02:02
e. integrating developer experiences uh and specifically
00:02:07
i'm talking about the build server protocol
00:02:10
and some of the implementation details
00:02:16
or design details of this protocol
00:02:20
which is supported now by intelligent a as a client are
00:02:25
by loop as a server also by metal so the client all of
00:02:29
us talking earlier about it but more to and what well that is later
00:02:36
so we're coming from a place where we have uh integrated
00:02:41
development environment where everything is in uh interface in front of you
00:02:47
but that doesn't work for everybody now all the work flows
00:02:53
that you'd like to have are supported by the i. d. so
00:02:58
yeah one size does not fit all so so we have uh we re
00:03:01
proliferation of special purpose tools such as
00:03:05
dependency managers build tools um test frameworks
00:03:12
and so on ah m. and they have the advantage of being to some
00:03:19
degree cross platform on they offer you a command line interface which is programmable
00:03:25
to some degree again and come and an override a count of contexts mostly
00:03:30
and uh that machine on a c. i. pipeline on different types of that machines
00:03:38
um and that's great but what about my a
00:03:44
general purpose development front and i'd still like to
00:03:47
keep using intelligent and like you to keep using
00:03:50
intelligent even when you speak specific tools for specific tasks
00:03:56
but there's some integration issues so many of these tools that really have matching interfaces
00:04:04
and i mean just look at these faces they they they don't really fit together right
00:04:09
yeah so so you you're gonna try to you gonna have a bill tool like s. p. t. which has
00:04:17
one way of interfacing with that you can dumb stuff to the
00:04:23
shallow or to some files and then you have to write a
00:04:26
special purpose integration in intelligent a or in lupin your compile server
00:04:33
and that's a lot of work a string things together
00:04:38
so i just as an example are the
00:04:43
s. p. t. integration in intelligent a involves a
00:04:48
a special purpose plugin or two or three for several purposes
00:04:52
uh that we inject over some special s. p. t.
00:04:56
mechanism which sometimes breaks because i don't know why uh
00:05:03
some old s. p. t. code doesn't do the caching write something
00:05:08
and and then some special code in intelligent a to map this
00:05:13
structure that if extracted and them to an x. m. l. file
00:05:17
and just read immediately again and then try to map the s. p. t. project
00:05:21
structure to be intelligent project structure which works most of the time but not always
00:05:28
um
00:05:30
yeah i uh another plug in just for show integration a ah
00:05:34
and that doesn't even give us nice compile output an intelligent a
00:05:40
and and we have this problem for a variety of tools
00:05:43
like for each test framework we have special purpose code for
00:05:46
each bill tool for our various libraries we have special support
00:05:51
like our cup play uh for for matters like stylised empty
00:05:57
and your other projects like group have the same integration issue
00:06:05
oh yeah i mentioned bill tools ah we have lots of choices in this space especially
00:06:13
in scholar nobody seems to be entirely happy with little tool so they create a new one
00:06:20
and i'd have to duplicate the s. p. t. a. uh that uh it s.
00:06:27
s. p. t. effort for every one of them if i wanted to support them
00:06:32
and
00:06:33
for the more we have yeah the little coherence between these tools
00:06:39
uh i'm between the interface so you
00:06:42
might have context switches terminal id id terminal
00:06:47
and that does interrupt your work flow a bit and i still wanna just work from intelligent and maybe not
00:06:55
switch around too much and learn different command line interfaces as well
00:07:01
say we we have this haphazard tangle of tools uh
00:07:06
and we have haphazard integrations the like sort of seem to keep stuff to get there
00:07:13
but you don't really know where less than your right arm is so that that's all not very very coherent
00:07:20
so i've been trying to to uh type these loose threads of
00:07:26
tooling together with other tooling developers to create a more integrated development experience
00:07:35
uh but first some prior art you you might have heard have you heard of a language server protocol
00:07:42
yeah that's most of you that's that's very sophisticated audience today
00:07:48
um yeah this address is one
00:07:51
major integration bottleneck between editors and compilers
00:07:57
so by offering a standard way to communicate between them and i offer a
00:08:04
specific actions and reporting errors and warnings and so on
00:08:09
this works pretty well for editor integrations
00:08:14
um but and it has been been very successful so much so
00:08:20
that people keep asking us if we're gonna support it in intelligent a
00:08:25
the the answers it'd probably not
00:08:28
because of various challenges
00:08:32
so one of them is that the l. s. p. doesn't really model the structure
00:08:37
of the co pays it just tells you you got a bunch of files here um
00:08:41
this this is your workspace within these files um you'll have a
00:08:46
you might have some errors warnings going on uh you can have certain
00:08:51
actions on points in the code but it doesn't give us the it
00:08:58
the model of the whole code base that we need an intelligent for some of the advanced features
00:09:05
or alternatively it would require each server to support and uh even though we have
00:09:13
we can support them relatively easily with our own infrastructure
00:09:17
and there's no they're a mapping from r. l. s. p. which just gives you
00:09:22
messages and actions too intelligent a which
00:09:25
has a whole syntax three that represents our
00:09:30
also errors so represents the syntax three of the
00:09:35
code and allows a re factoring actions on it
00:09:43
and well some other things like there is maybe no common cons of
00:09:48
building testing running to some degree that is fine you just save and the
00:09:53
server knows okay need to recompile the maybe like more fine grained support for that
00:10:01
okay i mention lots of bill tools so uh what if a new bill
00:10:05
to emerges from the desert sands such
00:10:09
as fury which john pretty is probably uh
00:10:14
talking about right now or later or earlier
00:10:19
anyway he he's here and talking about it um so fury doesn't really have
00:10:24
users yet so i can't quite get
00:10:28
justify spending several weeks on supporting yeah
00:10:34
a given that it would also increase the complexity of our code base
00:10:41
perry quite a bit for if i worked on this for each bill tool
00:10:47
and uh even s. p. t. is still an ongoing effort to support correctly
00:10:52
because there's always new age cases to be found where things don't work right together
00:11:00
but i wanna i wanna do this anyway right uh and adjusts opening here
00:11:07
intelligent the import project i i got fury here which is of
00:11:13
course a project defined in fury and widget wanna just import it
00:11:22
and let's see if this works
00:11:28
it up over on my screen
00:11:31
yep so i i just important fury
00:11:34
into into the j.
00:11:37
and you see here oh oh the whole structure now the the the
00:11:41
two modules to record few rebuild fury has this concept of source dependencies ah
00:11:48
the right and so you see all the source dependencies nap to a separate module here and uh i don't know
00:11:56
oops
00:11:58
f. o. open a file
00:12:03
oh
00:12:05
and
00:12:07
i have a hard time you in this okay but all the stuff seems to work like syntax highlighting or
00:12:16
navigating to sources and so for their available
00:12:24
so yeah how how did i do this
00:12:28
i i i did not if implement fury support in p. s.
00:12:33
p. uh i know i did not implement fury support in intelligent a
00:12:38
but what i did do is implemented d. s. p. servers support into fury
00:12:45
and i implemented a b. s. p. clients afford into intelligent so
00:12:52
neither of these know about the other so there's no special
00:12:55
purpose integration between them but they work together to some degree already
00:13:03
um
00:13:06
so
00:13:11
by the way of fury is also a a p.
00:13:15
s. p. client because it uses and loop and which
00:13:22
i don't know who is talking about look at this conference
00:13:26
or how he's not talking but he's here you can talk to him about look uh
00:13:32
anyway uh bloopers or compile server and which
00:13:37
supports various bill tools from maya export to loop
00:13:42
but fury is not special in that it uses loop as its compilation back and
00:13:48
so it doesn't have a compiler integrated like most bill tools
00:13:54
so um yeah i i was kind of inspired by a crowd of
00:14:00
supposedly and rambling cone give me six hours developed three
00:14:04
and i will spend the first for sharpening my x.
00:14:07
and uh well i i thought yeah i'd give me one month to support a bill
00:14:12
tool and i will spend the first year developing a protocol s. so here we are
00:14:19
the build server protocol
00:14:23
so this is um very much
00:14:26
inspired by ah oriented
00:14:30
to be compatible to compliment ellis p. language server protocol
00:14:37
uh it was originally wrote up as a draft by all off from this colour
00:14:42
centre he earlier talked about um metals metals
00:14:46
also uses the s. p. as it's uh
00:14:51
yeah the ford score compilation backend
00:14:56
and uh so yeah like a list he integrates editors and
00:15:04
compilers language servers uh p. s. p. integrates build tools
00:15:10
and i. d.s or language servers as in the case of metals which
00:15:14
is uh my language server so in the case of metals you have
00:15:17
a three step process you go from bill tool to language server to
00:15:20
editor a wit and intelligence aegis interface between bill tool and i. d.
00:15:28
so like alice p. o. abused he has the concept of a project structure which
00:15:35
it yeah abstracts interview generic concepts which i'll talk
00:15:39
about right away and a few common operations that
00:15:43
like most still tools will having common searches build or even test
00:15:51
and they have some defined a request reply semantics
00:15:59
so some
00:16:01
base concepts of the s. p.
00:16:05
it is technically appliance server protocol but you
00:16:09
can also send directions yeah you can send messages
00:16:13
in both directions either from the server to the client or from the client to the server
00:16:17
uh they can be messages which don't expect a
00:16:21
reply or typically requests which do expect a reply
00:16:26
and most of the time it it'll be the client
00:16:29
sending the request and the server replying or sending notifications
00:16:36
ah one edition over l. s. p. that we
00:16:40
have is the server discovery protocol so we'd like
00:16:46
servers inclines to be able to talk to each other without
00:16:50
implementing any client or server specific support unless we really want to
00:16:55
support this extensions for a specific tool so what we did is
00:17:02
we defined a few system specific locations where a bill to make place configuration files
00:17:10
uh and the client knows where to look for these configuration files and
00:17:16
reads information about how to start a build server so we wouldn't need any
00:17:24
ah ports or
00:17:27
when the need to configure any ports or i. p. addresses or anything it's all in the file
00:17:34
and the server is then simply uh
00:17:38
an executable
00:17:40
which communicates over standard input standard output
00:17:46
so this is very portable so we
00:17:49
don't need any assistance specific uh communication protocols
00:17:57
and the life cycle of that once the servers start it looks a bit like this
00:18:02
um the client will send start the server
00:18:05
and then send a build initialise request the server
00:18:10
well answer and will eventually give a initialised um message notification
00:18:20
and then all the interesting stuff happens and when they're done
00:18:23
the client request to the server to shut down and exit
00:18:30
now the interesting part is uh one of the interesting parts is the definition of the build structure
00:18:39
so most bill tools have some concept of projects and
00:18:44
modules i. n. b. s. p. we call that targets
00:18:50
uh our target comprises um set of source files that
00:18:55
works on ah it has some um capabilities so you
00:19:02
like can tester can compile can run some stuff you can do with this target and
00:19:10
it will give you some meta data uh which can be language specific so force colour
00:19:14
it could be the sky library that is being used and the or maybe julianne moore depending
00:19:20
and in the cases colours well it won't know in
00:19:24
the general case it will also have dependencies on other targets
00:19:30
and get any target can be associated with a set of
00:19:34
sources likewise any source can be associated with a number of targets
00:19:47
now on these targets you can do some predefined
00:19:50
actions the most common one will probably be compilation ah
00:19:56
and another justly you may some targets may support testing
00:20:01
or even running if it will tool is so inclined
00:20:09
now you see here
00:20:13
after a a compilation request for some targets
00:20:16
you may get numbers of diagnostics in progress
00:20:21
it's a diagnostics is the word for
00:20:27
messages relating to your source code so such
00:20:31
as ah errors warnings um anything else going on
00:20:38
at a specific code location
00:20:41
these are exactly an allergist to l. s. p. diagnostics
00:20:46
so they can be forwarded by a a language server
00:20:51
as well as p. diagnostics
00:20:55
oh
00:20:57
and yeah
00:21:01
another thing we have is a progress notifications tasks so
00:21:08
since of build server might have some long running tasks
00:21:11
such as compilation will for longer your projects take awhile
00:21:17
um we want to give progress updates so for that we have
00:21:22
to find a specific notifications task start past progress and task finish
00:21:32
and i'll show you have that looks in a minute
00:21:37
or about right now so i'll close fury here
00:21:44
and opened occur which i imported as a loop project just before
00:21:55
and i'll try to run a build on it
00:22:01
and that didn't quite work out because of some bug or other uh just to refresh
00:22:10
oh yeah ours i love those
00:22:14
okay now it should work
00:22:17
so you see here that there is ongoing progress notifications
00:22:21
about the individual targets within the our car project and
00:22:26
you see and uh progress thirty five forty percent fifty and so on
00:22:32
and how long they take so with this this can help you to actually
00:22:37
get an overview of which part of your
00:22:39
compilation is eating all the time which modules
00:22:45
or taking longer than all the others so we so here actors
00:22:49
to twenty three seconds to compile while others here uh uh something
00:22:57
a cowboys t. i. was not even a second
00:23:04
and in the same window you'll get a whole bunch
00:23:06
of yet the diagnostics so warnings which you can navigate
00:23:11
to directly so i think this is an him prove
00:23:14
meant over the existing integration we have four s. p. t.
00:23:25
so this this is basically already uh released feature you can use it
00:23:30
pretty well with blue put our it works with fury at least for import
00:23:35
uh after currently got an intern working on supporting it for no
00:23:41
uh i hope people are inspired to uh write
00:23:47
these p. support for other bill tools as well because
00:23:52
that will yeah enable integration with intelligent a and medals at the same time so
00:23:59
what more could you want
00:24:03
um some
00:24:07
sons intelligent a specific challenges i wanted to talk about
00:24:11
so mapping the loop and no mapping the d. s. p.
00:24:18
project structure too intelligent a
00:24:22
here's got another step in there because we have on the top we have s. p. t.
00:24:29
so it's p. t. has projects which have multiple scopes
00:24:34
uh most commonly you'll be using testing compile some people also use integration test or
00:24:41
something that they made up a but that doesn't really work so well so i at least four come
00:24:48
for importing intelligence so it's more a commendable to use separate projects for that
00:24:54
a a speech he can has approached concept of salsa shared between
00:25:01
projects
00:25:04
so that we can map straightforwardly onto
00:25:08
b. s. p. uh huh that's also how the structure looks in the loop which was built basically
00:25:17
to be compatible with p. s. p. directly
00:25:22
so if you install the blue plugin in s. p. t. and when the blue
00:25:27
install it will create some blue configuration files which basically match the p. s. p. structure
00:25:34
but in these p. we don't have scopes so we
00:25:38
map the uh compile and the test scope to separate targets
00:25:45
now once we start importing this into intelligent way
00:25:49
um what we currently do is we get these compile and test
00:25:53
targets and match them together to go back to a single intelligent module
00:25:59
that's
00:26:01
maybe not ideal well we're currently reconsidering if we want to happen to separate modules but
00:26:08
in tele j. does support compile and test skills but no other stops so
00:26:13
that's one reason it doesn't work well with integration tests go up in s. p. t. another challenge is
00:26:21
a shared sources intelligent for technical reasons doesn't support sharing
00:26:26
sources between models you have a one to one correspondence
00:26:30
so what we do instead is to create synthetic modules
00:26:35
that are depended depended upon by the modules that share the source
00:26:41
yeah
00:26:44
i shared sources i'm i'm with that i mean source
00:26:48
to shared between different modules or targets so so if
00:26:56
ooh it's pretty common so um
00:27:00
for example if you want to share sources between different versions
00:27:06
of a compilation so if you are compiling force colour g.
00:27:09
s. and uh it's got a native and j. v. and
00:27:13
then shortly wisdom can you ask your question into the microphone piece because everything
00:27:18
was recorded and if we don't speak into the mike no recording thank you
00:27:24
i i hope my answer had been of context for the question to be recorded right um
00:27:32
so yeah those are shared sources are you might also share
00:27:37
sources between different when you compile against different versions of skylark
00:27:43
and i think there's other reasons to like compiling certain tests
00:27:51
and different types of tests against the same sources maybe but usually you'd model that as a dependencies
00:27:59
right so intelligent doesn't support the shared sources between modules so
00:28:03
we map it to us pacific uh to a synthetic module
00:28:09
and that works as long as these
00:28:13
modules don't have conflicting dependencies which is
00:28:17
often the case so uh you might get rid
00:28:20
code where in intelligent where that doesn't happen in
00:28:26
uh the bill tool itself
00:28:29
that's unfortunate but right now it's what we got
00:28:36
so i've already showed you how a compile looks when
00:28:41
run from my obvious p. this is just a backup slide
00:28:47
and
00:28:49
that brings me ready to the
00:28:52
these p. road map ah so any questions so far
00:29:01
okay so the version two zero of the d. s.
00:29:06
p. protocol is more less around the corner how we're still
00:29:14
figuring out some details and need to coordinate with a blue
00:29:18
and metals to get the these minor breaking changes up to date
00:29:28
but other than that it it's already perfectly
00:29:30
viable to write support for specific build tools
00:29:36
uh as i said in tele jade loop on the fury
00:29:41
support this and we're coordinating to make them work
00:29:46
together on the same version uh once we got get
00:29:49
out version two zero we intend to have it
00:29:51
a stable for the forseeable future so only compatible changes
00:29:59
more build tools are already being worked on specifically mail
00:30:02
i think somebody was already ah so looking it pays off
00:30:07
and within intelligent they like the support is the love it it's body so
00:30:13
so we might want to support more languages have better java specific support right
00:30:18
now it's only this colour and it's the java is a subset of scholar
00:30:26
uh and some robustness and mapping projects
00:30:30
and i'm sure once you start using it you'll figure out find out lots of problems with that that our then six
00:30:40
ah going beyond that
00:30:44
tooling protocols it's uh
00:30:51
when we might want to think about if we can replicate the
00:30:56
l. s. p. success for more uh tooling specific tasks to integrate tooling
00:31:05
um so some of them might fall into the scope of existing
00:31:11
protocols and searches remote building up p. s. p. is currently not
00:31:17
dedicated to the remote but it's made to be a extensible enough that it would work with remote servers um
00:31:25
re support more fine grained support for testing so n. b.
00:31:29
s. p. we only have a basic support test this target but
00:31:35
if you want to support specific or different test frameworks in a
00:31:39
generic way you could think about how a protocol for that looks
00:31:46
um
00:31:48
challenges that are of course that sometimes tools are
00:31:54
not generic enough or don't share enough common surface area
00:32:00
to make this standardisation necessary or useful
00:32:08
and of course uh adoption so i'm talking to
00:32:11
you here in hopes of getting more adoption for
00:32:16
is he to make it actually useful because as long as it's otherwise it's just some that code sitting there
00:32:23
out of interest does anyone already use uh abuse p.
00:32:27
in some form like loop with intelligent a or with metals
00:32:32
right yeah that it's a few but i hope more if you will uh get to try it out soon
00:32:41
now that brings me to the end of the park as such
00:32:47
i'd very much welcome uh any discussion or questions or comments by you
00:32:55
yeah you do it
00:33:04
well thanks for the great introduction moral review obviously on
00:33:09
could you go back to the all the point you're making out the all the
00:33:12
input unhelpful was going to be basically standard and then standard i'll so desert me
00:33:18
can you live in dallas p. everything is basically all like a socket communication
00:33:24
you can just um some jeez on on this in the batteries all right
00:33:27
right on the it's the same with b. s. p.
00:33:31
and in principle you could use any communication protocol you want
00:33:35
but with a connection protocol specifically we defined the all the communication goes over
00:33:41
standard in standard out it will still be a jason's r. p. c. messages
00:33:46
that just get printed to standard instead of a
00:33:49
of a dedicated process so this process can be
00:33:54
part of the regular build tool are like you were
00:33:58
like with fury run and fury b. s. p. and then it
00:34:02
just waits for messages from standard in in response to stand about
00:34:07
with these days not busy messages uh or it could be a separate uh executable that
00:34:16
handles this are indirectly so connects to a remote server for instance and
00:34:23
buy a t. c. p. i. p. or anything really but this way
00:34:27
we have a standardised way to communicate with the t. s. p. server
00:34:34
so so does that mean the intelligent will always be dull on the start stop processes initially
00:34:40
right all yeah cause a loss p. e. you can basically
00:34:44
do the handshake after both all programs or come up and running
00:34:49
yeah um so in this case you just have a
00:34:53
you you could have a residence server for b. s. p.
00:34:56
and in this case you just have a separate executable that just connects to this resident process and just
00:35:02
pushes the messages to that standard out interest okay cool thank you
00:35:07
okay any other questions yes
00:35:12
with the uh yeah my question is for example metal c.
00:35:19
uses up and the d. s. p. two contact with blood
00:35:25
then again you another i. d. e. that also uses
00:35:28
metal an editor that supports metal that also wants to use
00:35:32
yeah yeah the should be used it to metals or should directly
00:35:39
and then the billboard b. s. p. supports
00:35:43
not till tools working on the same code base
00:35:47
contact it at the same time it's not possible on principle
00:35:51
sure that that depends on the little tool if you allow concurrent
00:35:56
access but you can have multiple b. s. p. servers uh
00:36:00
connecting to the same bill to or running on the same workspace
00:36:04
oh okay yeah it it's only a matter of the bill tool itself allows
00:36:10
concurrent access as or or or if it will get confused and
00:36:15
break caching when when you talk about the bill till you mean
00:36:18
below or you mean is big deal i'm a little tool in general
00:36:23
uh whatever you use um so to your first part of the question uh
00:36:27
so intelligent they can use loop as it's built service build server protocol server
00:36:35
and in which case you import directly from loop um no loop has its own
00:36:41
configuration file which you could write by hand but the more typical use cases that you
00:36:47
export and existing project from s. p. t. or mail or baseball or may even
00:36:53
to the blue configuration format and then you imported into intelligent via b. s. p.
00:37:02
so and and that's basically also uh metals does
00:37:07
oh well it's possible to work both with metals and then something and then let
00:37:13
it or that uses the l. the t. b. s. p. at the same thing
00:37:18
yeah in principle there's no problem with that okay thank you
00:37:25
a big hello huh thank you i'm i'm single things one uh i think is
00:37:30
an ins and a and a thesis or something that is is really fantastic that would
00:37:34
it be to one uh is the next user in and i can just printed
00:37:37
the terminal divide it and then you would also want and it seems to support multiple
00:37:42
not all the same time i think really actually um but i haven't had any to figure out like it's not gonna say so
00:37:47
but that was really cool not the module cutting solutions also fantastic 'cause
00:37:52
that was my main concern and wounded knee thinking about how did intelligent
00:37:55
we should sources and immediately you just described it so i wanted to and you can see that and then the next one to ask those
00:38:02
uh is it possible for a tool to support a subset uh oh the the protocol
00:38:08
uh is a snake we should can be declared that but also it's not me i mean it's not hardens too hard with
00:38:16
all that that subject so that that's the appropriate or opinion yes
00:38:19
uh so the other speed and use p. protocol already has a
00:38:25
messages and uh has that so you take the
00:38:29
the server and both server and client can declare there
00:38:33
capabilities there's some minimum set of capabilities that
00:38:36
you need to support to have reasonable integration
00:38:41
but for instance a bill to doesn't even need to be able to support compile
00:38:47
you can already import it into intelligent a if
00:38:50
it just uh supports a some basic messages like uh
00:38:56
give me targets and give me sources associate with the targets that's enough to imported and intelligent
00:39:03
i look at the disney critical now and thank you very much yeah
00:39:07
yeah
00:39:09
next question here
00:39:12
you can you give us a rough idea of the magnitude of the task of adding
00:39:19
a new build tool like you know how many lines of code to add to right
00:39:23
uh into in tele j. or creating a b. s. p.
00:39:28
support within the bill tool for creating d. s. p.s support
00:39:32
i think it would depend a lot on the bill to uh so the
00:39:35
basic use p. support i had in winter fury was of four five hundred lines
00:39:43
yeah
00:39:46
was not too complicated most of it was understanding how fury works i um
00:39:55
but it is it has to be the that if we have an imitation and you couldn't just begins i think there
00:40:00
probably is a invisible and there's a test kit in the
00:40:03
u. s. p. implementation or so there is a set of
00:40:09
tests that you can run against the server uh there is
00:40:15
it's colour checked generators for all the messages in p. s. p.
00:40:20
and a few other things oh it's not complete but it should give you a start yeah any other
00:40:31
well yeah
00:40:37
uh_huh so elsewhere with joe brains um
00:40:40
have you made strides with getting other language
00:40:45
support in ontology a chicken's m. b. s. p. to interact with their build tools
00:40:50
not him but we've talked about it internally so so um my approaches
00:40:55
if this had gets any traction and skylark and we have a reasonably
00:41:01
a stable support for the then we'll extracted to a separate
00:41:06
plugin which could also make it into the core distribution of intelligent
00:41:14
uh anymore comments
00:41:19
okay ah
00:41:22
that's it then i'll just ask a few things uh
00:41:26
want to give me honest feedback it's anonymous so you can be honest on the scale uh they zap
00:41:33
such as a vote
00:41:36
and the and the next is some shameless plug so i mentioned my
00:41:43
most of our team is here are and the two of them
00:41:48
will be giving talks that the developer jungle hosted by word of
00:41:52
slab if you have seen it the place without the plants uh one of them is going to be by dimitri his over here
00:42:00
uh on some cool new tools for big data if in developing
00:42:07
so check that out please and the other is about some
00:42:12
work we're doing and uh scallop plugin so about data flow analysis
00:42:20
so i'm looking forward to that one as well because i don't really understand what's going on there
00:42:26
so um i collect toby will explain more about
00:42:31
how we're gonna give more insight into the code
00:42:37
from the scallop like and these are both lightning talks ten to fifteen minutes
00:42:43
and lastly i oh we have a booth and this is
00:42:47
as so come talk to us any time during the conference
00:42:53
we have these two talks are tomorrow
00:42:56
fourteen forty sixteen thirty
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.
6375 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.
302 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.
245 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.
2707 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