Exploring the Scala's Tooling Ecosystem

https://nescala.io/proposed-talks.html#tooling
https://typelevel.org/event/2019-06-summit-lausanne/
https://schedule.scala.io/#/talk/77507
https://bee-scala.org/sessions/exploring-scala-tooling-ecosystem/

Transcript

1. Exploring the Scala's
   Tooling Ecosystem
   

   View Slide

2. Q&C
   sli.do
   #tooling
   

   View Slide

3. What are tools used for?
   Editing, compiling and testing.
   

   View Slide

4. What are tools used for?
   Editing, compiling and testing, ...
   learning, experimenting, analysing, reviewing code,
   debugging, dependency management, publishing, versioning,
   benchmarking, creating other tools, ...
   coffee, ...
   

   View Slide

5. Why do we care ?
   

   View Slide

6. Why do we care ?
   Productivity.
   

   View Slide

7. View Slide

8. LSP: Language Server Protocol
   The Language Server Protocol defines the protocol used
   between an editor or an IDE and a language server that
   provides language features like auto complete, go to
   definition, find all references, etc.
   

   View Slide

9. View Slide

10. View Slide

11. Language Servers in Scala lang
    ● Dragos-vscode-scala
    ● Metals
    ● Dotty LS
    ● SBT server (since SBT 1.x)
    

    View Slide

12. Editors/IDEs supporting the LSP
    ● IntelliJ
    ● Eclipse IDE
    ● VS code
    ● Atom
    ● Sublime Text 3
    ● vim8/neovim
    ● Emacs
    See ls.org and ms.lsp for the latest on adoption.
    

    View Slide

13. Metals @scalameta
    

    View Slide

14. Metals
    Works with: vs-code, atom, vim, Sublime text 3, Emacs.
    Uses BSP to communicate LS and build tools.
    “Our goal in Metals is to support code navigation with as low
    CPU and memory overhead as possible without sacrificing
    rich functionality.”
    

    View Slide

15. View Slide

16. View Slide

17. Metals v0.7.6 - Thorium
    ● Automatic import build for Gradle, Maven and Mill
    ● Auto-fill option to case classes and functions
    ● Code completions on keywords
    ● Deduce values for named parameters completion
    ● Compilation explorer
    ● Add support for Scala 2.13 and JDK 11
    ● Coming soon: rename symbols, find definition, DAP,
    worksheets and much more!
    

    View Slide

18. SemanticDB
    ● Portable semantic APIs
    ● Data schema for semantic information about code.
    ● Relevant for developer tools.
    ● Protobuf, JSON, SQL.
    ● Do not require a running compiler.
    

    View Slide

19. BSP: Build Server Protocol
    ● Extends LSP
    ● Formalizes interaction between editors, LS, editors, build
    tools
    ● Bidirectional notifications,
    ● Client/server architecture.
    ● Clients are build tools, LS and editors/IDEs.
    ● Language agnostic.
    ● JSON-RPC-protobuf.
    

    View Slide

20. Bloop
    ● sbt, Maven, Gradle and Mill.
    ● IntelliJ and Metals (VS Code, Sublime, vim and Atom)
    ● Provides fast compile, test and run
    ● Has a built-in command-line tool
    ● Integrates with most JVM build tools
    ● Supports JVM, Scala.js and Scala Native
    

    View Slide

21. Bloop (v1.3.2)
    ● Compiling deduplication
    ● Compiling isolation
    ● Zipkin tracing support
    ● Use Ammonite by default in bloop console
    ● Support for Scala 2.13.x, JDK 9 & 11
    ● Improvements in build tools support
    

    View Slide

22. (v1.3.3)
    ● Out-of-box Coursier integration
    ● Better classloader management
    ● IO improvements
    ● Custom incremental tasks
    ● SemanticDB support
    ● JDK 11 support
    

    View Slide

23. Mill
    ● Build tools as Pure functional programs
    ● Fast and Flexible
    ● Targets with caching
    ● Inspectale cache graph
    ● Commands
    ● Can act as a build server
    ● Play framework and Bloop support
    ● Support for Dotty
    

    View Slide

24. Seed
    ● Builds are describe in a TOML file
    ● Handles dependency resolution via coursier
    ● Generates configurations for Bloop and Intellij
    ● BSP server integration using ZIO
    ● Cross-platform support (JVM, JS, LLVM)
    ● Cross-compile modular projects made easy
    

    View Slide

25. Pants
    ● Source based dependency tool
    ● Aims to large code bases
    ● Monorepo environments
    ● Fast
    ● Scalable
    ● dependency inference from source-code imports
    

    View Slide

26. Bazel (v1.0)
    ● Reproducible builds
    ● Fast and correct builds
    ● Heavy caching
    ● Enforces build correctness
    ● Large codebases
    ● Multi-language, multi-platform
    ● Allow builds to scale
    

    View Slide

27. TL;DR
    Tools that work with my editor => LSP
    Build integrations => BSP
    Navigation => SemanticDB, TASTy
    Refactoring => scalafix
    Diagnostics => BSP + LSP
    

    View Slide

28. Additional links...
    https://www.scala-lang.org/blog/2018/02/14/tooling.html
    https://github.com/scalacenter/tooling-working-groups/blob/master/meetin
    gs/2018-01-17/minutes.md
    https://scala.sphere.it/#agenda
    https://contributors.scala-lang.org/c/tooling
    https://www.scala-lang.org/contribute/tools.html
    https://scala.epfl.ch/projects.html
    https://microsoft.github.io/language-server-protocol/
    https://github.com/Microsoft/language-server-protocol
    

    View Slide

29. ¡Gracias!
    @unyagami
    

    View Slide