Skip to content

Latest commit

 

History

History
854 lines (653 loc) · 33.4 KB

CONTRIBUTING.md

File metadata and controls

854 lines (653 loc) · 33.4 KB
Raw
layout title category tags order
developer-doc
Contributing to Enso
summary
summary
contributing
2

Contributing to Enso

Thank you for your interest in contributing to Enso! We believe that only through community involvement can Enso be the best it can be! There are a whole host of ways to contribute, and every single one is appreciated. The major sections of this document are linked below:

All contributions to Enso should be in keeping with our Code of Conduct.

The Contributor License Agreement

As part of your first contribution to this repository, you need to accept the Contributor License Agreement. You will automatically be asked to sign the CLA when you make your first pull request.

Any work intentionally submitted for inclusion in Enso shall be licensed under this CLA.

The CLA you sign applies to all repositories associated with the Enso project, so you will only have to sign it once at the start of your contributions.

Issues

If you're wanting to get involved with Enso's development and are looking for somewhere to start, you can check out the following tags in our issues:

You can use the "Size" and "Difficulty" labels that should be assigned to every issue to get a better idea of how much work a given issue might be.

Feature Enhancements

If you feel like you have a suggestion for a change to the way that Enso works as a language, please take a look at the Enso RFC process to learn how to file an RFC for the project.

In essence, the RFC process provides a way to propose major changes to the language, the compiler, and the runtime in a way that ensures that they get seen and discussed by all the major stakeholders involved.

If, on the other hand, you're asking for a smaller feature, please feel free to submit a feature request to the repository.

Bug Reports

While it's never great to find a bug, they are a reality of software and software development! We can't fix or improve on the things that we don't know about, so report as many bugs as you can! If you're not sure whether something is a bug, file it anyway!

If you are concerned that your bug publicly presents a security risk to the users of Enso, please look at our security guidelines.

Even though GitHub search can be a bit hard to use sometimes, we'd appreciate if you could search for your issue before filing a bug as it's possible that someone else has already reported the issue. We know the search isn't the best, and it can be hard to know what to search for, so we really don't mind if you do submit a duplicate!

Opening an issue is as easy as following this link and filling out the fields. The template is intended to collect all the information we need to best diagnose the issue, so please take the time to fill it out accurately.

The reproduction steps are particularly important, as the more easily we can reproduce it, the faster we can fix the bug! It's also helpful to have the output of enso --version, as that will let us know if the bug is Operating System or Architecture specific.

Hacking on Enso

This will get you up and running for Enso development, with only a minimal amount of setup required. Enso's build system is fairly simple, allowing you to bootstrap the compiler as long as you have a minimal set of tools.

Design Documentation

If you're going to start contributing to Enso, it is often a good idea to take a look at the design documentation for the language. These files explain provide both a rigorous specification of Enso's design, but also insight into the why behind the decisions that have been made.

These can be found in docs/, and are organised by the part of the compiler that they relate to.

System Requirements

The following operating systems are supported for developing Enso:

  • Windows 10
  • macOS 10.14 and above
  • Linux 4.4 and above

Currently we support x86_64 (all mentioned OS) and arm64 (Mac only) architectures. You may be able to develop Enso on other systems, but issues arising from unsupported configurations will not be fixed by the core team.

In order to build and run Enso you will need the following tools:

  • NodeJS with the latest LTS version. We recommend using nvm for managing NodeJS installation.
  • sbt with the same version as specified in project/build.properties.
  • Maven with version at least 3.6.3.
  • GraalVM with the same version as described in the build.sbt file, configured as your default JVM. GraalVM is distributed for different Java versions, so you need a GraalVM distribution for the same Java version as specified in build.sbt.
  • Flatbuffers Compiler with version 1.12.0.
  • Rustup, the rust toolchain management utility.
  • On MacOS and Linux, the tar command is required for running some tests. It should be installed by default on most distributions.
  • If you want to be able to build the Launcher Native Image, you will need a native C compiler for your platform as described in the Native Image Prerequisites. On Linux that will be gcc, on macOS you may need xcode and on Windows you need to configure the Developer Command Prompt for Microsoft Visual C++ for the x64 architecture.

Managing multiple JVM installations can be a pain, so you can consider using helper tools for that. We recommend:

For users of M1 Mac: installing GraalVM on M1 Mac requires manual actions, please refer to a dedicated documentation.

The flatbuffers flatc compiler can be installed from the following locations:

  • Using the conda package manager (conda install flatbuffers). This will work on all platforms, but requires some knowledge of conda and how its environments work.
  • Windows users can download binaries directly from the flatbuffers github releases.
  • MacOS users can install it via homebrew (brew install flatbuffers).

Getting the Sources

Given you've probably been reading this document on GitHub, you might have an inkling where to look!. You can clone Enso using two methods:

  • Via HTTPS: We recommend you only use HTTPS if checking out the sources as read-only.
git clone https://github.com/enso-org/enso.git
  • Via SSH: For those who plan on regularly making direct commits, cloning over SSH may provide a better user experience (but requires setting up your SSH Keys with GitHub).
git clone git@github.com:enso-org/enso.git

Getting Set Up (Rust)

The Rust code in this repository requires a specific nightly rust toolchain, as defined by rust-toolchain override file. The rustup will automatically download the appropriate compiler version along with the necessary components.

Please consult the GUI Contribution Guide to learn details on setting your system up. Quick summary:

enso$ rustup toolchain install stable                  # Stable toolchain required for the following tools.
enso$ cargo +stable install wasm-pack --version 0.10.2 # Install the wasm-pack toolkit.
enso$ cargo +stable install cargo-watch                # To enable `./run wasm watch` utility

The previous three steps shall be enough to build the IDE via ./run wasm build run wasm build --wasm-profile dev.

Getting Set Up (JVM)

In order to properly build the runtime component, the JVM running SBT needs to have some dependency JARs available in its module path at startup. To ensure they are available, before running any compilation or other tasks, these dependencies should be prepared. To do so, run the following command in the repository root directory:

sbt bootstrap

It is preferred to not run this command from the sbt shell, but in batch mode, because SBT has to be launched again anyway to pick up these JARs at startup.

Bootstrap has to be run only when building the project for the first time and after each change of Graal version.

Getting Set Up (Documentation)

We enforce automated formatting of all of our documentation and configuration using the fairly common prettier automatic formatter. You can install prettier for our project by running the following command:

npm install

This does, however, mean that you have to have node installed on your system. Please follow the guidelines above to install node if you have not already done so.

The version if prettier is forced by our package-lock.json in order for us to make formatting bumps all at once.

You can format all of our documentation and configuration as follows:

npx prettier --write <dir>

Building Enso

There are multiple projects in this repository, but all can be built, run and tested using sbt. As long as your configuration is correct, with the correct versions of SBT, Rust and GraalVM, the same steps can be followed on all of our supported platforms (Linux, MacOS and Windows).

SBT will handle downloading and building library dependencies as needed, meaning that you don't need to handle any of this manually.

Please note that at the current time, the Windows build of GraalVM is in an experimental state. This means that while it may function, we are not intending to provide work-arounds for building on that platform while it is still in an unstable state.

Building Enso Components

In order to build a specific component (e.g. runtime), please follow the following steps.

  1. Enter the sbt shell in the repository root directory by typing sbt.
  2. Change to the project you are concerned with (in our case runtime) by executing project runtime.
  3. Execute compile in order to compile the project in question. This will compile the project and all its dependencies as necessary.

You can substitute both bench and test for compile in step 3, and the sbt shell will execute the appropriate thing. Furthermore we have testOnly and benchOnly that accept a glob pattern that delineates some subset of the tests or benchmarks to run (e.g. testOnly *FunctionArguments*).

Building the Interpreter CLI Fat Jar

In order to build a fat jar with the CLI component, run the assembly task inside the runner subproject:

sbt "engine-runner/assembly"

This will produce an executable runner.jar fat jar and a runtime.jar fat jar in the repository root. The runner.jar depends only on the runtime.jar and a vanilla GraalVM distribution.

Building the Project Manager Fat Jar

In order to build a fat jar with the Project Manager component, run the assembly task on the project-manager subproject:

sbt "project-manager/assembly"

This will produce a project-manager fat jar and a runtime.jar fat jar in the repository root.

Building the Launcher Native Binary

If you want to build the native launcher binary, you need to ensure that the Native Image component is installed in your GraalVM distribution. To install it, run:

<path-to-graal-home>/bin/gu install native-image

Then, you can build the launcher using:

sbt launcher/buildNativeImage

Passing Debug Options

GraalVM provides some useful debugging options, including the ability to output the compilation graph during JIT optimisation, and the ASM generated by the JIT.

However, as we don't want these things polluting our standard builds, we provide a helper SBT command withDebug to allow for passing these options. It supports the following flags:

  • --dumpGraphs: This dumps the IGV (a Graal tool) graphs for the program to allow for manual analysis and discovery of optimisation failures.
  • --showCompilations: Prints the truffle compilation trace information.
  • --printAssembly: Prints the assembly output from the HotSpot JIT tier.

For more information on this sbt command, please see WithDebugCommand.scala.

It is used as an addendum to the basic sbt command you want to run (e.g. test from above). The format is withDebug COMMAND [OPTIONS...], and if you need to pass any additional options to COMMAND you must do so following a --. For example:

withDebug run --dumpGraphs --printAssembly -- --run MyFile.enso
withDebug benchOnly --showCompilations -- RecursionBenchmark

Step by step debugging can be triggered as

withDebug testOnly --debugger -- *FavoriteTest*

read more about debugging Java & Enso code.

Working with Assembly

In order to examine the assembly generated by GraalVM and HotSpot you need to provide your JVM install with a dynamic library that supports the dumping of assembly. It can be acquired for MacOS and Linux here, and for windows from here. There are other methods to acquire it, as well, so please choose one best suited for you.

Once you have a copy of the dynamic library, it needs to be placed in $JVM_HOME/lib/server.

Native Image

Native image is a capability provided alongside GraalVM that allows the generation of native executables from JVM language programs (such as the Enso interpreter itself). However, it results in significantly degraded peak performance, so it is not part of our roadmap currently.

If you would like to experiment with it, you can execute the buildNativeImage command in the sbt shell while inside the runner project. Please note that while the command is available at the moment, and you are welcome to report an issue with the functionality, any bugs you report will not be considered high priority.

WE CURRENTLY DO NOT SUPPORT THE NATIVE IMAGE BUILD.

Using IntelliJ

Internally, most of the developers working on the Enso project use IntelliJ as their primary IDE. To that end, what follows is a basic set of instructions for getting the project into a working state in IntelliJ.

  1. Clone the project sources.
  2. Open IntelliJ
  3. File -> New -> Project From Existing Sources.
  4. Navigate to the directory into which you cloned the project sources. By default this will be called enso. Select the directory, and not the build.sbt file it contains.
  5. In the 'Import Project' dialogue, select 'Import project from external model' and choose 'sbt'.
  6. Where it says 'Download:', ensure you check both 'Library Sources' and 'sbt sources'.
  7. In addition, check the boxes next to 'Use sbt shell:' such that it is used both 'for imports' and 'for builds'.
  8. Disallow the overriding of the sbt version.
  9. Under the 'Project JDK' setting, please ensure that it is set up to use a GraalVM version as described in System requirements. You may need to add it using the 'New' button if it isn't already set up.
  10. Click 'Finish'. This will prompt you as to whether you want to overwrite the project folder. Select 'Yes' to continue. The Enso project will load up with an open SBT shell, which can be interacted with as described above. You will want to use scalafmt for formatting of Scala code, and install Google Java Format for formatting Java code. For more information see the relevant Style Guides.

Depending on the version of GraalVM with which you are working, you may be required to add the following flags to the per-module overrides for IntelliJ's java compiler in order for it to not show spurious errors. This is because some versions of GraalVM export their own closed version of com.oracle.truffle.api that IntelliJ picks up preferentially to the version we use for development. You can find these options in Preferences -> Build, Execution, Deployment -> Compiler -> Java Compiler.

--add-exports org.graalvm.truffle/com.oracle.truffle.api=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.debug=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.dsl=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.exception=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.frame=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.instrumentation=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.interop=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.io=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.library=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.memory=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.nodes=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.object=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.profiles=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.source=ALL-UNNAMED
--add-exports org.graalvm.truffle/com.oracle.truffle.api.utilities=ALL-UNNAMED

However, as mentioned in the Troubleshooting section below, the forked nature of execution in the SBT shell means that we can't trivially make use of the IntelliJ debugger. In order to get debugging working, you will need to follow these steps:

  1. Go to Run -> Edit Configurations.
  2. Click the + button in the header of the 'Run/Debug Configurations' dialogue that pops up.
  3. Select 'Remote' and name the new configuration appropriately.
  4. In the options for that configuration select 'Listen to remote JVM' under 'Debugger mode:'
  5. Where it provides the command-line arguments for the remote JVM, copy these and add them to truffleRunOptions in build.sbt. Remove the portion of these options after suspend=y, including the comma. They are placeholders that we don't use.
  6. Alternatively, certain tasks, such as run, benchOnly and testOnly can be used through the withDebug SBT command. For this to work, your remote configuration must specify the host of localhost and the port 5005. The command syntax is withDebug --debugger TASK_NAME -- TASK_PARAMETERS, e.g. withDebug --debugger testOnly -- *AtomConstructors*.
  7. Now, when you want to debug something, you can place a breakpoint as usual in IntelliJ, and then execute your remote debugging configuration. Now, in the SBT shell, run a command to execute the code you want to debug (e.g. testOnly *CurryingTest*). This will open the standard debugger interface and will allow you to step through your code.

Please be careful to ensure that you don't commit these changes to the sbt configuration as they are specific to your machine.

Troubleshooting

If you are having issues building Enso, please check the list below before filing an issue with us.

  • StackOverflowError During Compilation: Please ensure that your version of sbt is respecting the project's .jvmopts settings. We make significant use of recursion when expanding macros for the parser, and these require use of additional stack. Alternatively, you can explicitly pass -Xss8M to the sbt invocation.
  • Debugging Not Working: The sbt tasks run the invoked programs in a forked JVM. This means that to attach a debugger to it you need to use the JVM remote debugging support. Follow Enso debugging instructions or see the Using IntelliJ section for instructions.

If your problem was not listed above, please file a bug report in our issue tracker and we will get back to you as soon as possible.

Testing Enso

Running the tests for the JVM enso components is as simple as running sbt / test. To test the Rust components you can run ./run wasm test.

Testing Enso Libraries

To test the libraries that are shipped with Enso you need to first build the engine, the easiest way to do so is to run sbt buildEngineDistribution. That will create a distribution in the directory built-distribution. The engine runner that can be used for running the tests is located at built-distribution/enso-engine-<VERSION>-linux-amd64/enso-<VERSION>/bin/enso (or enso.bat for Windows).

To run the tests you can run the following commands (where enso refers to the built runner executable as explained above):

enso --run test/Tests # for the Base library
enso --run test/Geo_Tests
enso --run test/Table_Tests

The Database tests will by default only test the SQLite backend, to test other backends see test/Table_Tests/src/Database/README.md for information on how to configure them.

The Base tests rely in a few places on the system language. On Linux you can set the LANG environment variable to C to make sure that the language is configured correctly and run the tests as following:

LANG=C enso --run test/Tests

Test Dependencies

Some test suites require extra setup and enabled only on CI. To replicate the CI environment you should install and run extra services:

# Httpbin
go get -v github.com/ahmetb/go-httpbin/cmd/httpbin
$(go env GOPATH)/bin/httpbin -host :8080

To run all the stdlib test suites, set CI=true environment variable:

env CI=true enso --run test/Tests/

For more details about the CI setup, you can check the .github/workflows/scala.yml GitHub workflow.

Running Enso

The language interpreter can be started by the bin/enso launcher script located inside of the Enso runtime distribution. Use the following sbt command to compile necessary bits (see Building the Interperter CLI Fat Jar) and generate the Enso distribution:

Bash
$ sbt buildEngineDistribution
...
Engine package created at built-distribution/enso-engine-0.0.0-dev-linux-amd64/enso-0.0.0-dev
PowerShell
sbt.bat buildEngineDistribution

Then one can execute the launcher:

$ built-distribution/enso-engine-0.0.0-dev-linux-amd64/enso-0.0.0-dev/bin/enso

Detailed information on the flags it supports can be shown with the --help flag, but the primary functionality is as follows:

  • --new PATH: Creates a new Enso project at the location specified by PATH.
  • --run PATH: Executes the interpreter on the Enso source specified by PATH. In this case, PATH must point to either a standalone Enso file or an Enso project.
Bash
distribution/bin/enso --new ~/Hello
distribution/bin/enso --run ~/Hello
Hello, World!
PowerShell
distribution/bin/enso.bat --new ~/Hello
distribution/bin/enso.bat --run ~/Hello
Hello, World!

Running IDE

You can start IDE with a development version of the language server. IDE executable has --external-backend flag that switches off the bundled backend. That requires you to run the project manager process yourself. You can either get a project manager from one of the latest releases on GitHub, or build one using SBT buildProjectManagerDistribution command.

Running development version of the IDE is possible via the ./run script in the root of the repository:

$ ./run ide start --wasm-profile dev --external-backend
Bash
sbt buildProjectManagerDistribution
PowerShell
sbt.bat buildProjectManagerDistribution

When the command is completed, a development version of the project manager will have appeared in the built-distribution directory.

The IDE will connect to the running project manager to look up the project and start the language server. The required version of the language server is specified in the edition field of the package.yaml project description. Enso projects are located in the ~/enso directory on Unix and %userprofile%\enso on Windows systems by default.

cat ~/enso/projects/Unnamed/package.yaml
name: Unnamed
namespace: local
version: 0.0.1
license: ""
authors: []
maintainers: []
edition: "2021.20-SNAPSHOT"
prefer-local-libraries: true

We need to set edition to a value that will represent the development version. It should be different from any Enso versions that have already been released. In this case, we chose the 2021.20-SNAPSHOT (the current development edition). The project manager will look for the appropriate subdirectory in the engines directory of the distribution folder. Distribution paths are printed when you run project manager with -v verbose logging.

Btw. you can specify ENSO_JVM_OPTS to turn debugging of the Engine runtime on:

$ export ENSO_JVM_OPTS=-agentlib:jdwp=transport=dt_socket,address=5005
$ ./built-distribution/enso-project-manager-0.0.0-dev-linux-amd64/enso/bin/project-manager --no-log-masking -v
[info] [2021-06-16T11:49:33.639Z] [org.enso.projectmanager.boot.ProjectManager$] Starting Project Manager...
[debug] [2021-06-16T11:49:33.639Z] [org.enso.runtimeversionmanager.distribution.DistributionManager] Detected paths: DistributionPaths(
  dataRoot = /home/dbv/.local/share/enso,
  runtimes = /home/dbv/.local/share/enso/runtime,
  engines  = /home/dbv/.local/share/enso/dist,
  bundle   = None,
  config   = /home/dbv/.config/enso,
  locks    = /run/user/1000/enso/lock,
  tmp      = /home/dbv/.local/share/enso/tmp
)

On Linux it looks for the ~/.local/share/enso/dist/0.2.32-SNAPSHOT/ directory.

We can build an engine distribution using the buildEngineDistribution command in SBT.

Bash
sbt buildEngineDistribution
PowerShell
sbt.bat buildEngineDistribution

And copy the result to the 0.2.32-SNAPSHOT engines directory of the distribution folder.

Bash
cp -r built-distribution/enso-engine-0.2.32-SNAPSHOT-linux-amd64/enso-0.2.32-SNAPSHOT ~/.local/share/enso/dist/0.2.32-SNAPSHOT
PowerShell
cp -r built-distribution/enso-engine-0.2.32-SNAPSHOT-linux-amd64/enso-0.2.32-SNAPSHOT ~/.local/share/enso/dist/0.2.32-SNAPSHOT

Now, when the project manager is running and the engines directory contains the required engine version, you can start IDE with the --no-backend flag. It will pick up the development version of the language server we just prepared.

To summarize, these are the steps required to run IDE with the development version of the language server.

  1. Run the project manager process.
  2. Copy or symlink the development version of the engine created with SBT's buildEnginedistribution command to the engines directory of the Enso distribution folder.
  3. Set the edition field of the package.yaml project definition to the version that you created in the previous step.
  4. Run the IDE with --no-backend flag.

Language Server Mode

The Language Server can be run using the --server option. It requires also a content root to be provided (--root-id and --path options). Command-line interface of the runner prints all server options when you execute it with --help option.

Below are options uses by the Language Server:

  • --server: Runs the Language Server
  • --root-id <uuid>: Content root id.
  • --path <path>: Path to the content root.
  • --interface <interface>: Interface for processing all incoming connections. Default value is 127.0.0.1
  • --rpc-port <port>: RPC port for processing all incoming connections. Default value is 8080.
  • --data-port <port>: Data port for visualisation protocol. Default value is 8081.

To run the Language Server on 127.0.0.1:8080 type:

distribution/bin/enso \
  --server \
  --root-id 3256d10d-45be-45b1-9ea4-7912ef4226b1 \
  --path /tmp/content-root

If you want to provide a socket that the server should listen to, you must specify the following options:

  • --interface: The interface on which the socket will exist (e.g. 0.0.0.0).
  • --port: The port on interface where the socket will be opened (e.g. 80).

Pull Requests

Pull Requests are the primary method for making changes to Enso. GitHub has fantastic documentation on using the pull request feature. Enso uses the 'fork-and-pull' model of development. It is as described here and involves people pushing changes to their own fork and creating pull requests to bring those changes into the main Enso repository.

Please make all pull requests against the develop branch.

  • We run CI on all contributions to Enso, but it's still useful for you to run the tests yourself locally first! This can be done by running test in the enso project in sbt.
  • Additionally, please ensure that your code conforms to the Enso style guides, particularly the Scala Style Guide and the Java Style Guide.

Make sure you perform these checks before every pull request. You can even add git hooks before every push to make sure that you can't forget.

  • Every pull request to the Enso repository is reviewed by a member of the core team! You'll get assigned a reviewer based on the areas your PR touches, but please feel free to ask for a specific person if you've worked with them in a specific area before!
  • If you have questions, or would like to begin the review process before your PR is 'done', please use the Draft Pull Requests feature on GitHub. Doing so will allow you to make use of our CI infrastructure as part of your development process.

Once the reviewer approves your pull request it will be tested by our continuous integration provider before being merged. If we request changes to your PR, please feel free to discuss the suggestions and comments! We can only achieve the best results through open collaboration.

Documentation

Documentation improvements are very welcome! For now, the main documentation available is the developer documentation for the language, which can be found at the dev docs site. The source for this documentation is found in the docs/ folder, and can be altered from there.

Documentation pull requests will be reviewed in exactly the same way as normal pull requests.

To find documentation-related issues, sort by the Category: Documentation label.

Issue Triage

Sometimes issues can be left open long after the bug has been fixed. Other times, a bug might go stale because something has changed in the meantime.

It can be helpful to go through older bug reports and make sure that they are still valid. Load up an older issue, double check that it's still true, and leave a comment letting us know if it is or is not. The least recently updated sort is good for finding issues like this.

Contributors with sufficient permissions can help by adding labels to help with issue triage.

If you're looking for somewhere to start, take a look at the Difficulty: Beginner issue label, as well as the Status: Help Wanted and Status: Good First Issue labels.

Out-of-Tree Contributions

As helpful as contributing to Enso directly is, it can also be just as helpful to contribute in other ways outside this repository:

Helpful Documentation and Links

For people new to Enso, and just starting to contribute, or even for more seasoned developers, some useful places to look for information are: