Go Language Server

go-langserver is a Go language server that speaks Language Server Protocol. It supports editor features such as go-to-definition, hover, and find-references for Go projects.

Open in Sourcegraph

To build and install the standalone go-langserver run

go get -u github.com/sourcegraph/go-langserver

Support

	Hover	Jump to def	Find references	Workspace symbols	VFS extension	Isolated	Parallel
Go	✔	✔	✔	✔	✔	✔	✔

InitializationOptions

If you are a client wanting to integrate go-langserver, you can use the following as initializationOptions in your initialize request to adjust the behaviour:

interface GoInitializationOptions {
  /**
   * funcSnippetEnabled enables the returning of argument snippets
   * on `func` completions, eg. func(foo string, arg2 bar).
   * Requires code completion to be enabled.
   *
   * Defaults to true if not specified.
   */
  funcSnippetEnabled?: boolean;

  /**
   * gocodeCompletionEnabled enables code completion feature (using gocode).
   *
   * Defaults to false if not specified.
   */
  gocodeCompletionEnabled?: boolean;

  /**
   * MaxParallelism controls the maximum number of goroutines that should be used
   * to fulfill requests. This is useful in editor environments where users do
   * not want results ASAP, but rather just semi quickly without eating all of
   * their CPU.
   *
   * Defaults to half of your CPU cores if not specified.
   */
  maxParallelism?: number;

  /**
   * useBinaryPkgCache controls whether or not $GOPATH/pkg binary .a files should
   * be used.
   *
   * Defaults to true if not specified.
   */
  useBinaryPkgCache?: boolean;
}

Profiling

If you run into performance issues while using the language server, it can be very helpful to attach a CPU or memory profile with the issue report. To capture one, first install Go and then:

Capture a heap (memory) profile:

go tool pprof -svg $GOPATH/bin/go-langserver http://localhost:6060/debug/pprof/heap > heap.svg

Capture a CPU profile:

go tool pprof -svg $GOPATH/bin/go-langserver http://localhost:6060/debug/pprof/profile > cpu.svg

Since these capture the active resource usage, it's best to run these commands while the issue is occurring (i.e. while memory or CPU is high).