An Erlang language server, providing IDE services.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.vscode
apps
.gitignore
.travis.yml
.vscodeignore
CONTRIBUTORS
COPYRIGHT.txt
Jenkinsfile
LICENSE
README.md
rebar.config
rebar.config.script
rebar.lock
rebar3
sourcer.code-workspace

README.md

Sourcer - language services for IDEs

This is a new incarnation of the sourcer project, the old code is still available on the legacy branch.

Sourcer aims to provide:

  • a generic language server implementation lsp_server
  • Erlang support for the above erlang_ls
  • Erlang language services library sourcer

Generic language server lsp_server

The generic language server uses a TCP connection to talk LSP with clients. It encodes/decodes the messages and delegates the actual work to a language specific server, using cancelable worker processes. It also supports making requests to the client.

Erlang server erlang_ls

This is the "real" server, connecting the generic and specific parts and an adapter from LSP to the 'sourcer' data format (see below) and back.

Supported Options

  • --help, -h, undefined, Shows help;
  • --dump, -d, string, Dump sourcer db for file;
  • --port, -p, integer, LSP server port;
  • --verbose, -v, integer, Verbosity level;
  • --indent, -i, string, Indent file(s) and exit;
  • --config, N/A, string, Configuration file.

Configuration

Configuration file contains property list.

Currently we only support these keys:

  • indent - controls indentation parameters.

Example:

{indent, [{indentW, 2}]}.

Language services library sourcer

The actual work is done by this application. It is meant to be LSP-agnostic so that it can be used in other contexts and tools.

An own parser (largely based on the legacy 'sourcer' code, but simpler) processes the source code and produces a ctags-like database containing information about all interesting code entities. The database should be distributed, in the sense that libraries should be able to provide the data about their own code (produced at build time).

Query facilities are provided so that the information in the database is presented in a way that the LSP expects.