Homepage: http://carboni.ca/projects/p/redtree
Git: https://github.com/michaeledgar/redtree
Author: Michael Edgar
Copyright: 2011
License: MIT
Latest Version: none
Release Date: None, yet.

Redtree is a Ruby parser in the spirit of Ripper and Melbourne, but with additional capabilities. It carries with it the entire shift-reduce sequence and token boundaries, so a consumer of a Redtree parse tree knows exactly where each token occurred.

Redtree was borne out of the Laser project, which needs to be able to rewrite source code token-by-token (including whitespace, a typically ignored token) for proper style-linting. It is also difficult to extract correct line numbers from a Ripper parse tree, and impossible to reliably obtain column numbers from Ripper or Melbourne. Hence: Redtree.

Details are always forthcoming, but:

1. Full Token Information: Each internal node of the parse tree knows exactly which tokens (or other internal nodes) composed the node's reduction during parsing. Every token is available, including its line/column start and end points.
2. Comment Nodes: Attaching comments to nodes is a task required of documentation tools, linters, static analyzers, and so on. Yet no Ruby parser adequately includes comments when necessary.
3. Compact Representation: Including this much information is possible with a compact, simple representation. Redtree incorporates this design to minimize its space use.
4. Fast Parsing: Parsing should be fast. Before converting to a more natural AST, Redtree runs at least 40% faster than Ripper. Of course, Ripper does more work converting to a simpler AST, but Redtree is stashing away a lot more information.

To install Redtree, use the following command:

$ gem install redtree --prerelease

(Add sudo if you're installing to a directory requiring root privileges to write)

When it works, I'll add it here. I promise.

• Aug.29.2011: Project initiated.

Redtree © 2011 by Michael Edgar. Redtree is licensed under the MIT license. See {file:LICENSE}.