The Info Tree
The Info Tree is another representation of a successful parse that can be generated by the parsers in the system. An info tree is similar to a parse tree in that it closely resembles how the grammar matched the input string. In contrast to a parse tree, an info tree is aimed to provide information about syntax highlighting and indentation. As such, there are a few differences between a parse tree and an info tree:
- The info tree captures all parts of the parse tree, regardless of what parts of the parse tree were captured in the grammar. It is therefore possible to reconstruct the parsed string from an info tree. This is not always true for a parse tree in Storm.
- The info tree is not typed. It is a generic, type-less version of the parse tree. It is thus not possible to directly transform it into an abstract syntax tree (one may of course traverse it manually, but that eliminates many benefits of the Syntax Language). On the other hand, the typeless nature makes it possible to create a generic parser for a set of syntax rules, and then adjust which rule is the start rule without having to re-create the parser.
- The info tree contains information about syntax highlighting and indentation.
The info tree representation is implemented by the classes lang.bnf.InfoNode,
lang.bnf.InfoInternal, and lang.bnf.InfoLeaf. This representation is used by the language
server to provide syntax highlighting and indentation of source code.