Some bugfixes, some new features.

[maxcalandrelli]

Changes summary:

• gocc now parses its own grammar and produces a working parser for itself
• bugfixes in handling of dots and regdefids in the lexical analyzer
• eliminated ambiguity between literals and labels of productions
• added the ability to specify negative char and char ranges to lexer
• added a user defined context as a parameter to parser
• added functions shorthand substitutions for SDT placeholders
• added the ability to parse only the longest possible prefix of data, returning the consumed bytes
• added support for a simple form of context-sensitive (non deterministic) parsing by invoking a different lexer/parser pair while scanning