---

6.4 Lexer & Parsing reference ¶

We have chosen Haskell to implement the lexing and parsing of the language, as functional languages like Haskell make it particularly easy to express, abstract, and combine parsers. This allows the Franc C to be very verbose, which was definitely needed.

As we wanted to speed things up and release a first version as quickly as possible, we decided to use a pre-existing parsing library ‘Megaparsec’. The library offers fast parsing performance, is an improved successor of ‘Parsec’, and provides high-quality error messages that greatly simplify debugging and development.

The ‘Lexer’ and ‘Parser’ both refers to two different parts:

• The ‘Lexer’, which uses ‘Megaparsec’ reads the source code’s syntax, checks that it follows the correct grammar, then converts the text into a list of tokens that the parser can later use. For example, the code ‘Si x est égale à 0, exécute le texte’ would be transformed into the list as ‘[If, Symbol "x", Operation Equals, Number 0, Then]’
• The ‘Parser’, on the other hand, takes this list of tokens to transform it into an Abstract Syntax Tree (AST). It uses Haskell’s recursive principles to build that tree so that the compiler can later use it to generate the final executable library.

The ‘Parser’ is only executed if the ‘Lexer’ did not fail to lex the targetted source code.

• Safety measures
• Error list
• Known caveats