The SYNTAX system

The SYNTAX system is a set of tools whose first aim is to ease the design and the implementation of the front-end part of translators -- mainly, but not exclusive, in the compilation field. These tools allow on one side to generate analysers (scanner, parser and semantic) and, on the other side, to compile source texts with these analysers.

Thus the goals of SYNTAX are identical with those of LEX and YACC. But SYNTAX is more powerful, especially when error processing is considered. It possess an automatic (though tunable) error repair and recovery mechanism. Moreover, though semantics can be processed by several means of its own (action, abstract tree, ...) it can be coupled with more sophisticated ones as the FNC-2 attribute grammar system.

The main modules of SYNTAX are:

• BNF: takes as input a Context-Free Grammar (CFG) and produces an internal form used by the other modules. Predicates can be specified in order to act upon the parsing process allowing to define non-deterministic or even context-sensitive (unambiguous) languages.
• CSYNT: checks that the input CFG is LALR(1). Conflicts can be resolved either by the user via a PRIO specification or by the system itself.
• LECL: is the SYNTAX scanner generator. Its input is the lexical specification level in regular expression form. Predefined or user predicates together with an unbounded number of characters of lookahead can be used when needed.
• RECOR: takes as input the lexical and syntactic error processing specifications. A standard input is available.
• TABLES-C: collects the outputs of the previous modules and generates a C program which must be linked with the SYNTAX library.