module Genlex: sig endThis module implements a simple ``standard'' lexical analyzer, presented as a function from character streams to token streams. It implements roughly the lexical conventions of Caml, but is parameterized by the set of keywords of your language.
Example: a lexer suitable for a desk calculator is obtained by
let lexer = make_lexer ["+";"-";"*";"/";"let";"="; "("; ")"]
The associated parser would be a function from token stream
to, for instance, int, and would have rules such as:
let parse_expr = parser
[< 'Int n >] -> n
| [< 'Kwd "("; n = parse_expr; 'Kwd ")" >] -> n
| [< n1 = parse_expr; n2 = parse_remainder n1 >] -> n2
and parse_remainder n1 = parser
[< 'Kwd "+"; n2 = parse_expr >] -> n1+n2
| ...
type token =
| |
Kwd of |
| |
Ident of |
| |
Int of |
| |
Float of |
| |
String of |
| |
Char of |
Int and Float
for integer and floating-point numbers; String for
string literals, enclosed in double quotes; Char for
character literals, enclosed in single quotes; Ident for
identifiers (either sequences of letters, digits, underscores
and quotes, or sequences of ``operator characters'' such as
+, *, etc); and Kwd for keywords (either identifiers or
single ``special characters'' such as (, }, etc).val make_lexer : string list -> char Stream.t -> token Stream.ts is returned as Kwd s if s
belongs to this list, and as Ident s otherwise.
A special character s is returned as Kwd s if s
belongs to this list, and cause a lexical error (exception
Parse_error) otherwise. Blanks and newlines are skipped.
Comments delimited by (* and *) are skipped as well,
and can be nested.