Module BatParserCo


module BatParserCo: sig .. end
A simple parser combinator library.

This module permits the simple definition of highly modular, dynamic parsers with unlimited backtracking. It may be used to parse any form of enumeration, including regular text, latin-1 text, bits, etc.

This library is vastly more powerful than Lexing, Str, Parsing or Scanf. It is also considerably slower.

Module CharParser contains pre-defined parsers to deal specifically with latin-1 text. Module Genlex contains a number of pre-defined parsers to deal specifically with programming languages.

Note This library is still very rough and needs much testing.



Base definitions


type 'a state =
| Eof (*The end of the source has been reached.*)
| State of 'a
The current state of the parser.

The actual set of states is defined by the user. States are typically used to convey informations, such as position in the file (i.e. line number and character).


type 'a report =
| Report of ('a state * string * 'a report) list (*The final result of parsing*)
module Source: sig .. end
A source for parsing.

Primitives

type ('a, 'b, 'c) t 
A parser for elements of type 'a, producing elements of type 'b, with user-defined states of type 'c.
val eof : ('a, unit, 'b) t
Accept the end of an enumeration.
val either : ('a, 'b, 'c) t list -> ('a, 'b, 'c) t
Accept one of several parsers.
val (<|>) : ('a, 'b, 'c) t ->
('a, 'b, 'c) t -> ('a, 'b, 'c) t
Accept one of two parsers
val maybe : ('a, 'b, 'c) t -> ('a, 'b option, 'c) t
Accept an optional argument.
val (~?) : ('a, 'b, 'c) t -> ('a, 'b option, 'c) t
As maybe
val bind : ('a, 'b, 'c) t ->
('b -> ('a, 'd, 'c) t) -> ('a, 'd, 'c) t
Monadic-style combination:

bind p f results in a new parser which behaves as p then, in case of success, applies f to the result.


Composition of two successive parsers.

compose p q results in a new parser which feeds the consecutive results of p into q. In case of error, positions are taken from p.

val (>>=) : ('a, 'b, 'c) t ->
('b -> ('a, 'd, 'c) t) -> ('a, 'd, 'c) t
As bind
val (>>>) : ('a, 'b, 'c) t ->
('a, 'd, 'c) t -> ('a, 'd, 'c) t
As bind, but ignoring the result
val cons : ('a, 'b, 'c) t ->
('a, 'b list, 'c) t -> ('a, 'b list, 'c) t
cons p q applies parser p then parser q and conses the results into a list.
val (>::) : ('a, 'b, 'c) t ->
('a, 'b list, 'c) t -> ('a, 'b list, 'c) t
As cons
val label : string -> ('a, 'b, 'c) t -> ('a, 'b, 'c) t
Give a name to a parser, for debugging purposes.
val state : ('a, 'b state, 'b) t
Succeed and return the state of the parser
val any : ('a, 'a, 'b) t
Accept any singleton value.
val return : 'a -> ('b, 'a, 'c) t
A parser which always succeds
val satisfy : ('a -> bool) -> ('a, 'a, 'b) t
satisfy p accepts one value p x such that p x = true
val filter : ('a -> bool) -> ('b, 'a, 'c) t -> ('b, 'a, 'c) t
filter f p is only accepts values x such that p accepts x and f (p x) is true
val run : ('a, 'b, 'c) t ->
('a, 'c) Source.t -> ('b, 'c report) BatStd.result
run p s executes parser p on source s. In case of success, returns Ok v, where v is the return value of p. In case of failure, returns Error f, with f containing details on the parsing error.
val fail : ('a, 'b, 'c) t
Always fail, without consuming anything.
val fatal : ('a, 'b, 'c) t
val lookahead : ('a, 'b, 'c) t -> ('a, 'b option, 'c) t
lookahead p behaves as maybe p but without consuming anything

Utilities

val exactly : 'a -> ('a, 'a, 'b) t
Singletons

Accept exactly one singleton.

val one_of : 'a list -> ('a, 'a, 'b) t
Accept one of several values. Faster and more convenient than combining satisfy and either.
val none_of : 'a list -> ('a, 'a, 'b) t
Accept any value not in a list Faster and more convenient than combining satisfy and either.
val range : 'a -> 'a -> ('a, 'a, 'b) t
Accept any element from a given range.
val zero_plus : ?sep:('a, 'b, 'c) t ->
('a, 'd, 'c) t -> ('a, 'd list, 'c) t
Repetitions

Accept a (possibly empty) list of expressions.

val ignore_zero_plus : ?sep:('a, 'b, 'c) t ->
('a, 'd, 'c) t -> ('a, unit, 'c) t
Ignore a (possibly empty) list of expressions. Optimized version of zero_plus, for use when the list of expressions is unimportant.
val ( ~* ) : ('a, 'b, 'c) t -> ('a, 'b list, 'c) t
As zero_plus without arguments.
val one_plus : ?sep:('a, 'b, 'c) t ->
('a, 'd, 'c) t -> ('a, 'd list, 'c) t
Accept a (non-empty) list of expressions
val ignore_one_plus : ?sep:('a, 'b, 'c) t ->
('a, 'd, 'c) t -> ('a, unit, 'c) t
Ignore a (non-empty) list of expressions. Optimized version of one_plus, for use when the list of expressions is unimportant.
val (~+) : ('a, 'b, 'c) t -> ('a, 'b list, 'c) t
As one_plus
val times : int -> ('a, 'b, 'c) t -> ('a, 'b list, 'c) t
times n p accepts a list of n expressions accepted by p
val (^^) : ('a, 'b, 'c) t -> int -> ('a, 'b list, 'c) t
p ^^ n is the same thing as times n p
val must : ('a, 'b, 'c) t -> ('a, 'b, 'c) t
Prevent backtracking.
val should : ('a, 'b, 'c) t -> ('a, 'b, 'c) t
Prevent backtracking.
val post_map : ('a -> 'b) -> ('c, 'a, 'd) t -> ('c, 'b, 'd) t
Maps

Pass the (successful) result of some parser through a map.

val source_map : ('a, 'b, 'c) t ->
('a, 'c) Source.t -> ('b, 'c) Source.t
val scan : ('a, 'b, 'c) t -> ('a, 'a list, 'c) t
Use a parser to extract list of tokens, but return that list of tokens instead of whatever the original parser returned.
val sat : ('a -> bool) -> ('a, unit, 'b) t
Others

satisfy p accepts one value p x such that p x = true

val debug_mode : bool Pervasives.ref
If set to true, debugging information will be printed to the standard error.