Module Bitstring_persistent

module Bitstring_persistent: sig .. end

Jump straight to the reference section for documentation on types and functions.

Introduction

Bitstring allows you to name sets of fields and reuse them elsewhere. For example if you frequently need to parse Pascal-style strings in the form length byte + string, then you could name the { strlen : 8 : int; str : strlen*8 : string } pattern and reuse it everywhere by name.

These are called persistent patterns.

The basic usage is:

(* Create a persistent pattern called 'pascal_string' which
 matches Pascal-style strings (length byte + string).
)
let bitmatch pascal_string =
  { strlen : 8 : int;
    str : strlen*8 : string }

let is_pascal_string bits =
  bitmatch bits with
  | { :pascal_string } ->
    printf "matches a Pascal string %s, len %d bytes\n"
      str strlen

or:

(* Load a persistent pattern from a file. *)
open bitmatch "pascal.bmpp"

let is_pascal_string bits =
  bitmatch bits with
  | { :pascal_string } ->
    printf "matches a Pascal string %s, len %d bytes\n"
      str strlen

Important notes

There are some important things you should know about persistent patterns before you decide to use them:

'Persistent' refers to the fact that they can be saved into binary files. However these binary files use OCaml Marshal module and depend (sometimes) on the version of OCaml used to generate them and (sometimes) the version of bitstring used. So your build system should rebuild these files from source when your code is rebuilt.

Persistent patterns are syntactic. They work in the same way as cutting and pasting (or #include-ing) code. For example if a persistent pattern binds a field named len, then any uses of len following in the surrounding pattern could be affected.

Programs which generate and manipulate persistent patterns have to link to camlp4. Since camlp4 in OCaml >= 3.10 is rather large, we have placed this code into this separate submodule, so that programs which just use bitstring don't need to pull in the whole of camlp4. This restriction does not apply to code which only uses persistent patterns but does not generate them. If the distinction isn't clear, use ocamlobjinfo to look at the dependencies of your *.cmo files.

Persistent patterns can be generated in several ways, but they can only be used by the pa_bitstring syntax extension. This means they are purely compile-time constructs. You cannot use them to make arbitrary patterns and run those patterns (not unless your program runs ocamlc to make a *.cmo file then dynamically links to the *.cmo file).

Named patterns

A named pattern is a way to name a pattern and use it later in the same source file. To name a pattern, use:

let bitmatch name = { fields ... } ;;

and you can then use the name later on inside another pattern, by prefixing the name with a colon. For example:

bitmatch bits with { :name } -> ...

You can nest named patterns within named patterns to any depth.

Currently the use of named patterns is somewhat limited. The restrictions are:

Named patterns can only be used within the same source file, and the names occupy a completely separate namespace from anything else in the source file.

The let bitmatch syntax only works at the top level. We may add a let bitmatch ... in for inner levels later.

Because you cannot rename the bound identifiers in named patterns, you can effectively only use them once in a pattern. For example, { :name; :name } is legal, but any bindings in the first name would be overridden by the second name.

There are no "named constructors" yet, but the machinery is in place to do this, and we may add them later.

Persistent patterns in files

More useful than just naming patterns, you can load persistent patterns from external files. The patterns in these external files can come from a variety of sources: for example, in the cil-tools subdirectory are some Cil-based tools for importing C structures from header files. You can also generate your own files or write your own tools, as described below.

To use the persistent pattern(s) from a file do:

open bitmatch "filename.bmpp" ;;

A list of zero or more Bitstring_persistent.named patterns are read from the file and each is bound to a name (as contained in the file), and then the patterns can be used with the usual :name syntax described above.

Extension

The standard extension is .bmpp. This is just a convention and you can use any extension you want.

Directory search order

If the filename is an absolute or explicit path, then we try to load it from that path and stop if it fails. See the Filename module in the standard OCaml library for the definitions of "absolute path" and "explicit path". Otherwise we use the following directory search order:

• Relative to the current directory
• Relative to the OCaml library directory

bitstring-objinfo

The bitstring-objinfo command can be run on a file in order to print out the patterns in the file.

Constructors

We haven't implemented persistent constructors yet, although the machinery is in place to make this happen. Any constructors found in the file are ignored.

Creating your own persistent patterns

If you want to write a tool to import bitstrings from an exotic location or markup language, you will need to use the functions found in the reference section.

I will describe using an example here of how you would programmatically create a persistent pattern which matches Pascal-style "length byte + data" strings. Firstly note that there are two fields, so our pattern will be a list of length 2 and type Bitstring_persistent.pattern.

You will need to create a camlp4 location object (Loc.t) describing the source file. This source file is used to generate useful error messages for the user, so you may want to set it to be the name and location in the file that your tool reads for input. By convention, locations are bound to name _loc:

   let _loc = Loc.move_line 42 (Loc.mk "input.xml")
  

Create a pattern field representing a length field which is 8 bits wide, bound to the identifier len:

   let len_field = create_pattern_field _loc
   let len_field = set_length_int len_field 8
   let len_field = set_lident_patt len_field "len"
  

Create a pattern field representing a string of len*8 bits. Note that the use of <:expr< >> quotation requires you to preprocess your source with camlp4of (see this page on Reflective OCaml).

   let str_field = create_pattern_field _loc
   let str_field = set_length str_field <:expr< len*8 >>
   let str_field = set_lident_patt str_field "str"
   let str_field = set_type_string str_field
  

Join the two fields together and name it:

   let pattern = [len_field; str_field]
   let named_pattern = "pascal_string", Pattern pattern
  

Save it to a file:

   let chan = open_out "output.bmpp" in
   named_to_channel chan named_pattern;
   close_out chan
  

You can now use this pattern in another program like this:

   open bitmatch "output.bmpp" ;;
   let parse_pascal_string bits =
   bitmatch bits with
   | { :pascal_string } -> str, len
   | { _ } -> invalid_arg "not a Pascal string"
  

You can write more than one named pattern to the output file, and they will all be loaded at the same time by open bitmatch ".." (obviously you should give each pattern a different name). To do this, just call Bitstring_persistent.named_to_channel as many times as needed.

Reference

Types

type patt = Camlp4.PreCast.Syntax.Ast.patt 

type expr = Camlp4.PreCast.Syntax.Ast.expr 

type loc_t = Camlp4.PreCast.Syntax.Ast.Loc.t 

type 'a field 

type pattern = patt field list 

type constructor = expr field list 

type named = string * alt 

type alt = 

Printers

val string_of_pattern : pattern -> string

val string_of_constructor : constructor -> string

val string_of_pattern_field : patt field -> string

val string_of_constructor_field : expr field -> string

Persistence

val named_to_channel : Pervasives.out_channel -> named -> unit

val named_to_string : named -> string

val named_to_buffer : string -> int -> int -> named -> int

val named_from_channel : Pervasives.in_channel -> named

val named_from_string : string -> int -> named

Create pattern fields

These fields are used in pattern matches (bitmatch).

val create_pattern_field : loc_t ->
       patt field

    let field = create_pattern_field _loc in
    let field = set_lident_patt field "len" in
    let field = set_length_int field 8 in

val set_lident_patt : patt field ->
       string -> patt field

val set_int_patt : patt field ->
       int -> patt field

val set_string_patt : patt field ->
       string -> patt field

val set_unbound_patt : patt field ->
       patt field

val set_patt : patt field ->
       patt ->
       patt field

val set_length_int : 'a field -> int -> 'a field

val set_length : 'a field ->
       expr -> 'a field

val set_endian : 'a field ->
       Bitstring.endian -> 'a field

val set_endian_expr : 'a field ->
       expr -> 'a field

val set_signed : 'a field -> bool -> 'a field

val set_type_int : 'a field -> 'a field

val set_type_string : 'a field -> 'a field

val set_type_bitstring : 'a field -> 'a field

val set_location : 'a field ->
       loc_t -> 'a field

val set_offset_int : 'a field -> int -> 'a field

val set_offset : 'a field ->
       expr -> 'a field

val set_no_offset : 'a field -> 'a field

val set_check : 'a field ->
       expr -> 'a field

val set_no_check : 'a field -> 'a field

val set_bind : 'a field ->
       expr -> 'a field

val set_no_bind : 'a field -> 'a field

val set_save_offset_to : 'a field ->
       patt -> 'a field

val set_save_offset_to_lident : 'a field -> string -> 'a field

val set_no_save_offset_to : 'a field -> 'a field

Create constructor fields

These fields are used in constructors (BITSTRING).

val create_constructor_field : loc_t ->
       expr field

val set_lident_expr : expr field ->
       string -> expr field

val set_int_expr : expr field ->
       int -> expr field

val set_string_expr : expr field ->
       string -> expr field

val set_expr : expr field ->
       expr ->
       expr field

Accessors

val get_patt : patt field ->
       patt

val get_expr : expr field ->
       expr

val get_length : 'a field -> expr

type endian_expr = 

val get_endian : 'a field -> endian_expr

val get_signed : 'a field -> bool

type field_type = 

val get_type : 'a field -> field_type

val get_location : 'a field -> loc_t

val get_offset : 'a field -> expr option

val get_check : 'a field -> expr option

val get_bind : 'a field -> expr option

val get_save_offset_to : 'a field -> patt option