src/Pure/General/yxml.ML
author wenzelm
Thu, 28 Aug 2008 00:33:09 +0200
changeset 28033 f03b5856f286
parent 28025 d9fcab768496
child 29326 a205acc94356
permissions -rw-r--r--
removed obsolete XML.Output workaround;
wenzelm@26528
     1
(*  Title:      Pure/General/yxml.ML
wenzelm@26528
     2
    ID:         $Id$
wenzelm@26528
     3
    Author:     Makarius
wenzelm@26528
     4
wenzelm@26540
     5
Efficient text representation of XML trees using extra characters X
wenzelm@26540
     6
and Y -- no escaping, may nest marked text verbatim.
wenzelm@26528
     7
wenzelm@26540
     8
Markup <elem att="val" ...>...body...</elem> is encoded as:
wenzelm@26540
     9
wenzelm@26540
    10
  X Y name Y att=val ... X
wenzelm@26528
    11
  ...
wenzelm@26528
    12
  body
wenzelm@26528
    13
  ...
wenzelm@26540
    14
  X Y X
wenzelm@26528
    15
*)
wenzelm@26528
    16
wenzelm@26528
    17
signature YXML =
wenzelm@26528
    18
sig
wenzelm@26528
    19
  val detect: string -> bool
wenzelm@26540
    20
  val output_markup: Markup.T -> string * string
wenzelm@26528
    21
  val element: string -> XML.attributes -> string list -> string
wenzelm@26528
    22
  val string_of: XML.tree -> string
wenzelm@26528
    23
  val parse_body: string -> XML.tree list
wenzelm@26528
    24
  val parse: string -> XML.tree
wenzelm@26528
    25
end;
wenzelm@26528
    26
wenzelm@26528
    27
structure YXML: YXML =
wenzelm@26528
    28
struct
wenzelm@26528
    29
wenzelm@26540
    30
(** string representation **)
wenzelm@26528
    31
wenzelm@26547
    32
(* markers *)
wenzelm@26547
    33
wenzelm@26540
    34
val X = Symbol.ENQ;
wenzelm@26540
    35
val Y = Symbol.ACK;
wenzelm@26540
    36
val XY = X ^ Y;
wenzelm@26540
    37
val XYX = XY ^ X;
wenzelm@26528
    38
wenzelm@26540
    39
val detect = String.isPrefix XY;
wenzelm@26528
    40
wenzelm@26528
    41
wenzelm@26547
    42
(* output *)
wenzelm@26540
    43
wenzelm@27884
    44
fun output_markup (markup as (name, atts)) =
wenzelm@27884
    45
  if Markup.is_none markup then ("", "")
wenzelm@27884
    46
  else (XY ^ name ^ implode (map (fn (a, x) => Y ^ a ^ "=" ^ x) atts) ^ X, XYX);
wenzelm@26540
    47
wenzelm@26528
    48
fun element name atts body =
wenzelm@26540
    49
  let val (pre, post) = output_markup (name, atts)
wenzelm@26540
    50
  in pre ^ implode body ^ post end;
wenzelm@26528
    51
wenzelm@26528
    52
fun string_of t =
wenzelm@26528
    53
  let
wenzelm@26528
    54
    fun attrib (a, x) =
wenzelm@26540
    55
      Buffer.add Y #>
wenzelm@26528
    56
      Buffer.add a #> Buffer.add "=" #> Buffer.add x;
wenzelm@26528
    57
    fun tree (XML.Elem (name, atts, ts)) =
wenzelm@26540
    58
          Buffer.add XY #> Buffer.add name #> fold attrib atts #> Buffer.add X #>
wenzelm@26528
    59
          fold tree ts #>
wenzelm@26540
    60
          Buffer.add XYX
wenzelm@28033
    61
      | tree (XML.Text s) = Buffer.add s;
wenzelm@26528
    62
  in Buffer.empty |> tree t |> Buffer.content end;
wenzelm@26528
    63
wenzelm@26528
    64
wenzelm@26540
    65
wenzelm@26540
    66
(** efficient YXML parsing **)
wenzelm@26528
    67
wenzelm@26528
    68
local
wenzelm@26528
    69
wenzelm@26540
    70
(* splitting *)
wenzelm@26540
    71
wenzelm@26540
    72
fun is_char s c = ord s = Char.ord c;
wenzelm@26540
    73
wenzelm@26540
    74
val split_string =
wenzelm@26540
    75
  Substring.full #>
wenzelm@26540
    76
  Substring.tokens (is_char X) #>
wenzelm@26540
    77
  map (Substring.fields (is_char Y) #> map Substring.string);
wenzelm@26540
    78
wenzelm@26540
    79
wenzelm@26540
    80
(* structural errors *)
wenzelm@26528
    81
wenzelm@26528
    82
fun err msg = raise Fail ("Malformed YXML encoding: " ^ msg);
wenzelm@26528
    83
fun err_attribute () = err "bad attribute";
wenzelm@26528
    84
fun err_element () = err "bad element";
wenzelm@26528
    85
fun err_unbalanced "" = err "unbalanced element"
wenzelm@26528
    86
  | err_unbalanced name = err ("unbalanced element " ^ quote name);
wenzelm@26528
    87
wenzelm@26528
    88
wenzelm@26528
    89
(* stack operations *)
wenzelm@26528
    90
wenzelm@26528
    91
fun add x ((elem, body) :: pending) = (elem, x :: body) :: pending;
wenzelm@26528
    92
wenzelm@26528
    93
fun push "" _ _ = err_element ()
wenzelm@26528
    94
  | push name atts pending = ((name, atts), []) :: pending;
wenzelm@26528
    95
wenzelm@26528
    96
fun pop ((("", _), _) :: _) = err_unbalanced ""
wenzelm@26528
    97
  | pop (((name, atts), body) :: pending) = add (XML.Elem (name, atts, rev body)) pending;
wenzelm@26528
    98
wenzelm@26528
    99
wenzelm@26540
   100
(* parsing *)
wenzelm@26528
   101
wenzelm@26528
   102
fun parse_attrib s =
wenzelm@28025
   103
  (case first_field "=" s of
wenzelm@28023
   104
    NONE => err_attribute ()
wenzelm@28025
   105
  | SOME ("", _) => err_attribute ()
wenzelm@28025
   106
  | SOME att => att);
wenzelm@26528
   107
wenzelm@26540
   108
fun parse_chunk ["", ""] = pop
wenzelm@26540
   109
  | parse_chunk ("" :: name :: atts) = push name (map parse_attrib atts)
wenzelm@26540
   110
  | parse_chunk txts = fold (add o XML.Text) txts;
wenzelm@26528
   111
wenzelm@26528
   112
in
wenzelm@26528
   113
wenzelm@26528
   114
fun parse_body source =
wenzelm@26540
   115
  (case fold parse_chunk (split_string source) [(("", []), [])] of
wenzelm@26528
   116
    [(("", _), result)] => rev result
wenzelm@26528
   117
  | ((name, _), _) :: _ => err_unbalanced name);
wenzelm@26528
   118
wenzelm@26528
   119
fun parse source =
wenzelm@26528
   120
  (case parse_body source of
wenzelm@27798
   121
    [result] => result
wenzelm@27798
   122
  | [] => XML.Text ""
wenzelm@27798
   123
  | _ => err "multiple results");
wenzelm@26528
   124
wenzelm@26528
   125
end;
wenzelm@26528
   126
wenzelm@26528
   127
end;
wenzelm@26528
   128