author | wenzelm |
Fri, 05 Jun 2009 21:55:08 +0200 | |
changeset 31469 | 40f815edbde4 |
parent 29606 | fedb8be05f24 |
child 34095 | c2f176a38448 |
permissions | -rw-r--r-- |
26528 | 1 |
(* Title: Pure/General/yxml.ML |
2 |
Author: Makarius |
|
3 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
4 |
Efficient text representation of XML trees using extra characters X |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
5 |
and Y -- no escaping, may nest marked text verbatim. |
26528 | 6 |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
7 |
Markup <elem att="val" ...>...body...</elem> is encoded as: |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
8 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
9 |
X Y name Y att=val ... X |
26528 | 10 |
... |
11 |
body |
|
12 |
... |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
13 |
X Y X |
26528 | 14 |
*) |
15 |
||
16 |
signature YXML = |
|
17 |
sig |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
18 |
val output_markup: Markup.T -> string * string |
26528 | 19 |
val element: string -> XML.attributes -> string list -> string |
20 |
val string_of: XML.tree -> string |
|
21 |
val parse_body: string -> XML.tree list |
|
22 |
val parse: string -> XML.tree |
|
23 |
end; |
|
24 |
||
25 |
structure YXML: YXML = |
|
26 |
struct |
|
27 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
28 |
(** string representation **) |
26528 | 29 |
|
26547 | 30 |
(* markers *) |
31 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
32 |
val X = Symbol.ENQ; |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
33 |
val Y = Symbol.ACK; |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
34 |
val XY = X ^ Y; |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
35 |
val XYX = XY ^ X; |
26528 | 36 |
|
37 |
||
26547 | 38 |
(* output *) |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
39 |
|
27884 | 40 |
fun output_markup (markup as (name, atts)) = |
29325 | 41 |
if Markup.is_none markup then Markup.no_output |
27884 | 42 |
else (XY ^ name ^ implode (map (fn (a, x) => Y ^ a ^ "=" ^ x) atts) ^ X, XYX); |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
43 |
|
26528 | 44 |
fun element name atts body = |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
45 |
let val (pre, post) = output_markup (name, atts) |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
46 |
in pre ^ implode body ^ post end; |
26528 | 47 |
|
48 |
fun string_of t = |
|
49 |
let |
|
50 |
fun attrib (a, x) = |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
51 |
Buffer.add Y #> |
26528 | 52 |
Buffer.add a #> Buffer.add "=" #> Buffer.add x; |
53 |
fun tree (XML.Elem (name, atts, ts)) = |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
54 |
Buffer.add XY #> Buffer.add name #> fold attrib atts #> Buffer.add X #> |
26528 | 55 |
fold tree ts #> |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
56 |
Buffer.add XYX |
28033 | 57 |
| tree (XML.Text s) = Buffer.add s; |
26528 | 58 |
in Buffer.empty |> tree t |> Buffer.content end; |
59 |
||
60 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
61 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
62 |
(** efficient YXML parsing **) |
26528 | 63 |
|
64 |
local |
|
65 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
66 |
(* splitting *) |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
67 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
68 |
fun is_char s c = ord s = Char.ord c; |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
69 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
70 |
val split_string = |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
71 |
Substring.full #> |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
72 |
Substring.tokens (is_char X) #> |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
73 |
map (Substring.fields (is_char Y) #> map Substring.string); |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
74 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
75 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
76 |
(* structural errors *) |
26528 | 77 |
|
78 |
fun err msg = raise Fail ("Malformed YXML encoding: " ^ msg); |
|
79 |
fun err_attribute () = err "bad attribute"; |
|
80 |
fun err_element () = err "bad element"; |
|
81 |
fun err_unbalanced "" = err "unbalanced element" |
|
82 |
| err_unbalanced name = err ("unbalanced element " ^ quote name); |
|
83 |
||
84 |
||
85 |
(* stack operations *) |
|
86 |
||
87 |
fun add x ((elem, body) :: pending) = (elem, x :: body) :: pending; |
|
88 |
||
89 |
fun push "" _ _ = err_element () |
|
90 |
| push name atts pending = ((name, atts), []) :: pending; |
|
91 |
||
92 |
fun pop ((("", _), _) :: _) = err_unbalanced "" |
|
93 |
| pop (((name, atts), body) :: pending) = add (XML.Elem (name, atts, rev body)) pending; |
|
94 |
||
95 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
96 |
(* parsing *) |
26528 | 97 |
|
98 |
fun parse_attrib s = |
|
28025 | 99 |
(case first_field "=" s of |
28023
92dd3ad302b7
simplified parse_attrib (find_substring instead of space_explode);
wenzelm
parents:
27932
diff
changeset
|
100 |
NONE => err_attribute () |
28025 | 101 |
| SOME ("", _) => err_attribute () |
102 |
| SOME att => att); |
|
26528 | 103 |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
104 |
fun parse_chunk ["", ""] = pop |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
105 |
| parse_chunk ("" :: name :: atts) = push name (map parse_attrib atts) |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
106 |
| parse_chunk txts = fold (add o XML.Text) txts; |
26528 | 107 |
|
108 |
in |
|
109 |
||
110 |
fun parse_body source = |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
111 |
(case fold parse_chunk (split_string source) [(("", []), [])] of |
26528 | 112 |
[(("", _), result)] => rev result |
113 |
| ((name, _), _) :: _ => err_unbalanced name); |
|
114 |
||
115 |
fun parse source = |
|
116 |
(case parse_body source of |
|
27798
b96c73f11a23
YXML.parse: allow text without markup, potentially empty;
wenzelm
parents:
26684
diff
changeset
|
117 |
[result] => result |
b96c73f11a23
YXML.parse: allow text without markup, potentially empty;
wenzelm
parents:
26684
diff
changeset
|
118 |
| [] => XML.Text "" |
b96c73f11a23
YXML.parse: allow text without markup, potentially empty;
wenzelm
parents:
26684
diff
changeset
|
119 |
| _ => err "multiple results"); |
26528 | 120 |
|
121 |
end; |
|
122 |
||
123 |
end; |
|
124 |