author | wenzelm |
Mon, 24 Oct 2016 15:00:13 +0200 | |
changeset 64375 | 74a2af7c5145 |
parent 64275 | ac2abc987cf9 |
child 68228 | 326f4bcc5abc |
permissions | -rw-r--r-- |
44698 | 1 |
(* Title: Pure/PIDE/yxml.ML |
26528 | 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 |
44698 | 5 |
and Y -- no escaping, may nest marked text verbatim. Suitable for |
6 |
direct inlining into plain text. |
|
26528 | 7 |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
8 |
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
|
9 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
10 |
X Y name Y att=val ... X |
26528 | 11 |
... |
12 |
body |
|
13 |
... |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
14 |
X Y X |
26528 | 15 |
*) |
16 |
||
17 |
signature YXML = |
|
18 |
sig |
|
43777 | 19 |
val X: Symbol.symbol |
20 |
val Y: Symbol.symbol |
|
43772
c825594fd0c1
clarified YXML.embed_controls -- this is idempotent and cannot be nested;
wenzelm
parents:
43731
diff
changeset
|
21 |
val embed_controls: string -> string |
43731
70072780e095
inner syntax supports inlined YXML according to Term_XML (particularly useful for producing text under program control);
wenzelm
parents:
43728
diff
changeset
|
22 |
val detect: string -> bool |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
23 |
val output_markup: Markup.T -> string * string |
43728 | 24 |
val string_of_body: XML.body -> string |
26528 | 25 |
val string_of: XML.tree -> string |
49656
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
26 |
val output_markup_elem: Markup.T -> (string * string) * string |
38266
492d377ecfe2
type XML.body as basic data representation language;
wenzelm
parents:
38265
diff
changeset
|
27 |
val parse_body: string -> XML.body |
26528 | 28 |
val parse: string -> XML.tree |
59433 | 29 |
val content_of: string -> string |
26528 | 30 |
end; |
31 |
||
32 |
structure YXML: YXML = |
|
33 |
struct |
|
34 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
35 |
(** string representation **) |
26528 | 36 |
|
38265
cc9fde54311f
renamed YXML.binary_text to YXML.escape_controls to emphasize what it actually does;
wenzelm
parents:
38228
diff
changeset
|
37 |
(* idempotent recoding of certain low ASCII control characters *) |
34095
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
38 |
|
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
39 |
fun pseudo_utf8 c = |
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
40 |
if Symbol.is_ascii_control c |
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
41 |
then chr 192 ^ chr (128 + ord c) |
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
42 |
else c; |
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
43 |
|
43772
c825594fd0c1
clarified YXML.embed_controls -- this is idempotent and cannot be nested;
wenzelm
parents:
43731
diff
changeset
|
44 |
fun embed_controls str = |
34095
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
45 |
if exists_string Symbol.is_ascii_control str |
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
46 |
then translate_string pseudo_utf8 str |
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
47 |
else str; |
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
48 |
|
c2f176a38448
robust representation of low ASCII control characters within XML/YXML text;
wenzelm
parents:
31469
diff
changeset
|
49 |
|
26547 | 50 |
(* markers *) |
51 |
||
64275
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
52 |
val X_char = Char.chr 5; |
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
53 |
val Y_char = Char.chr 6; |
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
54 |
|
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
55 |
val X = String.str X_char; |
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
56 |
val Y = String.str Y_char; |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
57 |
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
|
58 |
val XYX = XY ^ X; |
26528 | 59 |
|
64275
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
60 |
fun detect s = Char.contains s X_char orelse Char.contains s Y_char; |
43731
70072780e095
inner syntax supports inlined YXML according to Term_XML (particularly useful for producing text under program control);
wenzelm
parents:
43728
diff
changeset
|
61 |
|
26528 | 62 |
|
26547 | 63 |
(* output *) |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
64 |
|
27884 | 65 |
fun output_markup (markup as (name, atts)) = |
38474
e498dc2eb576
uniform Markup.empty/Markup.Empty in ML and Scala;
wenzelm
parents:
38266
diff
changeset
|
66 |
if Markup.is_empty markup then Markup.no_output |
27884 | 67 |
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
|
68 |
|
43728 | 69 |
fun string_of_body body = |
26528 | 70 |
let |
46831 | 71 |
fun attrib (a, x) = Buffer.add Y #> Buffer.add a #> Buffer.add "=" #> Buffer.add x; |
38228
ada3ab6b9085
simplified type XML.tree: embed Markup.T directly, avoid slightly odd triple;
wenzelm
parents:
34095
diff
changeset
|
72 |
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
|
73 |
Buffer.add XY #> Buffer.add name #> fold attrib atts #> Buffer.add X #> |
43728 | 74 |
trees ts #> |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
75 |
Buffer.add XYX |
43728 | 76 |
| tree (XML.Text s) = Buffer.add s |
77 |
and trees ts = fold tree ts; |
|
78 |
in Buffer.empty |> trees body |> Buffer.content end; |
|
79 |
||
80 |
val string_of = string_of_body o single; |
|
26528 | 81 |
|
82 |
||
49656
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
83 |
(* wrapped elements *) |
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
84 |
|
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
85 |
val Z = chr 0; |
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
86 |
val Z_text = [XML.Text Z]; |
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
87 |
|
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
88 |
fun output_markup_elem markup = |
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
89 |
let val [bg1, bg2, en] = space_explode Z (string_of (XML.wrap_elem ((markup, Z_text), Z_text))) |
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
90 |
in ((bg1, bg2), en) end; |
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
91 |
|
7ff712de5747
treat wrapped markup elements as raw markup delimiters;
wenzelm
parents:
46832
diff
changeset
|
92 |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
93 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
94 |
(** efficient YXML parsing **) |
26528 | 95 |
|
96 |
local |
|
97 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
98 |
(* splitting *) |
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
99 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
100 |
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
|
101 |
Substring.full #> |
64275
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
102 |
Substring.tokens (fn c => c = X_char) #> |
ac2abc987cf9
accomodate Poly/ML repository version, which treats singleton strings as boxed;
wenzelm
parents:
59433
diff
changeset
|
103 |
map (Substring.fields (fn c => c = Y_char) #> map Substring.string); |
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
104 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
105 |
|
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
106 |
(* structural errors *) |
26528 | 107 |
|
46832 | 108 |
fun err msg = raise Fail ("Malformed YXML: " ^ msg); |
26528 | 109 |
fun err_attribute () = err "bad attribute"; |
110 |
fun err_element () = err "bad element"; |
|
111 |
fun err_unbalanced "" = err "unbalanced element" |
|
112 |
| err_unbalanced name = err ("unbalanced element " ^ quote name); |
|
113 |
||
114 |
||
115 |
(* stack operations *) |
|
116 |
||
117 |
fun add x ((elem, body) :: pending) = (elem, x :: body) :: pending; |
|
118 |
||
119 |
fun push "" _ _ = err_element () |
|
120 |
| push name atts pending = ((name, atts), []) :: pending; |
|
121 |
||
122 |
fun pop ((("", _), _) :: _) = err_unbalanced "" |
|
38228
ada3ab6b9085
simplified type XML.tree: embed Markup.T directly, avoid slightly odd triple;
wenzelm
parents:
34095
diff
changeset
|
123 |
| pop ((markup, body) :: pending) = add (XML.Elem (markup, rev body)) pending; |
26528 | 124 |
|
125 |
||
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
126 |
(* parsing *) |
26528 | 127 |
|
128 |
fun parse_attrib s = |
|
28025 | 129 |
(case first_field "=" s of |
28023
92dd3ad302b7
simplified parse_attrib (find_substring instead of space_explode);
wenzelm
parents:
27932
diff
changeset
|
130 |
NONE => err_attribute () |
28025 | 131 |
| SOME ("", _) => err_attribute () |
132 |
| SOME att => att); |
|
26528 | 133 |
|
26540
173d548ce9d2
replaced ETX/EOT by ENQ/ACK, which are less likely to be interpreted by tty etc.;
wenzelm
parents:
26528
diff
changeset
|
134 |
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
|
135 |
| 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
|
136 |
| parse_chunk txts = fold (add o XML.Text) txts; |
26528 | 137 |
|
138 |
in |
|
139 |
||
140 |
fun parse_body source = |
|
43615
8e0f6cfa8eb2
reverted ce00462f,b3759dce, 7a165592: unwanted generalisation
noschinl
parents:
42355
diff
changeset
|
141 |
(case fold parse_chunk (split_string source) [(("", []), [])] of |
26528 | 142 |
[(("", _), result)] => rev result |
143 |
| ((name, _), _) :: _ => err_unbalanced name); |
|
144 |
||
145 |
fun parse source = |
|
146 |
(case parse_body source of |
|
27798
b96c73f11a23
YXML.parse: allow text without markup, potentially empty;
wenzelm
parents:
26684
diff
changeset
|
147 |
[result] => result |
b96c73f11a23
YXML.parse: allow text without markup, potentially empty;
wenzelm
parents:
26684
diff
changeset
|
148 |
| [] => XML.Text "" |
b96c73f11a23
YXML.parse: allow text without markup, potentially empty;
wenzelm
parents:
26684
diff
changeset
|
149 |
| _ => err "multiple results"); |
26528 | 150 |
|
151 |
end; |
|
152 |
||
59433 | 153 |
val content_of = parse_body #> XML.content_of; |
154 |
||
26528 | 155 |
end; |
156 |