1 ;;; nxml-parse.el --- XML parser, sharing infrastructure with nxml-mode
3 ;; Copyright (C) 2003, 2007, 2008, 2009 Free Software Foundation, Inc.
8 ;; This file is part of GNU Emacs.
10 ;; GNU Emacs is free software: you can redistribute it and/or modify
11 ;; it under the terms of the GNU General Public License as published by
12 ;; the Free Software Foundation, either version 3 of the License, or
13 ;; (at your option) any later version.
15 ;; GNU Emacs is distributed in the hope that it will be useful,
16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 ;; GNU General Public License for more details.
20 ;; You should have received a copy of the GNU General Public License
21 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
25 ;; Entry point is `nxml-parse-file'.
34 (defvar nxml-parse-file-name nil
)
36 (defvar nxml-validate-function nil
37 "Either nil or a function called by `nxml-parse-file' to perform validation.
38 The function will be called once for each start-tag or end-tag. The
39 function is passed two arguments TEXT and START-TAG. For a start-tag,
40 START-TAG is a list (NAME ATTRIBUTES) where NAME and ATTRIBUTES are in
41 the same form as returned by `nxml-parse-file'. For an end-tag,
42 START-TAG is nil. TEXT is a string containing the text immediately
43 preceding the tag, or nil if there was no such text. An empty element
44 is treated as a start-tag followed by an end-tag.
46 For a start-tag, the namespace state will be the state after
47 processing the namespace declarations in the start-tag. For an
48 end-tag, the namespace state will be the state before popping the
49 namespace declarations for the corresponding start-tag.
51 The function must return nil if no error is detected or a
52 cons (MESSAGE . LOCATION) where MESSAGE is a string containing
53 an error message and LOCATION indicates what caused the error
56 - nil indicates the tag as whole caused it; this is always allowed;
58 - text indicates the text caused it; this is allowed only if
61 - tag-close indicates the close of the tag caused it; this is
62 allowed only if START-TAG is non-nil;
64 - (attribute-name . N) indicates that the name of the Nth attribute
65 caused it; N counts from 0; this is allowed only if START-TAG is non-nil
66 and N must be less than the number of attributes;
68 - (attribute-value . N) indicates that the value of the Nth attribute
69 caused it; N counts from 0; this is allowed only if START-TAG is non-nil
70 and N must be less than the number of attributes.")
72 (defun nxml-parse-file (file)
73 "Parse the XML document in FILE and return it as a list.
74 An XML element is represented as a list (NAME ATTRIBUTES . CHILDREN).
75 NAME is either a string, in the case where the name does not have a
76 namespace, or a cons (NAMESPACE . LOCAL-NAME), where NAMESPACE is a
77 symbol and LOCAL-NAME is a string, in the case where the name does
78 have a namespace. NAMESPACE is a keyword whose name is `:URI', where
79 URI is the namespace name. ATTRIBUTES is an alist of attributes where
80 each attribute has the form (NAME . VALUE), where NAME has the same
81 form as an element name, and VALUE is a string. A namespace
82 declaration is represented as an attribute whose name is
83 \(:http://www.w3.org/2000/xmlns/ . LOCAL-NAME). CHILDREN is a list
84 containing strings and child elements; CHILDREN never contains two
85 consecutive strings and never contains an empty string. Processing
86 instructions and comments are not represented. The return value is a
87 list representing the document element.
89 If the XML document is not well-formed, an error having the condition
90 `nxml-file-parse-error' will be signaled; the error data will be a
91 list of the form \(FILE POSITION MESSAGE), where POSITION is an
92 integer specifying the position where the error was detected, and
93 MESSAGE is a string describing the error.
95 The current contents of FILE will be parsed even if there is a
96 modified buffer currently visiting FILE.
98 If the variable `nxml-validate-function' is non-nil, it will be called
99 twice for each element, and any reported error will be signaled in the
100 same way as well-formedness error."
102 (set-buffer (nxml-parse-find-file file
))
104 (let ((nxml-parse-file-name file
))
105 (nxml-parse-instance))
108 (defun nxml-parse-find-file (file)
110 (set-buffer (get-buffer-create " *nXML Parse*"))
112 (let ((set-auto-coding-function 'nxml-set-xml-coding
))
113 (insert-file-contents file
))
116 (defun nxml-parse-instance ()
119 (xmltok-forward-prolog)
120 (nxml-check-xmltok-errors)
122 (nxml-parse-instance-1)))))
124 (defun nxml-parse-instance-1 ()
125 (let* ((top (cons nil nil
))
126 ;; tail is a cons cell, whose cdr is nil
127 ;; additional elements will destructively appended to tail
129 ;; stack of tails one for each open element
131 ;; list of QNames of open elements
133 ;; list of strings buffering a text node, in reverse order
135 ;; position of beginning of first (in buffer) string in text
137 (while (xmltok-forward)
138 (nxml-check-xmltok-errors)
139 (cond ((memq xmltok-type
'(start-tag end-tag empty-element
))
141 (setq text
(apply 'concat
(nreverse text
)))
142 (setcdr tail
(cons text nil
))
143 (setq tail
(cdr tail
)))
144 (when (not (eq xmltok-type
'end-tag
))
145 (when (and (not open-element-tags
)
147 (nxml-parse-error nil
"Multiple top-level elements"))
148 (setq open-element-tags
149 (cons (xmltok-start-tag-qname)
152 (let ((tag (nxml-parse-start-tag)))
153 (nxml-validate-tag text text-pos tag
)
155 (setcdr tail
(cons tag nil
))
156 (setq tail
(cdr tail
))
157 (setq tail-stack
(cons tail tail-stack
))
158 (setq tail
(last tag
))))
159 (when (not (eq xmltok-type
'start-tag
))
160 (or (eq xmltok-type
'empty-element
)
161 (equal (car open-element-tags
)
162 (xmltok-end-tag-qname))
163 (if open-element-tags
164 (nxml-parse-error nil
165 "Unbalanced end-tag; expected </%s>"
166 (car open-element-tags
))
167 (nxml-parse-error nil
"Extra end-tag")))
168 (nxml-validate-tag text text-pos nil
)
171 (setq open-element-tags
(cdr open-element-tags
))
172 (setq tail
(car tail-stack
))
173 (setq tail-stack
(cdr tail-stack
)))
175 ((memq xmltok-type
'(space data entity-ref char-ref cdata-section
))
176 (cond (open-element-tags
178 (setq text-pos xmltok-start
))
180 (cons (nxml-current-text-string) text
)))
181 ((not (eq xmltok-type
'space
))
185 (cdr (assq xmltok-type
186 '((data .
"Text characters")
187 (entity-ref .
"Entity reference")
188 (char-ref .
"Character reference")
189 (cdata-section .
"CDATA section"))))))))))
191 (nxml-parse-error (point-max) "Missing document element"))
194 (defun nxml-parse-start-tag ()
195 (let (parsed-attributes
196 parsed-namespace-attributes
197 atts att prefixes prefix ns value name
)
198 (setq atts xmltok-namespace-attributes
)
200 (setq att
(car atts
))
201 (setq value
(or (xmltok-attribute-value att
)
202 (nxml-parse-error nil
"Invalid attribute value")))
203 (setq ns
(nxml-make-namespace value
))
204 (setq prefix
(and (xmltok-attribute-prefix att
)
205 (xmltok-attribute-local-name att
)))
206 (cond ((member prefix prefixes
)
207 (nxml-parse-error nil
"Duplicate namespace declaration"))
209 (nxml-ns-set-default ns
))
211 (nxml-ns-set-prefix prefix ns
))
212 (t (nxml-parse-error nil
"Cannot undeclare namespace prefix")))
213 (setq prefixes
(cons prefix prefixes
))
214 (setq parsed-namespace-attributes
215 (cons (cons (nxml-make-name nxml-xmlns-namespace-uri
216 (xmltok-attribute-local-name att
))
218 parsed-namespace-attributes
))
219 (setq atts
(cdr atts
)))
222 (let ((prefix (xmltok-start-tag-prefix)))
224 (or (nxml-ns-get-prefix prefix
)
225 (nxml-parse-error (1+ xmltok-start
)
226 "Prefix `%s' undeclared"
228 (nxml-ns-get-default)))
229 (xmltok-start-tag-local-name)))
230 (setq atts xmltok-attributes
)
232 (setq att
(car atts
))
234 (let ((prefix (xmltok-attribute-prefix att
)))
236 (or (nxml-ns-get-prefix prefix
)
237 (nxml-parse-error (xmltok-attribute-name-start att
)
238 "Prefix `%s' undeclared"
240 (setq parsed-attributes
241 (let ((nm (nxml-make-name ns
242 (xmltok-attribute-local-name att
))))
243 (when (assoc nm parsed-attributes
)
244 (nxml-parse-error (xmltok-attribute-name-start att
)
245 "Duplicate attribute"))
246 (cons (cons nm
(or (xmltok-attribute-value att
)
247 (nxml-parse-error nil
"Invalid attribute value")))
249 (setq atts
(cdr atts
)))
250 ;; We want to end up with the attributes followed by the
251 ;; the namespace attributes in the same order as
252 ;; xmltok-attributes and xmltok-namespace-attributes respectively.
253 (when parsed-namespace-attributes
254 (setq parsed-attributes
255 (nconc parsed-namespace-attributes parsed-attributes
)))
256 (list name
(nreverse parsed-attributes
))))
258 (defun nxml-validate-tag (text text-pos tag
)
259 (when nxml-validate-function
260 (let ((err (funcall nxml-validate-function text tag
))
263 (setq pos
(nxml-validate-error-position (cdr err
)
266 (or pos
(error "Incorrect return value from %s"
267 nxml-validate-function
))
268 (nxml-parse-error pos
(car err
))))))
270 (defun nxml-validate-error-position (location text-pos tag
)
271 (cond ((null location
) xmltok-start
)
272 ((eq location
'text
) text-pos
)
273 ((eq location
'tag-close
)
274 (and tag
(- (point) (if (eq xmltok-type
'empty-element
) 2 1))))
276 (let ((att (nth (cdr location
) xmltok-attributes
)))
278 (setq att
(nth (- (cdr location
) (length xmltok-attributes
))
279 xmltok-namespace-attributes
)))
281 ((eq (car location
) 'attribute-name
)
282 (xmltok-attribute-name-start att
))
283 ((eq (car location
) 'attribute-value
)
284 (xmltok-attribute-value-start att
)))))))
286 (defun nxml-make-name (ns local-name
)
291 (defun nxml-current-text-string ()
292 (cond ((memq xmltok-type
'(space data
))
293 (buffer-substring-no-properties xmltok-start
295 ((eq xmltok-type
'cdata-section
)
296 (buffer-substring-no-properties (+ xmltok-start
9)
298 ((memq xmltok-type
'(char-ref entity-ref
))
299 (unless xmltok-replacement
300 (nxml-parse-error nil
301 (if (eq xmltok-type
'char-ref
)
302 "Reference to unsupported Unicode character"
303 "Unresolvable entity reference")))
304 xmltok-replacement
)))
306 (defun nxml-parse-error (position &rest args
)
307 (nxml-signal-file-parse-error nxml-parse-file-name
308 (or position xmltok-start
)
309 (apply 'format args
)))
311 (defun nxml-check-xmltok-errors ()
313 (let ((err (car (last xmltok-errors
))))
314 (nxml-signal-file-parse-error nxml-parse-file-name
315 (xmltok-error-start err
)
316 (xmltok-error-message err
)))))
318 (provide 'nxml-parse
)
320 ;; arch-tag: fc19639b-1bff-4673-9992-f539da89ba1e
321 ;;; nxml-parse.el ends here