1 ;;; xml.el --- XML parser
3 ;; Copyright (C) 2000, 2001, 2003 Free Software Foundation, Inc.
5 ;; Author: Emmanuel Briot <briot@gnat.com>
6 ;; Maintainer: Mark A. Hershberger <mah@everybody.org>
9 ;; This file is part of GNU Emacs.
11 ;; GNU Emacs is free software; you can redistribute it and/or modify
12 ;; it under the terms of the GNU General Public License as published by
13 ;; the Free Software Foundation; either version 2, or (at your option)
16 ;; GNU Emacs is distributed in the hope that it will be useful,
17 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
18 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 ;; GNU General Public License for more details.
21 ;; You should have received a copy of the GNU General Public License
22 ;; along with GNU Emacs; see the file COPYING. If not, write to the
23 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
24 ;; Boston, MA 02111-1307, USA.
28 ;; This file contains a somewhat incomplete non-validating XML parser. It
29 ;; parses a file, and returns a list that can be used internally by
30 ;; any other lisp libraries.
34 ;; The document type declaration may either be ignored or (optionally)
35 ;; parsed, but currently the parsing will only accept element
36 ;; declarations. The XML file is assumed to be well-formed. In case
37 ;; of error, the parsing stops and the XML file is shown where the
40 ;; It also knows how to ignore comments and processing instructions.
42 ;; The XML file should have the following format:
43 ;; <node1 attr1="name1" attr2="name2" ...>value
44 ;; <node2 attr3="name3" attr4="name4">value2</node2>
45 ;; <node3 attr5="name5" attr6="name6">value3</node3>
47 ;; Of course, the name of the nodes and attributes can be anything. There can
48 ;; be any number of attributes (or none), as well as any number of children
51 ;; There can be only top level node, but with any number of children below.
55 ;; The functions `xml-parse-file' and `xml-parse-tag' return a list with
56 ;; the following format:
58 ;; xml-list ::= (node node ...)
59 ;; node ::= (tag_name attribute-list . child_node_list)
60 ;; child_node_list ::= child_node child_node ...
61 ;; child_node ::= node | string
62 ;; tag_name ::= string
63 ;; attribute_list ::= (("attribute" . "value") ("attribute" . "value") ...)
67 ;; Some macros are provided to ease the parsing of this list.
68 ;; Whitespace is preserved. Fixme: There should be a tree-walker that
73 ;; Note that {buffer-substring,match-string}-no-properties were
74 ;; formerly used in several places, but that removes composition info.
76 ;;*******************************************************************
78 ;;** Macros to parse the list
80 ;;*******************************************************************
82 (defsubst xml-node-name
(node)
83 "Return the tag associated with NODE.
84 The tag is a lower-case symbol."
87 (defsubst xml-node-attributes
(node)
88 "Return the list of attributes of NODE.
92 (defsubst xml-node-children
(node)
93 "Return the list of children of NODE.
94 This is a list of nodes, and it can be nil."
97 (defun xml-get-children (node child-name
)
98 "Return the children of NODE whose tag is CHILD-NAME.
99 CHILD-NAME should be a lower case symbol."
101 (dolist (child (xml-node-children node
))
103 (if (equal (xml-node-name child
) child-name
)
104 (push child match
))))
107 (defun xml-get-attribute-or-nil (node attribute
)
108 "Get from NODE the value of ATTRIBUTE.
109 Return `nil' if the attribute was not found.
111 See also `xml-get-attribute'."
112 (when (xml-node-attributes node
)
113 (let ((value (assoc attribute
(xml-node-attributes node
))))
117 (defsubst xml-get-attribute
(node attribute
)
118 "Get from NODE the value of ATTRIBUTE.
119 An empty string is returned if the attribute was not found.
121 See also `xml-get-attribute-or-nil'."
122 (or (xml-get-attribute-or-nil node attribute
) ""))
124 ;;*******************************************************************
126 ;;** Creating the list
128 ;;*******************************************************************
131 (defun xml-parse-file (file &optional parse-dtd parse-ns
)
132 "Parse the well-formed XML file FILE.
133 If FILE is already visited, use its buffer and don't kill it.
134 Returns the top node with all its children.
135 If PARSE-DTD is non-nil, the DTD is parsed rather than skipped.
136 If PARSE-NS is non-nil, then QNAMES are expanded."
138 (if (get-file-buffer file
)
140 (set-buffer (get-file-buffer file
))
142 (let (auto-mode-alist) ; no need for xml-mode
145 (let ((xml (xml-parse-region (point-min)
148 parse-dtd parse-ns
)))
151 (kill-buffer (current-buffer)))
154 ;; Note that this is setup so that we can do whitespace-skipping with
155 ;; `(skip-syntax-forward " ")', inter alia. Previously this was slow
156 ;; compared with `re-search-forward', but that has been fixed. Also
157 ;; note that the standard syntax table contains other characters with
158 ;; whitespace syntax, like NBSP, but they are invalid in contexts in
159 ;; which we might skip whitespace -- specifically, they're not
160 ;; NameChars [XML 4].
162 (defvar xml-syntax-table
163 (let ((table (make-syntax-table)))
164 ;; Get space syntax correct per XML [3].
166 (modify-syntax-entry c
"." table
)) ; all are space in standard table
167 (dolist (c '(?
\t ?
\n ?
\r)) ; these should be space
168 (modify-syntax-entry c
" " table
))
169 ;; For skipping attributes.
170 (modify-syntax-entry ?
\" "\"" table
)
171 (modify-syntax-entry ?
' "\"" table
)
172 ;; Non-alnum name chars should be symbol constituents (`-' and `_'
173 ;; are OK by default).
174 (modify-syntax-entry ?.
"_" table
)
175 (modify-syntax-entry ?
: "_" table
)
177 (dolist (c '(#x00B7
#x02D0
#x02D1
#x0387
#x0640
#x0E46
#x0EC6
#x3005
178 #x3031
#x3032
#x3033
#x3034
#x3035
#x309D
#x309E
#x30FC
180 (modify-syntax-entry (decode-char 'ucs c
) "w" table
))
181 ;; Fixme: rest of [4]
183 "Syntax table used by `xml-parse-region'.")
186 ;; Note that [:alpha:] matches all multibyte chars with word syntax.
188 (defconst xml-name-regexp
"[[:alpha:]_:][[:alnum:]._:-]*"))
190 ;; Fixme: This needs re-writing to deal with the XML grammar properly, i.e.
191 ;; document ::= prolog element Misc*
192 ;; prolog ::= XMLDecl? Misc* (doctypedecl Misc*)?
195 (defun xml-parse-region (beg end
&optional buffer parse-dtd parse-ns
)
196 "Parse the region from BEG to END in BUFFER.
197 If BUFFER is nil, it defaults to the current buffer.
198 Returns the XML list for the region, or raises an error if the region
199 is not well-formed XML.
200 If PARSE-DTD is non-nil, the DTD is parsed rather than skipped,
201 and returned as the first element of the list.
202 If PARSE-NS is non-nil, then QNAMES are expanded."
204 (narrow-to-region beg end
)
205 ;; Use fixed syntax table to ensure regexp char classes and syntax
207 (with-syntax-table (standard-syntax-table)
208 (let ((case-fold-search nil
) ; XML is case-sensitive.
213 (goto-char (point-min))
215 (if (search-forward "<" nil t
)
218 (setq result
(xml-parse-tag parse-dtd parse-ns
))
220 ;; translation of rule [1] of XML specifications
221 (error "XML files can have only one toplevel tag")
224 ((and (listp (car result
))
226 (setq dtd
(car result
))
227 (if (cdr result
) ; possible leading comment
228 (add-to-list 'xml
(cdr result
))))
230 (add-to-list 'xml result
)))))
231 (goto-char (point-max))))
233 (cons dtd
(nreverse xml
))
236 (defun xml-ns-parse-ns-attrs (attr-list &optional xml-ns
)
237 "Parse the namespace attributes and return a list of cons in the form:
238 \(namespace . prefix)"
242 (let* ((splitup (split-string (car attr
) ":"))
243 (prefix (nth 0 splitup
))
244 (lname (nth 1 splitup
)))
245 (when (string= "xmlns" prefix
)
246 (push (cons (if lname
250 xml-ns
)))) attr-list
)
253 ;; expand element names
254 (defun xml-ns-expand-el (el xml-ns
)
255 "Expand the XML elements from \"prefix:local-name\" to a cons in the form
256 \"(namespace . local-name)\"."
258 (let* ((splitup (split-string el
":"))
259 (lname (or (nth 1 splitup
)
261 (prefix (if (nth 1 splitup
)
263 (if (string= lname
"xmlns")
266 (ns (cdr (assoc-string prefix xml-ns
))))
269 (cons (intern (concat ":" ns
))
272 ;; expand attribute names
273 (defun xml-ns-expand-attr (attr-list xml-ns
)
274 "Expand the attribute list for a particular element from the form
275 \"prefix:local-name\" to the form \"{namespace}:local-name\"."
279 (let* ((splitup (split-string (car attr
) ":"))
280 (lname (or (nth 1 splitup
)
282 (prefix (if (nth 1 splitup
)
284 (if (string= (car attr
) "xmlns")
287 (ns (cdr (assoc-string prefix xml-ns
))))
291 (cons (intern (concat ":" ns
))
296 (defun xml-intern-attrlist (attr-list)
297 "Convert attribute names to symbols for backward compatibility."
298 (mapcar (lambda (attr)
299 (setcar attr
(intern (car attr
))))
303 (defun xml-parse-tag (&optional parse-dtd parse-ns
)
304 "Parse the tag at point.
305 If PARSE-DTD is non-nil, the DTD of the document, if any, is parsed and
306 returned as the first element in the list.
307 If PARSE-NS is non-nil, then QNAMES are expanded.
309 - a list : the matching node
310 - nil : the point is not looking at a tag.
311 - a pair : the first element is the DTD, the second is the node."
312 (let ((xml-ns (if (consp parse-ns
)
316 ;; Default no namespace
318 ;; We need to seed the xmlns namespace
319 (cons "xmlns" "http://www.w3.org/2000/xmlns/"))))))
321 ;; Processing instructions (like the <?xml version="1.0"?> tag at the
322 ;; beginning of a document).
324 (search-forward "?>")
325 (skip-syntax-forward " ")
326 (xml-parse-tag parse-dtd xml-ns
))
327 ;; Character data (CDATA) sections, in which no tag should be interpreted
328 ((looking-at "<!\\[CDATA\\[")
329 (let ((pos (match-end 0)))
330 (unless (search-forward "]]>" nil t
)
331 (error "CDATA section does not end anywhere in the document"))
332 (buffer-substring pos
(match-beginning 0))))
333 ;; DTD for the document
334 ((looking-at "<!DOCTYPE")
337 (setq dtd
(xml-parse-dtd))
339 (skip-syntax-forward " ")
341 (cons dtd
(xml-parse-tag nil xml-ns
))
342 (xml-parse-tag nil xml-ns
))))
345 (search-forward "-->")
351 ((looking-at "<\\([^/>[:space:]]+\\)")
352 (goto-char (match-end 1))
355 (let* ((node-name (match-string 1))
356 (attr-list (xml-parse-attlist))
357 (children (if (consp xml-ns
) ;; take care of namespace parsing
359 (setq xml-ns
(xml-ns-parse-ns-attrs
361 (list (xml-ns-expand-attr
365 (list (xml-intern-attrlist attr-list
)
366 (intern node-name
))))
369 ;; is this an empty element ?
370 (if (looking-at "/>")
375 ;; is this a valid start tag ?
376 (if (eq (char-after) ?
>)
379 ;; Now check that we have the right end-tag. Note that this
380 ;; one might contain spaces after the tag name
381 (let ((end (concat "</" node-name
"\\s-*>")))
382 (while (not (looking-at end
))
385 (error "XML: Invalid end tag (expecting %s) at pos %d"
388 (let ((tag (xml-parse-tag nil xml-ns
)))
390 (push tag children
))))
395 (let ((string (buffer-substring pos
(point)))
398 ;; Clean up the string. As per XML
399 ;; specifications, the XML processor should
400 ;; always pass the whole string to the
401 ;; application. But \r's should be replaced:
402 ;; http://www.w3.org/TR/2000/REC-xml-20001006#sec-line-ends
403 (while (string-match "\r\n?" string pos
)
404 (setq string
(replace-match "\n" t t string
))
405 (setq pos
(1+ (match-beginning 0))))
407 (setq string
(xml-substitute-special string
))
409 (if (stringp (car children
))
410 ;; The two strings were separated by a comment.
411 (cons (concat (car children
) string
)
413 (cons string children
))))))))
415 (goto-char (match-end 0))
417 ;; This was an invalid start tag
418 (error "XML: Invalid attribute list")))))
419 (t ;; This is not a tag.
420 (error "XML: Invalid character")))))
422 (defun xml-parse-attlist ()
423 "Return the attribute-list after point. Leave point at the
424 first non-blank character after the tag."
427 (skip-syntax-forward " ")
428 (while (looking-at (eval-when-compile
429 (concat "\\(" xml-name-regexp
"\\)\\s-*=\\s-*")))
430 (setq name
(match-string 1))
431 (goto-char (match-end 0))
433 ;; See also: http://www.w3.org/TR/2000/REC-xml-20001006#AVNormalize
435 ;; Do we have a string between quotes (or double-quotes),
436 ;; or a simple word ?
437 (if (looking-at "\"\\([^\"]*\\)\"")
438 (setq end-pos
(match-end 0))
439 (if (looking-at "'\\([^']*\\)'")
440 (setq end-pos
(match-end 0))
441 (error "XML: Attribute values must be given between quotes")))
443 ;; Each attribute must be unique within a given element
444 (if (assoc name attlist
)
445 (error "XML: each attribute must be unique within an element"))
447 ;; Multiple whitespace characters should be replaced with a single one
449 (let ((string (match-string 1))
451 (replace-regexp-in-string "\\s-\\{2,\\}" " " string
)
452 (push (cons name
(xml-substitute-special string
)) attlist
))
455 (skip-syntax-forward " "))
458 ;;*******************************************************************
460 ;;** The DTD (document type declaration)
461 ;;** The following functions know how to skip or parse the DTD of
464 ;;*******************************************************************
466 ;; Fixme: This fails at least if the DTD contains conditional sections.
468 (defun xml-skip-dtd ()
469 "Skip the DTD at point.
470 This follows the rule [28] in the XML specifications."
471 (forward-char (length "<!DOCTYPE"))
472 (if (looking-at "\\s-*>")
473 (error "XML: invalid DTD (excepting name of the document)"))
477 (skip-syntax-forward " ")
478 (if (looking-at "\\[")
479 (re-search-forward "]\\s-*>")
480 (search-forward ">")))
481 (error (error "XML: No end to the DTD"))))
483 (defun xml-parse-dtd ()
484 "Parse the DTD at point."
485 (forward-char (eval-when-compile (length "<!DOCTYPE")))
486 (skip-syntax-forward " ")
488 (error "XML: invalid DTD (excepting name of the document)"))
490 ;; Get the name of the document
491 (looking-at xml-name-regexp
)
492 (let ((dtd (list (match-string 0) 'dtd
))
493 type element end-pos
)
494 (goto-char (match-end 0))
496 (skip-syntax-forward " ")
498 (cond ((looking-at "PUBLIC\\s-+")
499 (goto-char (match-end 0))
500 (unless (or (re-search-forward
501 "\\=\"\\([[:space:][:alnum:]-'()+,./:=?;!*#@$_%]*\\)\""
504 "\\='\\([[:space:][:alnum:]-()+,./:=?;!*#@$_%]*\\)'"
506 (error "XML: missing public id"))
507 (let ((pubid (match-string 1)))
508 (unless (or (re-search-forward "\\='\\([^']*\\)'" nil t
)
509 (re-search-forward "\\=\"\\([^\"]*\\)\"" nil t
))
510 (error "XML: missing system id"))
511 (push (list pubid
(match-string 1) 'public
) dtd
)))
512 ((looking-at "SYSTEM\\s-+")
513 (goto-char (match-end 0))
514 (unless (or (re-search-forward "\\='\\([^']*\\)'" nil t
)
515 (re-search-forward "\\=\"\\([^\"]*\\)\"" nil t
))
516 (error "XML: missing system id"))
517 (push (list (match-string 1) 'system
) dtd
)))
518 (skip-syntax-forward " ")
519 (if (eq ?
> (char-after))
521 (skip-syntax-forward " ")
522 (if (not (eq (char-after) ?\
[))
523 (error "XML: bad DTD")
525 ;; Parse the rest of the DTD
526 ;; Fixme: Deal with ENTITY, ATTLIST, NOTATION, PIs.
527 (while (not (looking-at "\\s-*\\]"))
528 (skip-syntax-forward " ")
531 ;; Translation of rule [45] of XML specifications
533 "<!ELEMENT\\s-+\\([[:alnum:].%;]+\\)\\s-+\\([^>]+\\)>")
535 (setq element
(match-string 1)
536 type
(match-string-no-properties 2))
537 (setq end-pos
(match-end 0))
539 ;; Translation of rule [46] of XML specifications
541 ((string-match "^EMPTY[ \t\n\r]*$" type
) ;; empty declaration
543 ((string-match "^ANY[ \t\n\r]*$" type
) ;; any type of contents
545 ((string-match "^(\\(.*\\))[ \t\n\r]*$" type
) ;; children ([47])
546 (setq type
(xml-parse-elem-type (match-string 1 type
))))
547 ((string-match "^%[^;]+;[ \t\n\r]*$" type
) ;; substitution
550 (error "XML: Invalid element type in the DTD")))
552 ;; rule [45]: the element declaration must be unique
553 (if (assoc element dtd
)
554 (error "XML: element declarations must be unique in a DTD (<%s>)"
557 ;; Store the element in the DTD
558 (push (list element type
) dtd
)
561 (search-forward "-->"))
564 (error "XML: Invalid DTD item")))
566 ;; Skip the end of the DTD
567 (search-forward ">"))))
570 (defun xml-parse-elem-type (string)
571 "Convert element type STRING into a Lisp structure."
574 (if (string-match "(\\([^)]+\\))\\([+*?]?\\)" string
)
576 (setq elem
(match-string 1 string
)
577 modifier
(match-string 2 string
))
578 (if (string-match "|" elem
)
579 (setq elem
(cons 'choice
580 (mapcar 'xml-parse-elem-type
581 (split-string elem
"|"))))
582 (if (string-match "," elem
)
583 (setq elem
(cons 'seq
584 (mapcar 'xml-parse-elem-type
585 (split-string elem
",")))))))
586 (if (string-match "[ \t\n\r]*\\([^+*?]+\\)\\([+*?]?\\)" string
)
587 (setq elem
(match-string 1 string
)
588 modifier
(match-string 2 string
))))
590 (if (and (stringp elem
) (string= elem
"#PCDATA"))
594 ((string= modifier
"+")
596 ((string= modifier
"*")
598 ((string= modifier
"?")
603 ;;*******************************************************************
605 ;;** Substituting special XML sequences
607 ;;*******************************************************************
610 (defvar str
)) ; dynamic from replace-regexp-in-string
612 ;; Fixme: Take declared entities from the DTD when they're available.
613 (defun xml-substitute-entity (match)
614 "Subroutine of xml-substitute-special."
616 (let ((match1 (match-string 1 str
)))
617 (cond ((string= match1
"lt") "<")
618 ((string= match1
"gt") ">")
619 ((string= match1
"apos") "'")
620 ((string= match1
"quot") "\"")
621 ((string= match1
"amp") "&")
622 ((and (string-match "#\\([0-9]+\\)" match1
)
623 (let ((c (decode-char
625 (string-to-number (match-string 1 match1
)))))
626 (if c
(string c
))))) ; else unrepresentable
627 ((and (string-match "#x\\([[:xdigit:]]+\\)" match1
)
628 (let ((c (decode-char
630 (string-to-number (match-string 1 match1
) 16))))
632 ;; Default to asis. Arguably, unrepresentable code points
633 ;; might be best replaced with U+FFFD.
636 (defun xml-substitute-special (string)
637 "Return STRING, after subsituting entity references."
638 ;; This originally made repeated passes through the string from the
639 ;; beginning, which isn't correct, since then either "&amp;" or
640 ;; "&amp;" won't DTRT.
641 (replace-regexp-in-string "&\\([^;]+\\);"
642 #'xml-substitute-entity string t t
))
644 ;;*******************************************************************
646 ;;** Printing a tree.
647 ;;** This function is intended mainly for debugging purposes.
649 ;;*******************************************************************
651 (defun xml-debug-print (xml)
653 (xml-debug-print-internal node
"")))
655 (defun xml-debug-print-internal (xml indent-string
)
656 "Outputs the XML tree in the current buffer.
657 The first line is indented with INDENT-STRING."
660 (insert indent-string ?
< (symbol-name (xml-node-name tree
)))
662 ;; output the attribute list
663 (setq attlist
(xml-node-attributes tree
))
665 (insert ?\
(symbol-name (caar attlist
)) "=\"" (cdar attlist
) ?
\")
666 (setq attlist
(cdr attlist
)))
670 (setq tree
(xml-node-children tree
))
672 ;; output the children
677 (xml-debug-print-internal node
(concat indent-string
" ")))
678 ((stringp node
) (insert node
))
680 (error "Invalid XML tree"))))
682 (insert ?
\n indent-string
683 ?
< ?
/ (symbol-name (xml-node-name xml
)) ?
>)))
687 ;;; arch-tag: 5864b283-5a68-4b59-a20d-36a72b353b9b