lisp/xml.el

   1 ;;; xml.el --- XML parser
   2
   3 ;; Copyright (C) 2000, 2001, 2003 Free Software Foundation, Inc.
   4
   5 ;; Author: Emmanuel Briot  <briot@gnat.com>
   6 ;; Maintainer: Mark A. Hershberger <mah@everybody.org>
   7 ;; Keywords: xml, data
   8
   9 ;; This file is part of GNU Emacs.
  10
  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)
  14 ;; any later version.
  15
  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.
  20
  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.
  25
  26 ;;; Commentary:
  27
  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.
  31
  32 ;;; FILE FORMAT
  33
  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
  38 ;; parsing stopped.
  39 ;;
  40 ;; It also knows how to ignore comments and processing instructions.
  41 ;;
  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>
  46 ;;    </node1>
  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
  49 ;; below the nodes.
  50 ;;
  51 ;; There can be only top level node, but with any number of children below.
  52
  53 ;;; LIST FORMAT
  54
  55 ;; The functions `xml-parse-file' and `xml-parse-tag' return a list with
  56 ;; the following format:
  57 ;;
  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") ...)
  64 ;;                       | nil
  65 ;;    string     ::= "..."
  66 ;;
  67 ;; Some macros are provided to ease the parsing of this list.
  68 ;; Whitespace is preserved.  Fixme: There should be a tree-walker that
  69 ;; can remove it.
  70
  71 ;;; Code:
  72
  73 ;; Note that {buffer-substring,match-string}-no-properties were
  74 ;; formerly used in several places, but that removes composition info.
  75
  76 ;;*******************************************************************
  77 ;;**
  78 ;;**  Macros to parse the list
  79 ;;**
  80 ;;*******************************************************************
  81
  82 (defsubst xml-node-name (node)
  83   "Return the tag associated with NODE.
  84 The tag is a lower-case symbol."
  85   (car node))
  86
  87 (defsubst xml-node-attributes (node)
  88   "Return the list of attributes of NODE.
  89 The list can be nil."
  90   (nth 1 node))
  91
  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."
  95   (cddr node))
  96
  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."
 100   (let ((match ()))
 101     (dolist (child (xml-node-children node))
 102       (if child
 103           (if (equal (xml-node-name child) child-name)
 104               (push child match))))
 105     (nreverse match)))
 106
 107 (defun xml-get-attribute (node attribute)
 108   "Get from NODE the value of ATTRIBUTE.
 109 An empty string is returned if the attribute was not found."
 110   (if (xml-node-attributes node)
 111       (let ((value (assoc attribute (xml-node-attributes node))))
 112         (if value
 113             (cdr value)
 114           ""))
 115     ""))
 116
 117 ;;*******************************************************************
 118 ;;**
 119 ;;**  Creating the list
 120 ;;**
 121 ;;*******************************************************************
 122
 123 ;;;###autoload
 124 (defun xml-parse-file (file &optional parse-dtd parse-ns)
 125   "Parse the well-formed XML file FILE.
 126 If FILE is already visited, use its buffer and don't kill it.
 127 Returns the top node with all its children.
 128 If PARSE-DTD is non-nil, the DTD is parsed rather than skipped.
 129 If PARSE-NS is non-nil, then QNAMES are expanded."
 130   (let ((keep))
 131     (if (get-file-buffer file)
 132         (progn
 133           (set-buffer (get-file-buffer file))
 134           (setq keep (point)))
 135       (let (auto-mode-alist)            ; no need for xml-mode
 136         (find-file file)))
 137
 138     (let ((xml (xml-parse-region (point-min)
 139                                  (point-max)
 140                                  (current-buffer)
 141                                  parse-dtd parse-ns)))
 142       (if keep
 143           (goto-char keep)
 144         (kill-buffer (current-buffer)))
 145       xml)))
 146
 147 ;; Note that this is setup so that we can do whitespace-skipping with
 148 ;; `(skip-syntax-forward " ")', inter alia.  Previously this was slow
 149 ;; compared with `re-search-forward', but that has been fixed.  Also
 150 ;; note that the standard syntax table contains other characters with
 151 ;; whitespace syntax, like NBSP, but they are invalid in contexts in
 152 ;; which we might skip whitespace -- specifically, they're not
 153 ;; NameChars [XML 4].
 154
 155 (defvar xml-syntax-table
 156   (let ((table (make-syntax-table)))
 157     ;; Get space syntax correct per XML [3].
 158     (dotimes (c 31)
 159       (modify-syntax-entry c "." table)) ; all are space in standard table
 160     (dolist (c '(?\t ?\n ?\r))          ; these should be space
 161       (modify-syntax-entry c " " table))
 162     ;; For skipping attributes.
 163     (modify-syntax-entry ?\" "\"" table)
 164     (modify-syntax-entry ?' "\"" table)
 165     ;; Non-alnum name chars should be symbol constituents (`-' and `_'
 166     ;; are OK by default).
 167     (modify-syntax-entry ?. "_" table)
 168     (modify-syntax-entry ?: "_" table)
 169     ;; XML [89]
 170     (dolist (c '(#x00B7 #x02D0 #x02D1 #x0387 #x0640 #x0E46 #x0EC6 #x3005
 171                  #x3031 #x3032 #x3033 #x3034 #x3035 #x309D #x309E #x30FC
 172                  #x30FD #x30FE))
 173       (modify-syntax-entry (decode-char 'ucs c) "w" table))
 174     ;; Fixme: rest of [4]
 175     table)
 176   "Syntax table used by `xml-parse-region'.")
 177
 178 ;; XML [5]
 179 ;; Note that [:alpha:] matches all multibyte chars with word syntax.
 180 (eval-and-compile
 181   (defconst xml-name-regexp "[[:alpha:]_:][[:alnum:]._:-]*"))
 182
 183 ;; Fixme:  This needs re-writing to deal with the XML grammar properly, i.e.
 184 ;;   document    ::=    prolog element Misc*
 185 ;;   prolog    ::=    XMLDecl? Misc* (doctypedecl Misc*)?
 186
 187 ;;;###autoload
 188 (defun xml-parse-region (beg end &optional buffer parse-dtd parse-ns)
 189   "Parse the region from BEG to END in BUFFER.
 190 If BUFFER is nil, it defaults to the current buffer.
 191 Returns the XML list for the region, or raises an error if the region
 192 is not well-formed XML.
 193 If PARSE-DTD is non-nil, the DTD is parsed rather than skipped,
 194 and returned as the first element of the list.
 195 If PARSE-NS is non-nil, then QNAMES are expanded."
 196   (save-restriction
 197     (narrow-to-region beg end)
 198     ;; Use fixed syntax table to ensure regexp char classes and syntax
 199     ;; specs DTRT.
 200     (with-syntax-table (standard-syntax-table)
 201       (let ((case-fold-search nil)      ; XML is case-sensitive.
 202             xml result dtd)
 203         (save-excursion
 204           (if buffer
 205               (set-buffer buffer))
 206           (goto-char (point-min))
 207           (while (not (eobp))
 208             (if (search-forward "<" nil t)
 209                 (progn
 210                   (forward-char -1)
 211                   (if xml
 212                       ;;  translation of rule [1] of XML specifications
 213                       (error "XML files can have only one toplevel tag")
 214                     (setq result (xml-parse-tag parse-dtd parse-ns))
 215                     (cond
 216                      ((null result))
 217                      ((listp (car result))
 218                       (setq dtd (car result))
 219                       (if (cdr result)  ; possible leading comment
 220                           (add-to-list 'xml (cdr result))))
 221                      (t
 222                       (add-to-list 'xml result)))))
 223               (goto-char (point-max))))
 224           (if parse-dtd
 225               (cons dtd (nreverse xml))
 226             (nreverse xml)))))))
 227
 228
 229 (defun xml-parse-tag (&optional parse-dtd parse-ns)
 230   "Parse the tag at point.
 231 If PARSE-DTD is non-nil, the DTD of the document, if any, is parsed and
 232 returned as the first element in the list.
 233 If PARSE-NS is non-nil, then QNAMES are expanded.
 234 Returns one of:
 235  - a list : the matching node
 236  - nil    : the point is not looking at a tag.
 237  - a pair : the first element is the DTD, the second is the node."
 238   (let ((xml-ns (if (consp parse-ns)
 239                     parse-ns
 240                   (if parse-ns
 241                       (list
 242                        ;; Default no namespace
 243                        (cons "" "")
 244                        ;; We need to seed the xmlns namespace
 245                        (cons "xmlns" "http://www.w3.org/2000/xmlns/"))))))
 246     (cond
 247      ;; Processing instructions (like the <?xml version="1.0"?> tag at the
 248      ;; beginning of a document).
 249      ((looking-at "<\\?")
 250       (search-forward "?>")
 251       (skip-syntax-forward " ")
 252       (xml-parse-tag parse-dtd xml-ns))
 253      ;;  Character data (CDATA) sections, in which no tag should be interpreted
 254      ((looking-at "<!\\[CDATA\\[")
 255       (let ((pos (match-end 0)))
 256         (unless (search-forward "]]>" nil t)
 257           (error "CDATA section does not end anywhere in the document"))
 258         (buffer-substring pos (match-beginning 0))))
 259      ;;  DTD for the document
 260      ((looking-at "<!DOCTYPE")
 261       (let (dtd)
 262         (if parse-dtd
 263             (setq dtd (xml-parse-dtd))
 264           (xml-skip-dtd))
 265       (skip-syntax-forward " ")
 266       (if dtd
 267           (cons dtd (xml-parse-tag nil xml-ns))
 268         (xml-parse-tag nil xml-ns))))
 269      ;;  skip comments
 270      ((looking-at "<!--")
 271       (search-forward "-->")
 272       nil)
 273      ;;  end tag
 274      ((looking-at "</")
 275       '())
 276      ;;  opening tag
 277      ((looking-at "<\\([^/>[:space:]]+\\)")
 278       (goto-char (match-end 1))
 279       (let* ((node-name (match-string 1))
 280              ;; Parse the attribute list.
 281              (children (list (xml-parse-attlist) (intern node-name)))
 282              pos)
 283
 284         ;; add the xmlns:* attrs to our cache
 285         (when (consp xml-ns)
 286           (mapcar
 287            (lambda (attr)
 288              (let* ((splitup (split-string (symbol-name (car attr)) ":"))
 289                     (prefix (nth 0 splitup))
 290                     (lname (nth 1 splitup)))
 291                (when (string= "xmlns" prefix)
 292                  (setq xml-ns (append (list (cons (if lname
 293                                                       lname
 294                                                     "")
 295                                                   (cdr attr)))
 296                                       xml-ns)))))
 297            (car children))
 298
 299           ;; expand element names
 300           (let* ((splitup (split-string (symbol-name (cadr children)) ":"))
 301                  (lname (or (nth 1 splitup)
 302                             (nth 0 splitup)))
 303                  (prefix (if (nth 1 splitup)
 304                              (nth 0 splitup)
 305                            "")))
 306             (setcdr children (list
 307                               (intern (concat "{"
 308                                              (cdr (assoc-string prefix xml-ns))
 309                                              "}" lname)))))
 310
 311           ;; expand attribute names
 312           (mapcar
 313            (lambda (attr)
 314              (let* ((splitup (split-string (symbol-name (car attr)) ":"))
 315                     (lname (or (nth 1 splitup)
 316                                (nth 0 splitup)))
 317                     (prefix (if (nth 1 splitup)
 318                                 (nth 0 splitup)
 319                               (caar xml-ns))))
 320
 321                (setcar attr (intern (concat "{"
 322                                             (cdr (assoc-string prefix xml-ns))
 323                                             "}" lname)))))
 324            (car children)))
 325
 326         ;; is this an empty element ?
 327         (if (looking-at "/>")
 328         (progn
 329           (forward-char 2)
 330           (nreverse children))
 331
 332         ;; is this a valid start tag ?
 333         (if (eq (char-after) ?>)
 334             (progn
 335               (forward-char 1)
 336               ;;  Now check that we have the right end-tag. Note that this
 337               ;;  one might contain spaces after the tag name
 338               (let ((end (concat "</" node-name "\\s-*>")))
 339                 (while (not (looking-at end))
 340                   (cond
 341                    ((looking-at "</")
 342                     (error "XML: Invalid end tag (expecting %s) at pos %d"
 343                            node-name (point)))
 344                    ((= (char-after) ?<)
 345                     (let ((tag (xml-parse-tag nil xml-ns)))
 346                       (when tag
 347                         (push tag children))))
 348                    (t
 349                     (setq pos (point))
 350                     (search-forward "<")
 351                     (forward-char -1)
 352                     (let ((string (buffer-substring pos (point)))
 353                           (pos 0))
 354
 355                       ;; Clean up the string.  As per XML
 356                       ;; specifications, the XML processor should
 357                       ;; always pass the whole string to the
 358                       ;; application.  But \r's should be replaced:
 359                       ;; http://www.w3.org/TR/2000/REC-xml-20001006#sec-line-ends
 360                       (while (string-match "\r\n?" string pos)
 361                         (setq string (replace-match "\n" t t string))
 362                         (setq pos (1+ (match-beginning 0))))
 363
 364                       (setq string (xml-substitute-special string))
 365                       (setq children
 366                             (if (stringp (car children))
 367                                 ;; The two strings were separated by a comment.
 368                                 (cons (concat (car children) string)
 369                                       (cdr children))
 370                               (cons string children))))))))
 371
 372               (goto-char (match-end 0))
 373               (nreverse children))
 374           ;;  This was an invalid start tag
 375           (error "XML: Invalid attribute list")))))
 376      (t ;; This is not a tag.
 377       (error "XML: Invalid character")))))
 378
 379 (defun xml-parse-attlist ()
 380   "Return the attribute-list after point.Leave point at the first non-blank character after the tag."
 381   (let ((attlist ())
 382         start-pos name)
 383     (skip-syntax-forward " ")
 384     (while (looking-at (eval-when-compile
 385                          (concat "\\(" xml-name-regexp "\\)\\s-*=\\s-*")))
 386       (setq name (intern (match-string 1)))
 387       (goto-char (match-end 0))
 388
 389       ;; See also: http://www.w3.org/TR/2000/REC-xml-20001006#AVNormalize
 390
 391       ;; Do we have a string between quotes (or double-quotes),
 392       ;;  or a simple word ?
 393       (if (looking-at "\"\\([^\"]*\\)\"")
 394           (setq start-pos (match-beginning 0))
 395         (if (looking-at "'\\([^']*\\)'")
 396             (setq start-pos (match-beginning 0))
 397           (error "XML: Attribute values must be given between quotes")))
 398
 399       ;; Each attribute must be unique within a given element
 400       (if (assoc name attlist)
 401           (error "XML: each attribute must be unique within an element"))
 402
 403       ;; Multiple whitespace characters should be replaced with a single one
 404       ;; in the attributes
 405       (let ((string (match-string 1))
 406             (pos 0))
 407         (replace-regexp-in-string "\\s-\\{2,\\}" " " string)
 408         (push (cons name (xml-substitute-special string)) attlist))
 409
 410       (goto-char start-pos)
 411       (forward-sexp)                    ; we have string syntax
 412
 413       (skip-syntax-forward " "))
 414     (nreverse attlist)))
 415
 416 ;;*******************************************************************
 417 ;;**
 418 ;;**  The DTD (document type declaration)
 419 ;;**  The following functions know how to skip or parse the DTD of
 420 ;;**  a document
 421 ;;**
 422 ;;*******************************************************************
 423
 424 ;; Fixme: This fails at least if the DTD contains conditional sections.
 425
 426 (defun xml-skip-dtd ()
 427   "Skip the DTD at point.
 428 This follows the rule [28] in the XML specifications."
 429   (forward-char (length "<!DOCTYPE"))
 430   (if (looking-at "\\s-*>")
 431       (error "XML: invalid DTD (excepting name of the document)"))
 432   (condition-case nil
 433       (progn
 434         (forward-sexp)
 435         (skip-syntax-forward " ")
 436         (if (looking-at "\\[")
 437             (re-search-forward "]\\s-*>")
 438           (search-forward ">")))
 439     (error (error "XML: No end to the DTD"))))
 440
 441 (defun xml-parse-dtd ()
 442   "Parse the DTD at point."
 443   (forward-char (eval-when-compile (length "<!DOCTYPE")))
 444   (skip-syntax-forward " ")
 445   (if (looking-at ">")
 446       (error "XML: invalid DTD (excepting name of the document)"))
 447
 448   ;;  Get the name of the document
 449   (looking-at xml-name-regexp)
 450   (let ((dtd (list (match-string 0) 'dtd))
 451         type element end-pos)
 452     (goto-char (match-end 0))
 453
 454     (skip-syntax-forward " ")
 455     ;; XML [75]
 456     (cond ((looking-at "PUBLIC\\s-+")
 457            (goto-char (match-end 0))
 458            (unless (or (re-search-forward
 459                         "\\=\"\\([[:space:][:alnum:]-'()+,./:=?;!*#@$_%]*\\)\""
 460                         nil t)
 461                        (re-search-forward
 462                         "\\='\\([[:space:][:alnum:]-()+,./:=?;!*#@$_%]*\\)'"
 463                         nil t))
 464              (error "XML: missing public id"))
 465            (let ((pubid (match-string 1)))
 466              (unless (or (re-search-forward "\\='\\([^']*\\)'" nil t)
 467                          (re-search-forward "\\=\"\\([^\"]*\\)\"" nil t))
 468                (error "XML: missing system id"))
 469              (push (list pubid (match-string 1) 'public) dtd)))
 470           ((looking-at "SYSTEM\\s-+")
 471            (goto-char (match-end 0))
 472            (unless (or (re-search-forward "\\='\\([^']*\\)'" nil t)
 473                        (re-search-forward "\\=\"\\([^\"]*\\)\"" nil t))
 474              (error "XML: missing system id"))
 475            (push (list (match-string 1) 'system) dtd)))
 476     (skip-syntax-forward " ")
 477     (if (eq ?> (char-after))
 478         (forward-char)
 479       (skip-syntax-forward " ")
 480       (if (not (eq (char-after) ?\[))
 481           (error "XML: bad DTD")
 482         (forward-char)
 483         ;;  Parse the rest of the DTD
 484         ;;  Fixme: Deal with ENTITY, ATTLIST, NOTATION, PIs.
 485         (while (not (looking-at "\\s-*\\]"))
 486           (skip-syntax-forward " ")
 487           (cond
 488
 489            ;;  Translation of rule [45] of XML specifications
 490            ((looking-at
 491              "<!ELEMENT\\s-+\\([[:alnum:].%;]+\\)\\s-+\\([^>]+\\)>")
 492
 493             (setq element (intern (match-string 1))
 494                   type    (match-string-no-properties 2))
 495             (setq end-pos (match-end 0))
 496
 497             ;;  Translation of rule [46] of XML specifications
 498             (cond
 499              ((string-match "^EMPTY[ \t\n\r]*$" type) ;; empty declaration
 500               (setq type 'empty))
 501              ((string-match "^ANY[ \t\n\r]*$" type) ;; any type of contents
 502               (setq type 'any))
 503              ((string-match "^(\\(.*\\))[ \t\n\r]*$" type) ;; children ([47])
 504               (setq type (xml-parse-elem-type (match-string 1 type))))
 505              ((string-match "^%[^;]+;[ \t\n\r]*$" type) ;; substitution
 506               nil)
 507              (t
 508               (error "XML: Invalid element type in the DTD")))
 509
 510             ;;  rule [45]: the element declaration must be unique
 511             (if (assoc element dtd)
 512                 (error "XML: element declarations must be unique in a DTD (<%s>)"
 513                        (symbol-name element)))
 514
 515             ;;  Store the element in the DTD
 516             (push (list element type) dtd)
 517             (goto-char end-pos))
 518            ((looking-at "<!--")
 519             (search-forward "-->"))
 520
 521            (t
 522             (error "XML: Invalid DTD item")))
 523
 524           ;;  Skip the end of the DTD
 525           (search-forward ">"))))
 526     (nreverse dtd)))
 527
 528
 529 (defun xml-parse-elem-type (string)
 530   "Convert element type STRING into a Lisp structure."
 531
 532   (let (elem modifier)
 533     (if (string-match "(\\([^)]+\\))\\([+*?]?\\)" string)
 534         (progn
 535           (setq elem     (match-string 1 string)
 536                 modifier (match-string 2 string))
 537           (if (string-match "|" elem)
 538               (setq elem (cons 'choice
 539                                (mapcar 'xml-parse-elem-type
 540                                        (split-string elem "|"))))
 541             (if (string-match "," elem)
 542                 (setq elem (cons 'seq
 543                                  (mapcar 'xml-parse-elem-type
 544                                          (split-string elem ",")))))))
 545       (if (string-match "[ \t\n\r]*\\([^+*?]+\\)\\([+*?]?\\)" string)
 546           (setq elem     (match-string 1 string)
 547                 modifier (match-string 2 string))))
 548
 549     (if (and (stringp elem) (string= elem "#PCDATA"))
 550         (setq elem 'pcdata))
 551
 552     (cond
 553      ((string= modifier "+")
 554       (list '+ elem))
 555      ((string= modifier "*")
 556       (list '* elem))
 557      ((string= modifier "?")
 558       (list '\? elem))
 559      (t
 560       elem))))
 561
 562 ;;*******************************************************************
 563 ;;**
 564 ;;**  Substituting special XML sequences
 565 ;;**
 566 ;;*******************************************************************
 567
 568 (eval-when-compile
 569   (defvar str))                ; dynamic from replace-regexp-in-string
 570
 571 ;; Fixme:  Take declared entities from the DTD when they're available.
 572 (defun xml-substitute-entity (match)
 573   "Subroutine of xml-substitute-special."
 574   (save-match-data
 575     (let ((match1 (match-string 1 str)))
 576       (cond ((string= match1 "lt") "<")
 577             ((string= match1 "gt") ">")
 578             ((string= match1 "apos") "'")
 579             ((string= match1 "quot") "\"")
 580             ((string= match1 "amp") "&")
 581             ((and (string-match "#\\([0-9]+\\)" match1)
 582                   (let ((c (decode-char
 583                             'ucs
 584                             (string-to-number (match-string 1 match1)))))
 585                     (if c (string c))))) ; else unrepresentable
 586             ((and (string-match "#x\\([[:xdigit:]]+\\)" match1)
 587                   (let ((c (decode-char
 588                             'ucs
 589                             (string-to-number (match-string 1 match1) 16))))
 590                     (if c (string c)))))
 591             ;; Default to asis.  Arguably, unrepresentable code points
 592             ;; might be best replaced with U+FFFD.
 593             (t match)))))
 594
 595 (defun xml-substitute-special (string)
 596   "Return STRING, after subsituting entity references."
 597   ;; This originally made repeated passes through the string from the
 598   ;; beginning, which isn't correct, since then either "&amp;amp;" or
 599   ;; "&#38;amp;" won't DTRT.
 600   (replace-regexp-in-string "&\\([^;]+\\);"
 601                             #'xml-substitute-entity string t t))
 602
 603 ;;*******************************************************************
 604 ;;**
 605 ;;**  Printing a tree.
 606 ;;**  This function is intended mainly for debugging purposes.
 607 ;;**
 608 ;;*******************************************************************
 609
 610 (defun xml-debug-print (xml)
 611   (dolist (node xml)
 612     (xml-debug-print-internal node "")))
 613
 614 (defun xml-debug-print-internal (xml indent-string)
 615   "Outputs the XML tree in the current buffer.
 616 The first line is indented with INDENT-STRING."
 617   (let ((tree xml)
 618         attlist)
 619     (insert indent-string ?< (symbol-name (xml-node-name tree)))
 620
 621     ;;  output the attribute list
 622     (setq attlist (xml-node-attributes tree))
 623     (while attlist
 624       (insert ?\  (symbol-name (caar attlist)) "=\"" (cdar attlist) ?\")
 625       (setq attlist (cdr attlist)))
 626
 627     (insert ?>)
 628
 629     (setq tree (xml-node-children tree))
 630
 631     ;;  output the children
 632     (dolist (node tree)
 633       (cond
 634        ((listp node)
 635         (insert ?\n)
 636         (xml-debug-print-internal node (concat indent-string "  ")))
 637        ((stringp node) (insert node))
 638        (t
 639         (error "Invalid XML tree"))))
 640
 641     (insert ?\n indent-string
 642             ?< ?/ (symbol-name (xml-node-name xml)) ?>)))
 643
 644 (provide 'xml)
 645
 646 ;;; xml.el ends here