;;; xml.el --- XML parser
-;; Copyright (C) 2000-2012 Free Software Foundation, Inc.
+;; Copyright (C) 2000-2016 Free Software Foundation, Inc.
;; Author: Emmanuel Briot <briot@gnat.com>
;; Maintainer: Mark A. Hershberger <mah@everybody.org>
would be represented by
- '(\"\" . \"foo\")."
+ (\"\" . \"foo\").
+
+If you'd just like a plain symbol instead, use `symbol-qnames' in
+the PARSE-NS argument."
(car node))
See also `xml-get-attribute-or-nil'."
(or (xml-get-attribute-or-nil node attribute) ""))
-;;; Creating the list
-
-;;;###autoload
-(defun xml-parse-file (file &optional parse-dtd parse-ns)
- "Parse the well-formed XML file FILE.
-Return the top node with all its children.
-If PARSE-DTD is non-nil, the DTD is parsed rather than skipped.
-If PARSE-NS is non-nil, then QNAMES are expanded."
- (with-temp-buffer
- (insert-file-contents file)
- (xml--parse-buffer parse-dtd parse-ns)))
+;;; Regular expressions for XML components
+;; The following regexps are used as subexpressions in regexps that
+;; are `eval-when-compile'd for efficiency, so they must be defined at
+;; compile time.
(eval-and-compile
-(let* ((start-chars (concat "[:alpha:]:_"))
- (name-chars (concat "-[:digit:]." start-chars))
- ;;[3] S ::= (#x20 | #x9 | #xD | #xA)+
- (whitespace "[ \t\n\r]"))
- ;; [4] NameStartChar ::= ":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6]
- ;; | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] | [#x37F-#x1FFF]
- ;; | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF]
- ;; | [#x3001-#xD7FF] | [#xF900-#xFDCF] | [#xFDF0-#xFFFD]
- ;; | [#x10000-#xEFFFF]
- (defconst xml-name-start-char-re (concat "[" start-chars "]"))
- ;; [4a] NameChar ::= NameStartChar | "-" | "." | [0-9] | #xB7
- ;; | [#x0300-#x036F] | [#x203F-#x2040]
- (defconst xml-name-char-re (concat "[" name-chars "]"))
- ;; [5] Name ::= NameStartChar (NameChar)*
- (defconst xml-name-re (concat xml-name-start-char-re xml-name-char-re "*"))
- ;; [6] Names ::= Name (#x20 Name)*
- (defconst xml-names-re (concat xml-name-re "\\(?: " xml-name-re "\\)*"))
- ;; [7] Nmtoken ::= (NameChar)+
- (defconst xml-nmtoken-re (concat xml-name-char-re "+"))
- ;; [8] Nmtokens ::= Nmtoken (#x20 Nmtoken)*
- (defconst xml-nmtokens-re (concat xml-nmtoken-re "\\(?: " xml-name-re "\\)*"))
- ;; [66] CharRef ::= '&#' [0-9]+ ';' | '&#x' [0-9a-fA-F]+ ';'
- (defconst xml-char-ref-re "\\(?:&#[0-9]+;\\|&#x[0-9a-fA-F]+;\\)")
- ;; [68] EntityRef ::= '&' Name ';'
- (defconst xml-entity-ref (concat "&" xml-name-re ";"))
- ;; [69] PEReference ::= '%' Name ';'
- (defconst xml-pe-reference-re (concat "%" xml-name-re ";"))
- ;; [67] Reference ::= EntityRef | CharRef
- (defconst xml-reference-re (concat "\\(?:" xml-entity-ref "\\|" xml-char-ref-re "\\)"))
- ;; [10] AttValue ::= '"' ([^<&"] | Reference)* '"' | "'" ([^<&'] | Reference)* "'"
- (defconst xml-att-value-re (concat "\\(?:\"\\(?:[^&\"]\\|" xml-reference-re "\\)*\"\\|"
- "'\\(?:[^&']\\|" xml-reference-re "\\)*'\\)"))
- ;; [56] TokenizedType ::= 'ID' [VC: ID] [VC: One ID / Element Type] [VC: ID Attribute Default]
- ;; | 'IDREF' [VC: IDREF]
- ;; | 'IDREFS' [VC: IDREF]
- ;; | 'ENTITY' [VC: Entity Name]
- ;; | 'ENTITIES' [VC: Entity Name]
- ;; | 'NMTOKEN' [VC: Name Token]
- ;; | 'NMTOKENS' [VC: Name Token]
- (defconst xml-tokenized-type-re (concat "\\(?:ID\\|IDREF\\|IDREFS\\|ENTITY\\|"
- "ENTITIES\\|NMTOKEN\\|NMTOKENS\\)"))
- ;; [58] NotationType ::= 'NOTATION' S '(' S? Name (S? '|' S? Name)* S? ')'
- (defconst xml-notation-type-re
- (concat "\\(?:NOTATION" whitespace "(" whitespace "*" xml-name-re
- "\\(?:" whitespace "*|" whitespace "*" xml-name-re "\\)*"
- whitespace "*)\\)"))
- ;; [59] Enumeration ::= '(' S? Nmtoken (S? '|' S? Nmtoken)* S? ')'
- ;; [VC: Enumeration] [VC: No Duplicate Tokens]
- (defconst xml-enumeration-re (concat "\\(?:(" whitespace "*" xml-nmtoken-re
- "\\(?:" whitespace "*|" whitespace "*"
- xml-nmtoken-re "\\)*"
- whitespace ")\\)"))
- ;; [57] EnumeratedType ::= NotationType | Enumeration
- (defconst xml-enumerated-type-re (concat "\\(?:" xml-notation-type-re
- "\\|" xml-enumeration-re "\\)"))
- ;; [54] AttType ::= StringType | TokenizedType | EnumeratedType
- ;; [55] StringType ::= 'CDATA'
- (defconst xml-att-type-re (concat "\\(?:CDATA\\|" xml-tokenized-type-re
- "\\|" xml-notation-type-re
- "\\|" xml-enumerated-type-re "\\)"))
- ;; [60] DefaultDecl ::= '#REQUIRED' | '#IMPLIED' | (('#FIXED' S)? AttValue)
- (defconst xml-default-decl-re (concat "\\(?:#REQUIRED\\|#IMPLIED\\|\\(?:#FIXED"
- whitespace "\\)*" xml-att-value-re "\\)"))
- ;; [53] AttDef ::= S Name S AttType S DefaultDecl
- (defconst xml-att-def-re (concat "\\(?:" whitespace "*" xml-name-re
- whitespace "*" xml-att-type-re
- whitespace "*" xml-default-decl-re "\\)"))
- ;; [9] EntityValue ::= '"' ([^%&"] | PEReference | Reference)* '"'
- ;; | "'" ([^%&'] | PEReference | Reference)* "'"
- (defconst xml-entity-value-re (concat "\\(?:\"\\(?:[^%&\"]\\|" xml-pe-reference-re
- "\\|" xml-reference-re
- "\\)*\"\\|'\\(?:[^%&']\\|"
- xml-pe-reference-re "\\|"
- xml-reference-re "\\)*'\\)"))))
+
+;; [4] NameStartChar
+;; See the definition of word syntax in `xml-syntax-table'.
+(defconst xml-name-start-char-re (concat "[[:word:]:_]"))
+
+;; [4a] NameChar ::= NameStartChar | "-" | "." | [0-9] | #xB7
+;; | [#x0300-#x036F] | [#x203F-#x2040]
+(defconst xml-name-char-re (concat "[-0-9.[:word:]:_·̀-ͯ‿-⁀]"))
+
+;; [5] Name ::= NameStartChar (NameChar)*
+(defconst xml-name-re (concat xml-name-start-char-re xml-name-char-re "*"))
+
+;; [6] Names ::= Name (#x20 Name)*
+(defconst xml-names-re (concat xml-name-re "\\(?: " xml-name-re "\\)*"))
+
+;; [7] Nmtoken ::= (NameChar)+
+(defconst xml-nmtoken-re (concat xml-name-char-re "+"))
+
+;; [8] Nmtokens ::= Nmtoken (#x20 Nmtoken)*
+(defconst xml-nmtokens-re (concat xml-nmtoken-re "\\(?: " xml-name-re "\\)*"))
+
+;; [66] CharRef ::= '&#' [0-9]+ ';' | '&#x' [0-9a-fA-F]+ ';'
+(defconst xml-char-ref-re "\\(?:&#[0-9]+;\\|&#x[0-9a-fA-F]+;\\)")
+
+;; [68] EntityRef ::= '&' Name ';'
+(defconst xml-entity-ref (concat "&" xml-name-re ";"))
+
+(defconst xml-entity-or-char-ref-re (concat "&\\(?:#\\(x\\)?\\([0-9a-fA-F]+\\)\\|\\("
+ xml-name-re "\\)\\);"))
+
+;; [69] PEReference ::= '%' Name ';'
+(defconst xml-pe-reference-re (concat "%\\(" xml-name-re "\\);"))
+
+;; [67] Reference ::= EntityRef | CharRef
+(defconst xml-reference-re (concat "\\(?:" xml-entity-ref "\\|" xml-char-ref-re "\\)"))
+
+;; [10] AttValue ::= '"' ([^<&"] | Reference)* '"'
+;; | "'" ([^<&'] | Reference)* "'"
+(defconst xml-att-value-re (concat "\\(?:\"\\(?:[^&\"]\\|"
+ xml-reference-re "\\)*\"\\|"
+ "'\\(?:[^&']\\|" xml-reference-re
+ "\\)*'\\)"))
+
+;; [56] TokenizedType ::= 'ID'
+;; [VC: ID] [VC: One ID / Element Type] [VC: ID Attribute Default]
+;; | 'IDREF' [VC: IDREF]
+;; | 'IDREFS' [VC: IDREF]
+;; | 'ENTITY' [VC: Entity Name]
+;; | 'ENTITIES' [VC: Entity Name]
+;; | 'NMTOKEN' [VC: Name Token]
+;; | 'NMTOKENS' [VC: Name Token]
+(defconst xml-tokenized-type-re (concat "\\(?:ID\\|IDREF\\|IDREFS\\|ENTITY\\|"
+ "ENTITIES\\|NMTOKEN\\|NMTOKENS\\)"))
+
+;; [58] NotationType ::= 'NOTATION' S '(' S? Name (S? '|' S? Name)* S? ')'
+(defconst xml-notation-type-re
+ (concat "\\(?:NOTATION\\s-+(\\s-*" xml-name-re
+ "\\(?:\\s-*|\\s-*" xml-name-re "\\)*\\s-*)\\)"))
+
+;; [59] Enumeration ::= '(' S? Nmtoken (S? '|' S? Nmtoken)* S? ')'
+;; [VC: Enumeration] [VC: No Duplicate Tokens]
+(defconst xml-enumeration-re (concat "\\(?:(\\s-*" xml-nmtoken-re
+ "\\(?:\\s-*|\\s-*" xml-nmtoken-re
+ "\\)*\\s-+)\\)"))
+
+;; [57] EnumeratedType ::= NotationType | Enumeration
+(defconst xml-enumerated-type-re (concat "\\(?:" xml-notation-type-re
+ "\\|" xml-enumeration-re "\\)"))
+
+;; [54] AttType ::= StringType | TokenizedType | EnumeratedType
+;; [55] StringType ::= 'CDATA'
+(defconst xml-att-type-re (concat "\\(?:CDATA\\|" xml-tokenized-type-re
+ "\\|" xml-notation-type-re
+ "\\|" xml-enumerated-type-re "\\)"))
+
+;; [60] DefaultDecl ::= '#REQUIRED' | '#IMPLIED' | (('#FIXED' S)? AttValue)
+(defconst xml-default-decl-re (concat "\\(?:#REQUIRED\\|#IMPLIED\\|"
+ "\\(?:#FIXED\\s-+\\)*"
+ xml-att-value-re "\\)"))
+
+;; [53] AttDef ::= S Name S AttType S DefaultDecl
+(defconst xml-att-def-re (concat "\\(?:\\s-*" xml-name-re
+ "\\s-*" xml-att-type-re
+ "\\s-*" xml-default-decl-re "\\)"))
+
+;; [9] EntityValue ::= '"' ([^%&"] | PEReference | Reference)* '"'
+;; | "'" ([^%&'] | PEReference | Reference)* "'"
+(defconst xml-entity-value-re (concat "\\(?:\"\\(?:[^%&\"]\\|"
+ xml-pe-reference-re
+ "\\|" xml-reference-re
+ "\\)*\"\\|'\\(?:[^%&']\\|"
+ xml-pe-reference-re "\\|"
+ xml-reference-re "\\)*'\\)"))
+) ; End of `eval-when-compile'
+
;; [75] ExternalID ::= 'SYSTEM' S SystemLiteral
;; | 'PUBLIC' S PubidLiteral S SystemLiteral
;; Note that this is setup so that we can do whitespace-skipping with
;; `(skip-syntax-forward " ")', inter alia. Previously this was slow
-;; compared with `re-search-forward', but that has been fixed. Also
-;; note that the standard syntax table contains other characters with
-;; whitespace syntax, like NBSP, but they are invalid in contexts in
-;; which we might skip whitespace -- specifically, they're not
-;; NameChars [XML 4].
+;; compared with `re-search-forward', but that has been fixed.
(defvar xml-syntax-table
- (let ((table (make-syntax-table)))
- ;; Get space syntax correct per XML [3].
- (dotimes (c 31)
- (modify-syntax-entry c "." table)) ; all are space in standard table
- (dolist (c '(?\t ?\n ?\r)) ; these should be space
+ ;; By default, characters have symbol syntax.
+ (let ((table (make-char-table 'syntax-table '(3))))
+ ;; The XML space chars [3], and nothing else, have space syntax.
+ (dolist (c '(?\s ?\t ?\r ?\n))
(modify-syntax-entry c " " table))
- ;; For skipping attributes.
- (modify-syntax-entry ?\" "\"" table)
- (modify-syntax-entry ?' "\"" table)
- ;; Non-alnum name chars should be symbol constituents (`-' and `_'
- ;; are OK by default).
- (modify-syntax-entry ?. "_" table)
- (modify-syntax-entry ?: "_" table)
- ;; XML [89]
- (unless (featurep 'xemacs)
- (dolist (c '(#x00B7 #x02D0 #x02D1 #x0387 #x0640 #x0E46 #x0EC6 #x3005
- #x3031 #x3032 #x3033 #x3034 #x3035 #x309D #x309E #x30FC
- #x30FD #x30FE))
- (modify-syntax-entry (decode-char 'ucs c) "w" table)))
- ;; Fixme: rest of [4]
+ ;; The characters in NameStartChar [4], aside from ':' and '_',
+ ;; have word syntax. This is used by `xml-name-start-char-re'.
+ (modify-syntax-entry '(?A . ?Z) "w" table)
+ (modify-syntax-entry '(?a . ?z) "w" table)
+ (modify-syntax-entry '(#xC0 . #xD6) "w" table)
+ (modify-syntax-entry '(#xD8 . #XF6) "w" table)
+ (modify-syntax-entry '(#xF8 . #X2FF) "w" table)
+ (modify-syntax-entry '(#x370 . #X37D) "w" table)
+ (modify-syntax-entry '(#x37F . #x1FFF) "w" table)
+ (modify-syntax-entry '(#x200C . #x200D) "w" table)
+ (modify-syntax-entry '(#x2070 . #x218F) "w" table)
+ (modify-syntax-entry '(#x2C00 . #x2FEF) "w" table)
+ (modify-syntax-entry '(#x3001 . #xD7FF) "w" table)
+ (modify-syntax-entry '(#xF900 . #xFDCF) "w" table)
+ (modify-syntax-entry '(#xFDF0 . #xFFFD) "w" table)
+ (modify-syntax-entry '(#x10000 . #xEFFFF) "w" table)
table)
- "Syntax table used by `xml-parse-region'.")
+ "Syntax table used by the XML parser.
+In this syntax table, the XML space characters [ \\t\\r\\n], and
+only those characters, have whitespace syntax.")
-;; XML [5]
+;;; Entry points:
-;; Fixme: This needs re-writing to deal with the XML grammar properly, i.e.
-;; document ::= prolog element Misc*
-;; prolog ::= XMLDecl? Misc* (doctypedecl Misc*)?
+;;;###autoload
+(defun xml-parse-file (file &optional parse-dtd parse-ns)
+ "Parse the well-formed XML file FILE.
+Return the top node with all its children.
+If PARSE-DTD is non-nil, the DTD is parsed rather than skipped.
+
+If PARSE-NS is non-nil, then QNAMES are expanded. By default,
+the variable `xml-default-ns' is the mapping from namespaces to
+URIs, and expanded names will be returned as a cons
+
+ (\"namespace:\" . \"foo\").
+
+If PARSE-NS is an alist, it will be used as the mapping from
+namespace to URIs instead.
+
+If it is the symbol `symbol-qnames', expanded names will be
+returned as a plain symbol `namespace:foo' instead of a cons.
+
+Both features can be combined by providing a cons cell
+
+ (symbol-qnames . ALIST)."
+ (with-temp-buffer
+ (insert-file-contents file)
+ (xml--parse-buffer parse-dtd parse-ns)))
;;;###autoload
(defun xml-parse-region (&optional beg end buffer parse-dtd parse-ns)
"Parse the region from BEG to END in BUFFER.
+Return the XML parse tree, or raise an error if the region does
+not contain well-formed XML.
+
If BEG is nil, it defaults to `point-min'.
If END is nil, it defaults to `point-max'.
If BUFFER is nil, it defaults to the current buffer.
-Returns the XML list for the region, or raises an error if the region
-is not well-formed XML.
-If PARSE-DTD is non-nil, the DTD is parsed rather than skipped,
-and returned as the first element of the list.
-If PARSE-NS is non-nil, then QNAMES are expanded."
+If PARSE-DTD is non-nil, parse the DTD and return it as the first
+element of the list.
+If PARSE-NS is non-nil, then QNAMES are expanded. By default,
+the variable `xml-default-ns' is the mapping from namespaces to
+URIs, and expanded names will be returned as a cons
+
+ (\"namespace:\" . \"foo\").
+
+If PARSE-NS is an alist, it will be used as the mapping from
+namespace to URIs instead.
+
+If it is the symbol `symbol-qnames', expanded names will be
+returned as a plain symbol `namespace:foo' instead of a cons.
+
+Both features can be combined by providing a cons cell
+
+ (symbol-qnames . ALIST)."
;; Use fixed syntax table to ensure regexp char classes and syntax
;; specs DTRT.
(unless buffer
(insert-buffer-substring-no-properties buffer beg end)
(xml--parse-buffer parse-dtd parse-ns)))
+;; XML [5]
+
+;; Fixme: This needs re-writing to deal with the XML grammar properly, i.e.
+;; document ::= prolog element Misc*
+;; prolog ::= XMLDecl? Misc* (doctypedecl Misc*)?
+
(defun xml--parse-buffer (parse-dtd parse-ns)
- (with-syntax-table (standard-syntax-table)
+ (with-syntax-table xml-syntax-table
(let ((case-fold-search nil) ; XML is case-sensitive.
;; Prevent entity definitions from changing the defaults
(xml-entity-alist xml-entity-alist)
During namespace-aware parsing, any name without a namespace is
put into the namespace identified by DEFAULT. nil is used to
-specify that the name shouldn't be given a namespace."
+specify that the name shouldn't be given a namespace.
+Expanded names will by default be returned as a cons. If you
+would like to get plain symbols instead, provide a cons cell
+
+ (symbol-qnames . ALIST)
+
+in the XML-NS argument."
(if (consp xml-ns)
- (let* ((nsp (string-match ":" name))
+ (let* ((symbol-qnames (eq (car-safe xml-ns) 'symbol-qnames))
+ (nsp (string-match ":" name))
(lname (if nsp (substring name (match-end 0)) name))
(prefix (if nsp (substring name 0 (match-beginning 0)) default))
(special (and (string-equal lname "xmlns") (not prefix)))
;; Setting default to nil will insure that there is not
;; matching cons in xml-ns. In which case we
(ns (or (cdr (assoc (if special "xmlns" prefix)
- xml-ns))
+ (if symbol-qnames (cdr xml-ns) xml-ns)))
"")))
- (cons ns (if special "" lname)))
+ (if (and symbol-qnames
+ (not (string= prefix "xmlns")))
+ (intern (concat ns lname))
+ (cons ns (if special "" lname))))
(intern name)))
-(defun xml-parse-fragment (&optional parse-dtd parse-ns)
- "Parse xml-like fragments."
- (let ((xml-sub-parser t)
- ;; Prevent entity definitions from changing the defaults
- (xml-entity-alist xml-entity-alist)
- (xml-parameter-entity-alist xml-parameter-entity-alist)
- children)
- (while (not (eobp))
- (let ((bit (xml-parse-tag-1 parse-dtd parse-ns)))
- (if children
- (setq children (append (list bit) children))
- (if (stringp bit)
- (setq children (list bit))
- (setq children bit)))))
- (reverse children)))
-
(defun xml-parse-tag (&optional parse-dtd parse-ns)
"Parse the tag at point.
If PARSE-DTD is non-nil, the DTD of the document, if any, is parsed and
returned as the first element in the list.
-If PARSE-NS is non-nil, expand QNAMES; if the value of PARSE-NS
-is a list, use it as an alist mapping namespaces to URIs.
+If PARSE-NS is non-nil, expand QNAMES; for further details, see
+`xml-parse-region'.
Return one of:
- a list : the matching node
- nil : the point is not looking at a tag.
- a pair : the first element is the DTD, the second is the node."
- (let ((buf (current-buffer))
- (pos (point)))
+ (let* ((case-fold-search nil)
+ ;; Prevent entity definitions from changing the defaults
+ (xml-entity-alist xml-entity-alist)
+ (xml-parameter-entity-alist xml-parameter-entity-alist)
+ (buf (current-buffer))
+ (pos (point)))
(with-temp-buffer
- (insert-buffer-substring-no-properties buf pos)
- (goto-char (point-min))
- (xml-parse-tag-1 parse-dtd parse-ns))))
+ (with-syntax-table xml-syntax-table
+ (insert-buffer-substring-no-properties buf pos)
+ (goto-char (point-min))
+ (xml-parse-tag-1 parse-dtd parse-ns)))))
(defun xml-parse-tag-1 (&optional parse-dtd parse-ns)
"Like `xml-parse-tag', but possibly modify the buffer while working."
- (let ((xml-validating-parser (or parse-dtd xml-validating-parser))
- (xml-ns (cond ((consp parse-ns) parse-ns)
- (parse-ns xml-default-ns))))
+ (let* ((xml-validating-parser (or parse-dtd xml-validating-parser))
+ (xml-ns
+ (cond ((eq parse-ns 'symbol-qnames)
+ (cons 'symbol-qnames xml-default-ns))
+ ((or (consp (car-safe parse-ns))
+ (and (eq (car-safe parse-ns) 'symbol-qnames)
+ (listp (cdr parse-ns))))
+ parse-ns)
+ (parse-ns
+ xml-default-ns))))
(cond
;; Processing instructions, like <?xml version="1.0"?>.
- ((looking-at "<\\?")
+ ((looking-at-p "<\\?")
(search-forward "?>")
(skip-syntax-forward " ")
(xml-parse-tag-1 parse-dtd xml-ns))
(buffer-substring-no-properties pos (match-beginning 0))
(xml-parse-string))))
;; DTD for the document
- ((looking-at "<!DOCTYPE[ \t\n\r]")
+ ((looking-at-p "<!DOCTYPE[ \t\n\r]")
(let ((dtd (xml-parse-dtd parse-ns)))
(skip-syntax-forward " ")
(if xml-validating-parser
(cons dtd (xml-parse-tag-1 nil xml-ns))
(xml-parse-tag-1 nil xml-ns))))
;; skip comments
- ((looking-at "<!--")
+ ((looking-at-p "<!--")
(search-forward "-->")
;; FIXME: This loses the skipped-over spaces.
(skip-syntax-forward " ")
(let ((xml-sub-parser t))
(xml-parse-tag-1 parse-dtd xml-ns))))
;; end tag
- ((looking-at "</")
+ ((looking-at-p "</")
'())
;; opening tag
((looking-at (eval-when-compile (concat "<\\(" xml-name-re "\\)")))
(equal "http://www.w3.org/2000/xmlns/"
(caar attr)))
(push (cons (cdar attr) (cdr attr))
- xml-ns))))
+ (if (symbolp (car xml-ns))
+ (cdr xml-ns)
+ xml-ns)))))
(setq children (list attrs (xml-maybe-do-ns node-name "" xml-ns)))
(cond
;; is this an empty element ?
- ((looking-at "/>")
+ ((looking-at-p "/>")
(forward-char 2)
(nreverse children))
;; is this a valid start tag ?
((eobp)
(error "XML: (Not Well-Formed) End of document while reading element `%s'"
node-name))
- ((looking-at "</")
+ ((looking-at-p "</")
(forward-char 2)
(error "XML: (Not Well-Formed) Invalid end tag `%s' (expecting `%s')"
(let ((pos (point)))
(error "XML: (Well-Formed) Invalid character"))
;; However, if we're parsing incrementally, then we need to deal
;; with stray CDATA.
- (xml-parse-string)))))
+ (let ((s (xml-parse-string)))
+ (when (zerop (length s))
+ ;; We haven't consumed any input! We must throw an error in
+ ;; order to prevent looping forever.
+ (error "XML: (Not Well-Formed) Could not parse: %s"
+ (buffer-substring-no-properties
+ (point) (min (+ (point) 10) (point-max)))))
+ s)))))
(defun xml-parse-string ()
"Parse character data at point, and return it as a string.
(old-remaining-size (- (buffer-size) (point)))
ref val)
(while (and (not (eobp))
- (not (looking-at "<")))
+ (not (looking-at-p "<")))
;; Find the next < or & character.
(skip-chars-forward "^<&")
(when (eq (char-after) ?&)
;; If we find an entity or character reference, expand it.
- (unless (looking-at (eval-when-compile
- (concat "&\\(?:#\\([0-9]+\\)\\|#x\\([0-9a-fA-F]+\\)\\|\\("
- xml-name-re "\\)\\);")))
+ (unless (looking-at xml-entity-or-char-ref-re)
(error "XML: (Not Well-Formed) Invalid entity reference"))
;; For a character reference, the next entity or character
;; reference must be after the replacement. [4.6] "Numerical
;; character references are expanded immediately when
;; recognized and MUST be treated as character data."
- (cond ((setq ref (match-string 1))
- ;; Decimal character reference
- (setq val (save-match-data
- (decode-char 'ucs (string-to-number ref))))
- (and (null val)
- xml-validating-parser
- (error "XML: (Validity) Invalid character `%s'" ref))
- (replace-match (or (string val) xml-undefined-entity) t t))
- ;; Hexadecimal character reference
- ((setq ref (match-string 2))
- (setq val (save-match-data
- (decode-char 'ucs (string-to-number ref 16))))
- (and (null val)
- xml-validating-parser
- (error "XML: (Validity) Invalid character `x%s'" ref))
- (replace-match (or (string val) xml-undefined-entity) t t))
- ;; For an entity reference, search again from the start
- ;; of the replaced text, since the replacement can
- ;; contain entity or character references, or markup.
- ((setq ref (match-string 3))
- (setq val (assoc ref xml-entity-alist))
- (and (null val)
- xml-validating-parser
- (error "XML: (Validity) Undefined entity `%s'" ref))
- (replace-match (cdr val) t t)
- (goto-char (match-beginning 0))))
+ (if (setq ref (match-string 2))
+ (progn ; Numeric char reference
+ (setq val (save-match-data
+ (decode-char 'ucs (string-to-number
+ ref (if (match-string 1) 16)))))
+ (and (null val)
+ xml-validating-parser
+ (error "XML: (Validity) Invalid character reference `%s'"
+ (match-string 0)))
+ (replace-match (if val (string val) xml-undefined-entity) t t))
+ ;; For an entity reference, search again from the start of
+ ;; the replaced text, since the replacement can contain
+ ;; entity or character references, or markup.
+ (setq ref (match-string 3)
+ val (assoc ref xml-entity-alist))
+ (and (null val)
+ xml-validating-parser
+ (error "XML: (Validity) Undefined entity `%s'" ref))
+ (replace-match (or (cdr val) xml-undefined-entity) t t)
+ (goto-char (match-beginning 0)))
;; Check for XML bombs.
(and xml-entity-expansion-limit
(> (- (buffer-size) (point))
(replace-regexp-in-string "\\s-\\{2,\\}" " " string)
(let ((expansion (xml-substitute-special string)))
(unless (stringp expansion)
- ; We say this is the constraint. It is actually that neither
- ; external entities nor "<" can be in an attribute value.
+ ;; We say this is the constraint. It is actually that
+ ;; neither external entities nor "<" can be in an
+ ;; attribute value.
(error "XML: (Not Well-Formed) Entities in attributes cannot expand into elements"))
(push (cons name expansion) attlist)))
(let ((xml-validating-parser nil))
(xml-parse-dtd)))
-(defun xml-parse-dtd (&optional parse-ns)
+(defun xml-parse-dtd (&optional _parse-ns)
"Parse the DTD at point."
(forward-char (eval-when-compile (length "<!DOCTYPE")))
(skip-syntax-forward " ")
- (if (and (looking-at ">")
+ (if (and (looking-at-p ">")
xml-validating-parser)
(error "XML: (Validity) Invalid DTD (expecting name of the document)"))
(looking-at xml-name-re)
(let ((dtd (list (match-string-no-properties 0) 'dtd))
(xml-parameter-entity-alist xml-parameter-entity-alist)
- (parameter-entity-re (eval-when-compile
- (concat "%\\(" xml-name-re "\\);")))
next-parameter-entity)
(goto-char (match-end 0))
(skip-syntax-forward " ")
;; and try again.
(setq next-parameter-entity
(save-excursion
- (if (re-search-forward parameter-entity-re nil t)
+ (if (re-search-forward xml-pe-reference-re nil t)
(match-beginning 0))))
;; Parse the rest of the DTD
;; Fixme: Deal with NOTATION, PIs.
- (while (not (looking-at "\\s-*\\]"))
+ (while (not (looking-at-p "\\s-*\\]"))
(skip-syntax-forward " ")
(cond
((eobp)
(end-pos (match-end 0)))
;; Translation of rule [46] of XML specifications
(cond
- ((string-match "\\`EMPTY\\s-*\\'" type) ; empty declaration
+ ((string-match-p "\\`EMPTY\\s-*\\'" type) ; empty declaration
(setq type 'empty))
- ((string-match "\\`ANY\\s-*$" type) ; any type of contents
+ ((string-match-p "\\`ANY\\s-*$" type) ; any type of contents
(setq type 'any))
((string-match "\\`(\\(.*\\))\\s-*\\'" type) ; children ([47])
(setq type (xml-parse-elem-type
(match-string-no-properties 1 type))))
- ((string-match "^%[^;]+;[ \t\n\r]*\\'" type) ; substitution
+ ((string-match-p "^%[^;]+;[ \t\n\r]*\\'" type) ; substitution
nil)
(xml-validating-parser
(error "XML: (Validity) Invalid element type in the DTD")))
(goto-char (match-end 0)))
;; Comments (skip to end, ignoring parameter entity):
- ((looking-at "<!--")
+ ((looking-at-p "<!--")
(search-forward "-->")
(and next-parameter-entity
(> (point) next-parameter-entity)
(setq next-parameter-entity
(save-excursion
- (if (re-search-forward parameter-entity-re nil t)
+ (if (re-search-forward xml-pe-reference-re nil t)
(match-beginning 0))))))
;; Internal entity declarations:
(next-parameter-entity
(save-excursion
(goto-char next-parameter-entity)
- (unless (looking-at parameter-entity-re)
+ (unless (looking-at xml-pe-reference-re)
(error "XML: Internal error"))
(let* ((entity (match-string 1))
- (beg (point-marker))
(elt (assoc entity xml-parameter-entity-alist)))
(if elt
(progn
(goto-char next-parameter-entity))
(goto-char (match-end 0))))
(setq next-parameter-entity
- (if (re-search-forward parameter-entity-re nil t)
+ (if (re-search-forward xml-pe-reference-re nil t)
(match-beginning 0)))))
;; Anything else is garbage (ignored if not validating).
(progn
(setq elem (match-string-no-properties 1 string)
modifier (match-string-no-properties 2 string))
- (if (string-match "|" elem)
+ (if (string-match-p "|" elem)
(setq elem (cons 'choice
(mapcar 'xml-parse-elem-type
(split-string elem "|"))))
- (if (string-match "," elem)
+ (if (string-match-p "," elem)
(setq elem (cons 'seq
(mapcar 'xml-parse-elem-type
(split-string elem ",")))))))
(defun xml-substitute-special (string)
"Return STRING, after substituting entity and character references.
STRING is assumed to occur in an XML attribute value."
- (let ((ref-re (eval-when-compile
- (concat "&\\(?:#\\(x\\)?\\([0-9]+\\)\\|\\("
- xml-name-re "\\)\\);")))
- (strlen (length string))
+ (let ((strlen (length string))
children)
- (while (string-match ref-re string)
+ (while (string-match xml-entity-or-char-ref-re string)
(push (substring string 0 (match-beginning 0)) children)
(let* ((remainder (substring string (match-end 0)))
- (ref (match-string 2 string)))
+ (is-hex (match-string 1 string)) ; Is it a hex numeric reference?
+ (ref (match-string 2 string))) ; Numeric part of reference
(if ref
;; [4.6] Character references are included as
;; character data.
- (let ((val (decode-char 'ucs (string-to-number
- ref (if (match-string 1 string) 16)))))
+ (let ((val (decode-char 'ucs (string-to-number ref (if is-hex 16)))))
(push (cond (val (string val))
(xml-validating-parser
(error "XML: (Validity) Undefined character `x%s'" ref))
;; [4.4.5] Entity references are "included in literal".
;; Note that we don't need do anything special to treat
;; quotes as normal data characters.
- (setq ref (match-string 3 string))
+ (setq ref (match-string 3 string)) ; entity name
(let ((val (or (cdr (assoc ref xml-entity-alist))
(if xml-validating-parser
(error "XML: (Validity) Undefined entity `%s'" ref)
(if (and string (stringp string))
(let ((start 0))
(while (string-match "&#\\([0-9]+\\);" string start)
- (condition-case nil
- (setq string (replace-match
- (string (read (substring string
- (match-beginning 1)
- (match-end 1))))
- nil nil string))
- (error nil))
+ (ignore-errors
+ (setq string (replace-match
+ (string (read (substring string
+ (match-beginning 1)
+ (match-end 1))))
+ nil nil string)))
(setq start (1+ (match-beginning 0))))
string)
nil))
(defalias 'xml-print 'xml-debug-print)
(defun xml-escape-string (string)
- "Return STRING with entity substitutions made from `xml-entity-alist'."
- (mapconcat (lambda (byte)
- (let ((char (char-to-string byte)))
- (if (rassoc char xml-entity-alist)
- (concat "&" (car (rassoc char xml-entity-alist)) ";")
- char)))
- string ""))
+ "Convert STRING into a string containing valid XML character data.
+Replace occurrences of &<>\\='\" in STRING with their default XML
+entity references (e.g., replace each & with &).
+
+XML character data must not contain & or < characters, nor the >
+character under some circumstances. The XML spec does not impose
+restriction on \" or \\=', but we just substitute for these too
+\(as is permitted by the spec)."
+ (with-temp-buffer
+ (insert string)
+ (dolist (substitution '(("&" . "&")
+ ("<" . "<")
+ (">" . ">")
+ ("'" . "'")
+ ("\"" . """)))
+ (goto-char (point-min))
+ (while (search-forward (car substitution) nil t)
+ (replace-match (cdr substitution) t t nil)))
+ (buffer-string)))
(defun xml-debug-print-internal (xml indent-string)
"Outputs the XML tree in the current buffer.