annotate lisp/nxml/rng-parse.el @ 112427:bc872de587fa

* play/landmark.el: Change `lm-' prefix to `landmark-' (Bug#7672). (lm): Rename to landmark. (lm-test-run): Rename to landmark-test-run.
author Chong Yidong <cyd@stupidchicken.com>
date Sat, 22 Jan 2011 15:12:51 -0500
parents ef719132ddfa
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
86361
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
1 ;;; rng-parse.el --- parse an XML file and validate it against a schema
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
2
112218
376148b31b5e Add 2011 to FSF/AIST copyright years.
Glenn Morris <rgm@gnu.org>
parents: 106815
diff changeset
3 ;; Copyright (C) 2003, 2007, 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
86361
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
4
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
5 ;; Author: James Clark
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
6 ;; Keywords: XML, RelaxNG
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
7
86551
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
8 ;; This file is part of GNU Emacs.
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
9
94666
d495d4d5452f Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 87665
diff changeset
10 ;; GNU Emacs is free software: you can redistribute it and/or modify
86551
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
11 ;; it under the terms of the GNU General Public License as published by
94666
d495d4d5452f Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 87665
diff changeset
12 ;; the Free Software Foundation, either version 3 of the License, or
d495d4d5452f Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 87665
diff changeset
13 ;; (at your option) any later version.
86361
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
14
86551
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
15 ;; GNU Emacs is distributed in the hope that it will be useful,
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
18 ;; GNU General Public License for more details.
86361
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
19
86551
f6a10b4524d2 Add 2007 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 86379
diff changeset
20 ;; You should have received a copy of the GNU General Public License
94666
d495d4d5452f Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 87665
diff changeset
21 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
86361
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
22
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
23 ;;; Commentary:
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
24
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
25 ;; This combines the validation machinery in rng-match.el with the
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
26 ;; parser in nxml-parse.el by using the `nxml-validate-function' hook.
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
27
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
28 ;;; Code:
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
29
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
30 (require 'nxml-parse)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
31 (require 'rng-match)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
32 (require 'rng-dt)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
33
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
34 (defvar rng-parse-prev-was-start-tag nil)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
35
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
36 (defun rng-parse-validate-file (schema file)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
37 "Parse and validate the XML document in FILE and return it as a list.
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
38 The returned list has the same form as that returned by
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
39 `nxml-parse-file'. SCHEMA is a list representing the schema to use
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
40 for validation, such as returned by the function `rng-c-load-schema'.
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
41 If the XML document is invalid with respect to schema, an error will
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
42 be signaled in the same way as when it is not well-formed."
105866
3367f0022cf2 * nxml/xsd-regexp.el (xsdre-gen-categories):
Stefan Monnier <monnier@iro.umontreal.ca>
parents: 100908
diff changeset
43 (with-current-buffer (nxml-parse-find-file file)
86361
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
44 (unwind-protect
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
45 (let ((nxml-parse-file-name file)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
46 (nxml-validate-function 'rng-parse-do-validate)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
47 (rng-dt-namespace-context-getter '(nxml-ns-get-context))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
48 rng-parse-prev-was-start-tag)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
49 ;; We don't simply call nxml-parse-file, because
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
50 ;; we want to do rng-match-with-schema in the same
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
51 ;; buffer in which we will call the other rng-match-* functions.
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
52 (rng-match-with-schema schema
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
53 (nxml-parse-instance)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
54 (kill-buffer nil))))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
55
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
56 (defun rng-parse-do-validate (text start-tag)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
57 (cond ((and (let ((tem rng-parse-prev-was-start-tag))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
58 (setq rng-parse-prev-was-start-tag (and start-tag t))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
59 tem)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
60 (not start-tag)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
61 (rng-match-text-typed-p))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
62 (unless (rng-match-element-value (or text ""))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
63 (cons "Invalid data" (and text 'text))))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
64 ((and text
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
65 (not (rng-blank-p text))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
66 (not (rng-match-mixed-text)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
67 (cons "Text not allowed" 'text))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
68 ((not start-tag)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
69 (unless (rng-match-end-tag)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
70 (cons "Missing elements" nil)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
71 ((not (rng-match-start-tag-open
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
72 (rng-parse-to-match-name (car start-tag))))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
73 (cons "Element not allowed" nil))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
74 (t
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
75 (let ((atts (cadr start-tag))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
76 (i 0)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
77 att err)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
78 (while (and atts (not err))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
79 (setq att (car atts))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
80 (when (not (and (consp (car att))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
81 (eq (caar att) nxml-xmlns-namespace-uri)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
82 (setq err
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
83 (cond ((not (rng-match-attribute-name
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
84 (rng-parse-to-match-name (car att))))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
85 (cons "Attribute not allowed"
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
86 (cons 'attribute-name i)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
87 ((not (rng-match-attribute-value (cdr att)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
88 (cons "Invalid attribute value"
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
89 (cons 'attribute-value i))))))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
90 (setq atts (cdr atts))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
91 (setq i (1+ i)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
92 (or err
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
93 (unless (rng-match-start-tag-close)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
94 (cons "Missing attributes" 'tag-close)))))))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
95
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
96 (defun rng-parse-to-match-name (name)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
97 (if (consp name)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
98 name
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
99 (cons nil name)))
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
100
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
101 (provide 'rng-parse)
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
102
38f93f3d00a2 Initial merge of nxml
Mark A. Hershberger <mah@everybody.org>
parents:
diff changeset
103 ;;; rng-parse.el ends here