space.lisp

   1 ;;; -*- show-trailing-whitespace: t; indent-tabs: nil -*-
   2
   3 ;;; Copyright (c) 2007,2008 David Lichteblau, Ivan Shvedunov.
   4 ;;; All rights reserved.
   5
   6 ;;; Redistribution and use in source and binary forms, with or without
   7 ;;; modification, are permitted provided that the following conditions
   8 ;;; are met:
   9 ;;;
  10 ;;;   * Redistributions of source code must retain the above copyright
  11 ;;;     notice, this list of conditions and the following disclaimer.
  12 ;;;
  13 ;;;   * Redistributions in binary form must reproduce the above
  14 ;;;     copyright notice, this list of conditions and the following
  15 ;;;     disclaimer in the documentation and/or other materials
  16 ;;;     provided with the distribution.
  17 ;;;
  18 ;;; THIS SOFTWARE IS PROVIDED BY THE AUTHOR 'AS IS' AND ANY EXPRESSED
  19 ;;; OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  20 ;;; WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  21 ;;; ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
  22 ;;; DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  23 ;;; DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE
  24 ;;; GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
  25 ;;; INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
  26 ;;; WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
  27 ;;; NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
  28 ;;; SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  29
  30
  31 ;;; This file implements whitespace stripping.
  32 ;;;
  33 ;;; Although the spec presents a unified algorithm for whitespace stripping
  34 ;;; of stylesheets and source documents, we implement them separately.
  35 ;;;
  36 ;;; For stylesheets, the STP parse tree of the stylesheet is modified
  37 ;;; directly according the its xml:space declarations and xsl:text elements.
  38 ;;;
  39 ;;; For source documents, the strip-space and preserve-space declarations
  40 ;;; from the stylesheet are taken into account.  To avoid processing
  41 ;;; parts of the document that XPath would not otherwise have navigated
  42 ;;; to, we do whitespace stripping lazily using a proxy implementation
  43 ;;; of the XPath protocol.
  44
  45 (in-package :xuriella)
  46
  47 #+sbcl
  48 (declaim (optimize (debug 2)))
  49
  50
  51 ;;;; Helper functions
  52
  53 (eval-when (:compile-toplevel :load-toplevel :execute)
  54   (defparameter *whitespace*
  55     (format nil "~C~C~C~C"
  56             (code-char 9)
  57             (code-char 32)
  58             (code-char 13)
  59             (code-char 10))))
  60
  61 (defun normalize-whitespace (str)
  62   (cl-ppcre:regex-replace-all #.(format nil "[~A]+" *whitespace*)
  63                               (string-trim *whitespace* str)
  64                               " "))
  65
  66 (defun whitespacep (str)
  67   (cl-ppcre:all-matches #.(format nil "^[~A]+$" *whitespace*) str))
  68
  69 (defun words (str)
  70   (when str
  71     (cl-ppcre:split #.(format nil "[~A]+" *whitespace*)
  72                     (string-trim *whitespace* str))))
  73
  74
  75 ;;;; Strip whitespace in stylesheets
  76
  77 ;; Also strips comments and PIs.
  78 (defun strip-stylesheet (parent &optional preserve)
  79   (let ((i 0))
  80     (loop while (< i (length (cxml-stp-impl::%children parent))) do
  81          (let ((child (stp:nth-child i parent)))
  82            (etypecase child
  83              (stp:text
  84               (if (and (whitespacep (stp:data child))
  85                        (not preserve))
  86                   (stp:delete-nth-child i parent)
  87                   (incf i)))
  88              ((or stp:comment stp:processing-instruction)
  89               (stp:delete-nth-child i parent))
  90              (stp:element
  91                  (stp:with-attributes ((space "space" *xml*))
  92                      child
  93                    (let ((new-preserve
  94                           (cond
  95                             ((namep child "text") t)
  96                             ((not space) preserve)
  97                             ((equal space "preserve") t)
  98                             (t nil))))
  99                      (strip-stylesheet child new-preserve)))
 100                (incf i)))))))
 101
 102
 103 ;;;; Strip whitespace in source documents
 104
 105 (defun make-whitespace-stripper (node tests)
 106   (if tests
 107       (make-stripping-node nil node tests nil)
 108       node))
 109
 110 (defstruct stripping-node
 111   parent
 112   target
 113   children)
 114
 115 (defstruct (leaf-stripping-node
 116              (:constructor make-leaf-stripping-node (parent target))
 117              (:include stripping-node)))
 118
 119 (defstruct (parent-stripping-node
 120              (:constructor make-parent-stripping-node (parent target))
 121              (:include stripping-node)))
 122
 123 (defmethod print-object ((object stripping-node) stream)
 124   (print-unreadable-object (object stream :type t :identity nil)
 125     (let ((target (write-to-string (stripping-node-target object))))
 126       (if (and (alexandria:starts-with-subseq target "#<")
 127                (alexandria:ends-with #\> target))
 128           (write-sequence target stream :start 3 :end (1- (length target)))
 129           (write-sequence target stream)))))
 130
 131 (defun strip-under-qname-p (node tests)
 132   (let ((local-name (xpath-protocol:local-name node))
 133         (uri (xpath-protocol:namespace-uri node)))
 134     (dolist (test tests nil)
 135       (let ((result (funcall test local-name uri)))
 136         (when result
 137           (return (eq result :strip)))))))
 138
 139 (defun xpath-protocol/attribute-value (node local-name uri)
 140   (do-pipe (a (xpath-protocol:attribute-pipe node))
 141     (when (and (equal (xpath-protocol:local-name a) local-name)
 142                (equal (xpath-protocol:namespace-uri a) uri))
 143       (return (xpath-protocol:string-value a)))))
 144
 145 (defun make-stripping-node (parent target tests force-preserve)
 146   (let ((result (make-parent-stripping-node parent target))
 147         (xml-space (xpath-protocol/attribute-value target "space" *xml*)))
 148     (when xml-space
 149       (setf force-preserve (equal xml-space "preserve")))
 150     (labels ((recurse (child-node)
 151                (if (xpath-protocol:node-type-p child-node :element)
 152                    (make-stripping-node result child-node tests force-preserve)
 153                    (make-leaf-stripping-node result child-node)))
 154              (maybe-recurse (child-node)
 155                (if (and (xpath-protocol:node-type-p child-node :text)
 156                         (whitespacep (xpath-protocol:string-value child-node)))
 157                    nil
 158                    (recurse child-node))))
 159       (let ((all-children (xpath-protocol:child-pipe target)))
 160         (setf (stripping-node-children result)
 161               (if (or force-preserve
 162                       (not (xpath-protocol:node-type-p target :element))
 163                       (not (strip-under-qname-p target tests)))
 164                   (xpath::map-pipe-filtering #'recurse all-children)
 165                   (xpath::map-pipe-filtering #'maybe-recurse all-children)))))
 166     result))
 167
 168 (macrolet ((defproxy (name &rest args)
 169              `(defmethod ,name ((node stripping-node) ,@args)
 170                 (,name (stripping-node-target node) ,@args))))
 171   (defproxy xpath-protocol:local-name)
 172   (defproxy xpath-protocol:namespace-uri)
 173   (defproxy xpath-protocol:namespace-prefix)
 174   (defproxy xpath-protocol:qualified-name)
 175   (defproxy xpath-protocol:attribute-pipe)
 176   (defproxy xpath-protocol:namespace-pipe)
 177   (defproxy xpath-protocol:node-type-p type))
 178
 179 (defmethod xpath-protocol:node-p ((node stripping-node))
 180   t)
 181
 182 (defmethod xpath-protocol:child-pipe ((node stripping-node))
 183   (stripping-node-children node))
 184
 185 (defmethod xpath-protocol:parent-node ((node stripping-node))
 186   (stripping-node-parent node))
 187
 188 (defmethod xpath-protocol:string-value ((node stripping-node))
 189   (with-output-to-string (s)
 190     (write-string-value node s)))
 191
 192 (defmethod write-string-value ((node parent-stripping-node) stream)
 193   (do-pipe (child (xpath-protocol:child-pipe node))
 194     (unless (or (xpath-protocol:node-type-p child :comment)
 195                 (xpath-protocol:node-type-p child :processing-instruction))
 196       (write-string-value child stream))))
 197
 198 (defmethod write-string-value ((node leaf-stripping-node) stream)
 199   (write-string-value (stripping-node-target node) stream))
 200
 201 (defmethod write-string-value (node stream)
 202   (write-string (xpath-protocol:string-value node) stream))
 203
 204
 205 ;;;; TEXT NORMALIZER, from cxml-rng
 206
 207 ;;; FIXME: cxml should do that
 208
 209 (defun make-text-normalizer (next)
 210   (make-instance 'text-normalizer :chained-handler next))
 211
 212 (defclass text-normalizer (cxml:sax-proxy)
 213   ((pending-text-node :initform (make-string-output-stream)
 214                       :accessor pending-text-node)))
 215
 216 (defmethod sax:characters ((handler text-normalizer) data)
 217   (write-string data (pending-text-node handler)))
 218
 219 (defun flush-pending (handler)
 220   (let ((str (get-output-stream-string (pending-text-node handler))))
 221     (unless (zerop (length str))
 222       (sax:characters (cxml:proxy-chained-handler handler) str))))
 223
 224 (defmethod sax:start-element :before
 225     ((handler text-normalizer) uri lname qname attributes)
 226   (declare (ignore uri lname qname attributes))
 227   (flush-pending handler))
 228
 229 (defmethod sax:end-element :before
 230     ((handler text-normalizer) uri lname qname)
 231   (declare (ignore uri lname qname))
 232   (flush-pending handler))