UTF-8 fix, thanks to Francis Leboutte

doc/GNUmakefile

@@ -2,5 +2,5 @@ all: dom.html index.html installation.html klacks.html quickstart.html sax.html
 
 %.html: %.xml html.xsl
 	xsltproc html.xsl $< >$@.tmp
-	chmod -w *.html
 	mv $@.tmp $@
+	chmod -w $@

doc/index.xml

@@ -4,26 +4,13 @@
     <p>An XML parser written in Common Lisp.</p>
 
     <p>
-      Closure XML was written by <a
+      Closure XML was written
-      href="http://www.stud.uni-karlsruhe.de/~unk6/">Gilbert Baumann</a>
+      by <a href="http://www.stud.uni-karlsruhe.de/~unk6/">Gilbert
-      (unk6 at rz.uni-karlsruhe.de) as part of the Closure web
+      Baumann</a> as part of the Closure web browser and is now
-      browser.<br/>
+      maintained by
-      Contributions to the parser by
+      <a href="mailto:david@lichteblau.com">David Lichteblau</a>.
+      It is licensed under Lisp-LGPL.
     </p>      
-    <ul>
-      <li>
-        Henrik Motakef (hmot at henrik-motakef.de)<br/>
-        (SAX layer; namespace support)
-      </li>
-      <li>
-        <a href="mailto:david@lichteblau.com">David Lichteblau</a> for <a
-        href="http://www.knowledgetools.de">knowledgeTools</a>
-        (conversion into an independent package; DOM bug fixing; validation)
-        and <a href="http://www.headcraft.de/">headcraft</a>
-        (most september/october 2004 changes) and privately (changes
-        since then).
-      </li>
-    </ul>
 
     <p>
       CXML implements a <a
@@ -36,10 +23,6 @@
       other similar to StAX.
     </p>
 
-    <p>
-      CXML is licensed under Lisp-LGPL.
-    </p>
-
     <p>
       Send bug reports to <a
       href="mailto:cxml-devel@common-lisp.net">cxml-devel@common-lisp.net</a>
@@ -48,15 +31,31 @@
       information</a>).
     </p>
 
-    <h3>See also</h3>
+    <h3>Add-on features</h3>
     <p>
-      Relax NG validation is available as a separate
+      The following libraries are available as separate downloads:
-      project: <a href="http://www.lichteblau.com/cxml-rng/">cxml-rng</a>.
+    </p>
+    <p>
+      &#x2b17;&#160;
+      <a href="http://www.lichteblau.com/cxml-rng/">cxml-rng</a>
+      &#160;&#160;
+      Relax NG validation
+    </p>
+    <p>
+      &#x2b17;&#160;
+      <a href="http://www.lichteblau.com/cxml-stp/">cxml-stp</a>
+      &#160;&#160;
+      STP, an alternative to DOM
     </p>
 
 
     <a name="changes"/>
     <h3>Recent Changes</h3>
+    <p class="nomargin"><tt>rel-2007-xx-yy</tt></p>
+    <ul class="nomargin">
+      <li>Various DTD serialization fixes</li>
+      <li>UTF-8 fix, thanks to Francis Leboutte</li>
+    </ul>
     <p class="nomargin"><tt>rel-2007-07-07</tt></p>
     <ul class="nomargin">
       <li>

dom/dom-builder.lisp

@@ -135,7 +135,7 @@
       (cond
         ((eq (dom:node-type parent) :cdata-section)
           (setf (dom:data parent) data))
-        ((and last-child (eq (dom:node-type last-child) :text))
+	((and last-child (eq (dom:node-type last-child) :text))
           ;; um entities herum wird SAX:CHARACTERS mehrfach aufgerufen fuer
           ;; den gleichen Textknoten.  Hier muessen wir den bestehenden Knoten
           ;; erweitern, sonst ist das Dokument nicht normalisiert.

runes/encodings.lisp

@@ -250,7 +250,7 @@
                (setf rptr (%+ rptr 1)))
             
               ((%<= #|#b11000000|# byte0 #b11011111)
-               (cond ((< (%+ rptr 2) in-end)
+               (cond ((<= (%+ rptr 2) in-end)
                       (put
                        (dpb (ldb (byte 5 0) byte0) (byte 5 6)
                             (dpb (ldb (byte 6 0) (aref in (%+ rptr 1))) (byte 6 0)
@@ -260,7 +260,7 @@
                       (return))))
             
               ((%<= #|#b11100000|# byte0 #b11101111)
-               (cond ((< (%+ rptr 3) in-end)
+               (cond ((<= (%+ rptr 3) in-end)
                       (put
                        (dpb (ldb (byte 4 0) byte0) (byte 4 12)
                             (dpb (ldb (byte 6 0) (aref in (%+ 1 rptr))) (byte 6 6)
@@ -271,7 +271,7 @@
                       (return))))
             
               ((%<= #|#b11110000|# byte0 #b11110111)
-               (cond ((< (%+ rptr 4) in-end)
+               (cond ((<= (%+ rptr 4) in-end)
                       (put
                        (dpb (ldb (byte 3 0) byte0) (byte 3 18)
                             (dpb (ldb (byte 6 0) (aref in (%+ 1 rptr))) (byte 6 12)
@@ -283,7 +283,7 @@
                       (return))))
             
               ((%<= #|#b11111000|# byte0 #b11111011)
-               (cond ((< (%+ rptr 5) in-end)
+               (cond ((<= (%+ rptr 5) in-end)
                       (put
                        (dpb (ldb (byte 2 0) byte0) (byte 2 24)
                             (dpb (ldb (byte 6 0) (aref in (%+ 1 rptr))) (byte 6 18)
@@ -296,7 +296,7 @@
                       (return))))
             
               ((%<= #|#b11111100|# byte0 #b11111101)
-               (cond ((< (%+ rptr 6) in-end)
+               (cond ((<= (%+ rptr 6) in-end)
                       (put
                        (dpb (ldb (byte 1 0) byte0) (byte 1 30)
                             (dpb (ldb (byte 6 0) (aref in (%+ 1 rptr))) (byte 6 24)

test/misc.lisp

@@ -0,0 +1,29 @@
+;;;
+;;; When I'll grow up, I'll be a complete test suite.
+
+(deftest utf-8
+    (flet ((doit (from below)
+	     (loop for code from from below below do
+		  (when (and (code-char code)
+			     (not (eql code #xfffe))
+			     (not (eql code #xffff)))
+		    (let* ((a (if (< code #x10000)
+				  (format nil "abc~C" (code-char code))
+				  (let* ((x (- code #x10000))
+					 (lo (ldb (byte 10 0) x))
+					 (hi (ldb (byte 10 10) x)))
+				    (format nil "abc~C~C"
+					    (code-char (logior #xD800 hi))
+					    (code-char 
+					     (logior #xDC00 lo))))))
+			   (b (cxml:utf8-string-to-rod
+			       (cxml:rod-to-utf8-string
+				a))))
+		      (unless (string= a b)
+			(format t "FAIL: ~S ~A ~A~%"
+				(code-char code)
+				(map 'vector #'char-code a)
+				(map 'vector #'char-code b))))))))
+      (doit 32 #xD800)
+      (doit #x10000 char-code-limit)
+      (values)))