akregator-librss: use namespacesimprovement/akregator-librss-namespaces

Check element namespace rather than assuming a prefix, which can result
in incorrect metadata displaying.

Signed-off-by: mio <[email protected]>

6 files changed, 93 insertions, 25 deletions

diff --git a/akregator/src/librss/article.cpp b/akregator/src/librss/article.cpp
index 18522fe3e..88d42a7dc 100644
--- a/akregator/src/librss/article.cpp
+++ b/akregator/src/librss/article.cpp
@@ -92,13 +92,16 @@ Article::Article(const TQDomNode &node, Format format, Version version) : d(new
 			d->link = elemText;
 	}
 
+	// prefer content/content:encoded over summary/description for feeds that provide it
+	if (format == AtomFeed)
+	{
+		d->description = extractNode(node, TQString::fromLatin1("content"), false);
+	}
+	else
+	{
+		d->description = extractElementTextNS(node, ContentNamespace, TQString::fromLatin1("encoded"), false);
+	}
 
-    // prefer content/content:encoded over summary/description for feeds that provide it
-    TQString tagName=(format==AtomFeed)? TQString::fromLatin1("content"): TQString::fromLatin1("content:encoded");
-    
-    if (!(elemText = extractNode(node, tagName, false)).isNull())
-        d->description = elemText;
-    
     if (d->description.isEmpty())
     {
 		if (!(elemText = extractNode(node, TQString::fromLatin1("body"), false)).isNull())
@@ -130,7 +133,7 @@ Article::Article(const TQDomNode &node, Format format, Version version) : d(new
 			time = KRFCDate::parseDate(elemText);
 	}
 	
-	if (!(elemText = extractNode(node, TQString::fromLatin1("dc:date"))).isNull())
+	if (!(elemText = extractElementTextNS(node, DublinCoreNamespace, TQString::fromLatin1("date"))).isNull())
 	{
 		time = parseISO8601Date(elemText);
 	}
@@ -139,27 +142,22 @@ Article::Article(const TQDomNode &node, Format format, Version version) : d(new
         if (time != 0)
 		d->pubDate.setTime_t(time);
 
-	if (!(elemText = extractNode(node, TQString::fromLatin1("wfw:comment"))).isNull()) {
-		d->commentsLink = elemText;
-	}
-
-    if (!(elemText = extractNode(node, TQString::fromLatin1("slash:comments"))).isNull()) {
-        d->numComments = elemText.toInt();
-    }
+	d->commentsLink = extractElementTextNS(node, CommentAPINamespace, TQString::fromLatin1("comment"));
+	d->numComments = extractElementTextNS(node, SlashNamespace, TQString::fromLatin1("comments")).toInt();
 
     TQDomElement element = TQDomNode(node).toElement();
 
     // in RSS 1.0, we use <item about> attribute as ID
     // FIXME: pass format version instead of checking for attribute
 
-    if (!element.isNull() && element.hasAttribute(TQString::fromLatin1("rdf:about")))
+    if (!element.isNull() && element.hasAttributeNS(RDFNamespace, TQString::fromLatin1("about")))
     {
-        d->guid = element.attribute(TQString::fromLatin1("rdf:about")); // HACK: using ns properly did not work
+        d->guid = element.attributeNS(RDFNamespace, TQString::fromLatin1("about"), TQString::null);
         d->guidIsPermaLink = false;
     }
     else
     {
-        tagName=(format==AtomFeed)? TQString::fromLatin1("id"): TQString::fromLatin1("guid");
+        TQString tagName=(format==AtomFeed)? TQString::fromLatin1("id"): TQString::fromLatin1("guid");
         TQDomNode n = node.namedItem(tagName);
 	    if (!n.isNull())
         {
diff --git a/akregator/src/librss/document.cpp b/akregator/src/librss/document.cpp
index 3bc64d000..7d94a252a 100644
--- a/akregator/src/librss/document.cpp
+++ b/akregator/src/librss/document.cpp
@@ -224,7 +224,7 @@ Document::Document(const TQDomDocument &doc) : d(new Private)
         d->copyright = elemText;
 
     if (d->format == AtomFeed)
-        elemText = rootNode.toElement().attribute(TQString::fromLatin1("xml:lang"), TQString());
+        elemText = rootNode.toElement().attributeNS(XMLNamespace, "lang", TQString::null);
     else
         elemText = extractNode(channelNode, TQString::fromLatin1("language"));
 
@@ -441,7 +441,7 @@ Document::Document(const TQDomDocument &doc) : d(new Private)
         d->pubDate.setTime_t(_time);
     }
 
-    if (!(elemText = extractNode(channelNode, TQString::fromLatin1("dc:date"))).isNull()) {
+    if (!(elemText = extractElementTextNS(channelNode, DublinCoreNamespace, "date")).isNull()) {
         time_t _time = parseISO8601Date(elemText);
         /* \bug This isn't really the right way since it will set the date to
          * Jan 1 1970, 1:00:00 if the passed date was invalid; this means that
diff --git a/akregator/src/librss/global.h b/akregator/src/librss/global.h
index 16c08178c..8ff4e406d 100644
--- a/akregator/src/librss/global.h
+++ b/akregator/src/librss/global.h
@@ -18,6 +18,21 @@ class TQValueList;
 
 namespace RSS
 {
+	/// The Atom 1.0 XML namespace.
+	constexpr const char *AtomNamespace = "http://www.w3.org/2005/Atom";
+	/// The CommentAPI XML namespace.
+	constexpr const char *CommentAPINamespace = "http://wellformedweb.org/CommentAPI/";
+	/// The Content XML namespace.
+	constexpr const char *ContentNamespace = "http://purl.org/rss/1.0/modules/content/";
+	/// The Dublin Core XML namespace.
+	constexpr const char *DublinCoreNamespace = "http://purl.org/dc/elements/1.1/";
+	/// The RDF Concepts Vocabulary (RDF) namespace.
+	constexpr const char *RDFNamespace = "http://www.w3.org/1999/02/22-rdf-syntax-ns#";
+	/// The Slash XML namespace.
+	constexpr const char *SlashNamespace = "http://purl.org/rss/1.0/modules/slash/";
+	/// The XML namespace.
+	constexpr const char *XMLNamespace = "http://www.w3.org/XML/1998/namespace";
+
 	/**
 	 * Versions currently supported by this library. This enumeration is
 	 * subject to be extended in the future and used by Document::version() to
diff --git a/akregator/src/librss/loader.cpp b/akregator/src/librss/loader.cpp
index 75dd22fd8..8674dfb79 100644
--- a/akregator/src/librss/loader.cpp
+++ b/akregator/src/librss/loader.cpp
@@ -351,7 +351,7 @@ void Loader::slotRetrieverDone(const TQByteArray &data, bool success)
       TQByteArray tmpData;
       tmpData.setRawData(charData, len);
 
-      if (doc.setContent(tmpData))
+      if (doc.setContent(tmpData, /* namespaceProcessing */ true))
       {
          rssDoc = Document(doc);
          if (!rssDoc.isValid())
diff --git a/akregator/src/librss/tools_p.cpp b/akregator/src/librss/tools_p.cpp
index 9303bdf50..04dc570b4 100644
--- a/akregator/src/librss/tools_p.cpp
+++ b/akregator/src/librss/tools_p.cpp
@@ -117,6 +117,59 @@ static TQString extractAtomContent(const TQDomElement& e)
     return TQString();
 }
 
+TQDomElement extractElementNS(const TQDomNode &parent, const TQString &nameSpace, const TQString &localName)
+{
+    TQDomElement element;
+
+    if (parent.isNull())
+    {
+        return element;
+    }
+
+    TQDomNodeList children = parent.childNodes();
+    for (size_t i = 0; i < children.count(); ++i)
+    {
+        TQDomNode node = children.item(i);
+        if (node.isElement() && node.namespaceURI() == nameSpace && node.localName() == localName)
+        {
+            element = node.toElement();
+            break;
+        }
+    }
+
+    return element;
+}
+
+TQString extractElementTextNS(const TQDomNode &parent, const TQString &namespaceURI, const TQString &localName, bool isInlined)
+{
+    TQDomElement element = extractElementNS(parent, namespaceURI, localName);
+
+    if (element.isNull())
+    {
+        return TQString::null;
+    }
+
+    TQString result = element.text().stripWhiteSpace();
+    if (localName == "content")
+    {
+        // Atom content
+        result = extractAtomContent(element);
+    }
+    else
+    {
+        // Check for HTML; not necessary for atom:content
+        // Taken from extractNode below
+        bool hasPre = result.contains("<pre>", false) || result.contains("<pre ", false);
+        bool hasHtml = hasPre || result.contains("<");
+        if (!isInlined && !hasHtml)
+            result = result = result.replace(TQChar('\n'), "<br />");
+        if (!hasPre)
+            result = result.simplifyWhiteSpace();
+    }
+
+    return result.isEmpty() ? TQString::null : result;
+}
+
 TQString extractNode(const TQDomNode &parent, const TQString &elemName, bool isInlined)
 {
 	TQDomNode node = parent.namedItem(elemName);
@@ -208,25 +261,25 @@ TQString parseItemAuthor(const TQDomElement& element, Format format, Version ver
     TQString name;
     TQString email;
 
-    TQDomElement dcCreator = element.namedItem("dc:creator").toElement();
-    
+    TQDomElement dcCreator = extractElementNS(element, DublinCoreNamespace, "creator");
+
     if (!dcCreator.isNull())
          authorFromString(dcCreator.text(), name, email);
     else if (format == AtomFeed)
     {
         TQDomElement atomAuthor = element.namedItem("author").toElement();
         if (atomAuthor.isNull())
-            atomAuthor = element.namedItem("atom:author").toElement();
+            atomAuthor = extractElementNS(element, AtomNamespace, "author");
         if (!atomAuthor.isNull())
         {
             TQDomElement atomName = atomAuthor.namedItem("name").toElement();
             if (atomName.isNull())
-                atomName = atomAuthor.namedItem("atom:name").toElement();
+                atomName = extractElementNS(atomAuthor, AtomNamespace, "name");
             name = atomName.text().stripWhiteSpace();
             
             TQDomElement atomEmail = atomAuthor.namedItem("email").toElement();
             if (atomEmail.isNull())
-                atomEmail = atomAuthor.namedItem("atom:email").toElement();
+                atomEmail = extractElementNS(atomAuthor, AtomNamespace, "email");
             email = atomEmail.text().stripWhiteSpace();
         }
     }
diff --git a/akregator/src/librss/tools_p.h b/akregator/src/librss/tools_p.h
index 0ec9013f3..1b89fc85d 100644
--- a/akregator/src/librss/tools_p.h
+++ b/akregator/src/librss/tools_p.h
@@ -29,6 +29,8 @@ namespace RSS
 		unsigned int count;
 	};
 
+	TQDomElement extractElementNS(const TQDomNode& parent, const TQString& namespaceURI, const TQString& localName);
+	TQString extractElementTextNS(const TQDomNode& parent, const TQString& namespaceURI, const TQString& localName, bool isInlined = true);
 	TQString extractNode(const TQDomNode &parent, const TQString &elemName, bool isInlined=true);
 	TQString extractTitle(const TQDomNode &parent);
 	TQString childNodesAsXML(const TQDomNode& parent);