[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-commits
Subject:    kdenetwork/librss
From:       Stanislav Karchebny <berk () upnet ! ru>
Date:       2004-04-07 16:19:05
Message-ID: 20040407161905.2C0139A08 () office ! kde ! org
[Download RAW message or body]

CVS commit by karchebny: 


* Make librss _really_ work.

Before, XML constructs like &amp;lt;b&amp;gt;test&lt;/b&gt; would be incorrectly
processed to become <b>test</b>. The error resulted from QDom unescaping these
entities for us, and then librss doing that over again. The proper unescaping
would be &lt;b&gt;test</b>, which can be achieved by simply disabling decodeEntities
by default.

My proposal is to get rid of decodeEntities and boolean flag to extractNode \
altogether.

CCMAIL: raabe@kde.org


  M +4 -0      tools_p.cpp   1.7
  M +1 -1      tools_p.h   1.5


--- kdenetwork/librss/tools_p.cpp  #1.6:1.7
@@ -16,7 +16,11 @@
 #include <qdom.h>
 
+/* <berkus> frerich: please explain in what circumstances this function might be \
needed +            and if there are none, i think its safe to just remove it \
altogether. +*/
 QString RSS::decodeEntities(const QString &s)
 {
         QString result = s;
+        // btw: next line is wrong, all wrong (1st: &amp; not &amp 2nd: QDom does it \
                all for us)
         result.replace(QString::fromLatin1("&amp"), QString::fromLatin1("&"));
         for (int p = result.find(QString::fromLatin1("&")); p >= 0; p = \
result.find(QString::fromLatin1("&"), p)) {

--- kdenetwork/librss/tools_p.h  #1.4:1.5
@@ -28,5 +28,5 @@ namespace RSS
 
         QString decodeEntities(const QString &s);
-        QString extractNode(const QDomNode &parent, const QString &elemName, bool \
entities = true); +        QString extractNode(const QDomNode &parent, const QString \
&elemName, bool entities = false);  }
 


[prev in list] [next in list] [prev in thread] [next in thread]