1 /*

     2 * Copyright (c) 2005 Nokia Corporation and/or its subsidiary(-ies).

     3 * All rights reserved.

     4 * This component and the accompanying materials are made available

     5 * under the terms of the License "Eclipse Public License v1.0"

     6 * which accompanies this distribution, and is available

     7 * at the URL "http://www.eclipse.org/legal/epl-v10.html".

     8 *

     9 * Initial Contributors:

    10 * Nokia Corporation - initial contribution.

    11 *

    12 * Contributors:

    13 *

    14 * Description:  Misc. libxml2 related utilities.

    15 *

    16 */

    17 

    18 

    19 #ifndef XML_UTILS_H

    20 #define XML_UTILS_H

    21 

    22 // INCLUDES

    23 #include <e32std.h>

    24 #include <xmlengdom.h>

    25 

    26 // CONSTANTS

    27 

    28 // MACROS

    29 

    30 // DATA TYPES

    31 

    32 // FUNCTION PROTOTYPES

    33 

    34 // FORWARD DECLARATIONS

    35 class CXmlEncoding;

    36 class CXmlEntity;

    37 class RXmlEngDocument;

    38 class TXmlEngElement;

    39 

    40 // CLASS DECLARATION

    41 

    42 

    43 /**

    44 *  Misc. libxml2 related utilities.

    45 *

    46 *  \b Library: FeedsEngine.lib

    47 *

    48 *  @since 3.0

    49 */

    50 class CXmlUtils: public CBase

    51     {

    52     public: // Constructors and destructor

    53         /**

    54         * Two-phased constructor.

    55         */

    56         static CXmlUtils* NewL();

    57         

    58         /**

    59         * Destructor.

    60         */        

    61         virtual ~CXmlUtils();

    62 

    63 

    64     public:  // New methods

    65         /**

    66         * Returns a document from the provided buffer.

    67         *

    68         * @since 3.0

    69         * @param aBuffer A buffer in any char-encoding.

    70         * @param aCharSet The char-set.

    71         * @return A document.

    72         */

    73         RXmlEngDocument ParseBufferL(const TDesC8& aBuffer,

    74                 const TDesC& aCharSet) const;

    75 

    76         /**

    77         * Resolves any entities and escaped chars in the given url.

    78         *

    79         * @since 3.0

    80         * @param aUrl The url to clean up.

    81         * @return void.

    82         */

    83         void CleanupUrlL(TDes& aUrl) const;

    84 

    85         /**

    86         * Remove any markup found in the given descriptor.

    87         *

    88         * @since 3.0

    89         * @param aBuffer The buffer to clean up.

    90         * @param aNewLineChar The char used to insert a newline.

    91         * @return ETrue if markup was found in aOrig.

    92         */

    93         TBool CleanupMarkupL(TDes& aBuffer, TInt aNewLineChar) const;

    94 

    95         /**

    96         * Resolves any entities found in aOrig the named entity into its char-value.

    97         *

    98         * @since 3.0

    99         * @param aBuffer The buffer to clean up.

   100         * @return ETrue if entities were found in aOrig.

   101         */

   102         TBool ResolveEntitiesL(TDes& aBuffer) const;

   103 

   104         /**

   105         * Performs a deep extraction of the text children of the given node.  The result

   106         * is returned as a 16-bit descriptor.  The char encoding is always ucs2.

   107         *

   108         * @since 3.0

   109         * @param aElement A element.

   110         * @return The text or NULL.

   111         */

   112         HBufC* ExtractTextL(TXmlEngElement aElement) const;

   113 

   114         /**

   115         * Performs a deep extraction of the text children of the given node.  The result

   116         * is returned as a 16-bit descriptor.    The char encoding is always ucs2.  This

   117         * method is different from ExtractTextL in that it doesn't resolve entities or

   118         * do any other clean up.  It also has a length param.

   119         *

   120         * @since 3.1

   121         * @param aElement A element.

   122         * @param aMaxLength The max length to extract or 0 to extract the entire string.

   123         * @param aFromEnd If ETrue the aMaxLength chars are taken from the end of the string.

   124         * @return The text or NULL.

   125         */

   126         HBufC* ExtractSimpleTextL(TXmlEngElement aElement, TInt aMaxLength,

   127                 TBool aFromEnd = EFalse) const;

   128 

   129         /**

   130         * Returns the first element in the given document.

   131         *

   132         * @since 3.0

   133         * @param aDocument A document.

   134         * @return A element or NULL-Element.

   135         */

   136         TXmlEngElement GetDocumentFirstElement(RXmlEngDocument aDocument) const;

   137         

   138         /**

   139         * Returns the first child of the given node.

   140         *

   141         * @since 3.0

   142         * @param aElement A element.

   143         * @return A element or NULL-Element.

   144         */

   145         TXmlEngElement GetFirstElementChild(TXmlEngElement aElement) const;

   146 

   147         /**

   148         * Returns the first child of the given node with the provided name.  aUtf8Name

   149         * MUST be null terminated.

   150         *

   151         * @since 3.0

   152         * @param aElement A element.

   153         * @param aName The name of the node.

   154         * @return A element or NULL-Element.

   155         */

   156         TXmlEngElement GetFirstNamedChild(TXmlEngElement aElement, const TDesC8& aUtf8Name) const;

   157 

   158         /**

   159         * Returns the next sibling of the given node.

   160         *

   161         * @since 3.0

   162         * @param aNode A element.

   163         * @return A element or NULL-Element.

   164         */

   165         TXmlEngElement GetNextSiblingElement(TXmlEngElement aElement);

   166 

   167         /**

   168         * Returns true if the given node has a name of aName.  This method doesn't 

   169         * compare the namespace.

   170         *

   171         * @since 3.0

   172         * @param aElement A element.

   173         * @param aName The name of the node.

   174         * @return ETrue if they match.

   175         */

   176         TBool IsNamed(TXmlEngElement aElement, const TDesC8& aUtf8Name) const;

   177 

   178         /**

   179         * Returns true if the given node has match the given namespace and name.

   180         *

   181         * @since 3.0

   182         * @param aElement A element.

   183         * @param aNameSpace The namespace of the node.

   184         * @param aName The name of the node.

   185         * @return ETrue if they match.

   186         */

   187         TBool IsNamed(TXmlEngElement aElement, const TDesC8& aUtf8NameSpace, 

   188                 const TDesC8& aUtf8Name) const;

   189 

   190         /**

   191         * Returns the attribute's value or NULL if the attribute wasn't present.

   192         *

   193         * @since 3.0

   194         * @param aElement A element.

   195         * @param aAttribute The attribute to extract.

   196         * @return The value or NULL.

   197         */

   198         HBufC* AttributeL(TXmlEngElement aElement, const TDesC8& aAttribute) const;

   199 

   200 

   201     private:  // New methods

   202         /**

   203         * C++ default constructor.

   204         */

   205         CXmlUtils();

   206 

   207         /**

   208         * By default Symbian 2nd phase constructor is private.

   209         */

   210         void ConstructL();

   211                 

   212         /**

   213         * Cleanup stack callback method to cleanup ParseBufferL.

   214         *

   215         * @since 3.0

   216         * @param aPtr A libxml2 parser context ptr.

   217         * @return void.

   218         */

   219         static void CleanupParseBuffer(TAny *aPtr);

   220 

   221         /**

   222         * Skip any chars before the XML-prolog.

   223         *

   224         * @since 3.0

   225         * @param aBuffer A xml buffer.

   226         * @param aLen The length of the xml buffer.

   227         * @return void.

   228         */

   229         void SkipCharsBeforeXmlProlog(const TUint8** aString, TInt& aLen) const;

   230         

   231         /**

   232         * Remove any markup found in the given descriptor.

   233         *

   234         * @since 3.0

   235         * @param aBuffer The string to be stripped of markup.

   236         * @param aNewLineChar The char used to insert a newline.

   237         * @return ETrue if markup was found in aBuffer.

   238         */

   239         TBool StripMarkupL(TDes& aBuffer, TInt aNewLineChar) const;        

   240 

   241         /**

   242         * Remove any CDATA markers in the given descriptor.

   243         *

   244         * @since 3.1

   245         * @param aBuffer The string to be stripped of markup.

   246         * @return ETrue if CDATA markers were found in aBuffer.

   247         */

   248         TBool StripCDataMarkers(TDes& aBuffer) const;

   249         

   250         

   251     private:  // Data

   252         TLeakTracker  iLeakTracker;

   253         

   254         RXmlEngDOMImplementation iImpl;

   255         CXmlEncoding*  iXmlEncoding;

   256         CXmlEntity*    iXmlEntity;

   257     };

   258 

   259 #endif      // XML_UTILS_H

   260             

   261 // End of File