NativeXml: A native Delphi XML parser and writer

http://www.simdesign.nl/xml.html

This software component contains a small-footprint Object Pascal (Delphi) XML implementation that allows to read and write XML documents. You basically only need one unit and you can simply add it to the "uses" clause.

You can use this software to read XML documents from files, streams or strings. The load routine generates events that can be used to display load progress on the fly. You can also use it to create and save XML documents.

Features

• Read XML directly from string, stream or file with fast parser (1Mb/sec on 1.5GHz P3), writes to string, stream and file even faster.
  
  
• Import of all common tags, including comments, CDATA, XML declarations and stylesheets.
  
  
• Load from streaming media is supported; you can set up events to work with the node data while it comes in. See example 1. No need to keep the whole XML document in memory! Works well for huge documents. This is the SAX-like approach.
  
  
• Save to file, string or stream in readable, preserve or compact format. Readable format will add linefeeds and indents to make the file readable in a text editor. Compact format sticks to the Xml specification and adds minimal control characters. Preserve format keeps the markup of the document exactly as it was.
  
  
• Directly add blobs or streams in the XML file, they will be encoded as Binhex or Base64 (but you don't have to worry about that).
  
  
• Many native Delphi types can be stored directly to the nodes, making this code very suitable to use while writing storage methods.
  Store and load boolean, single, double, word, integer, string, TFont, TColor, TPen, TBrush, TDateTime.
  You don't have to worry about escaping the strings; this is done automatically.
  
  
• Loads and saves ANSI, UTF-8 as well as Unicode (UTF-16) documents and implements proper conversion warnings. Recognises streams even without byte order marks. NativeXml works internally with UTF8 encoded strings to preserve extended characters, and supports widestrings.
  
  
• Unique feature: Store, read and create any TPersistent object to/from XML directly (see Example5). This is done by iterating through all of the objects' published properties by use of RTTI (runtime type information). This feature is only available for D5 and up.
• New! Now you can use the binary format of NativeXml. This avoids the expensive parsing process to get the textual xml data into the document (DOM). And even better, the binary file size is much smaller, ~50% of file size with uncompressed binary files, and ~15% of the file size in zlib-compressed files. And you can use AES encryption to keep your sensitive files safe from prying eyes.
  The interface is the same whether textual xml or binary xml files are used, so you can work with binary files for speed and size purposes, then always send them to the public as normal xml files. 

  // TsdElementType enumerates the different kinds of elements that can be found
  // in the XML document.
  TsdElementType = (
    xeElement,     //  0 normal element <name {attr}> [value] [sub-elements] </name>

    xeAttribute,   //  1 attribute ( name='value' or name="value") : <element name='value'>....</element>

    xeCharData,    //  2 character data in a node : <element>char data</element>

    xeComment,     //  3 comment <!--{comment}-->

    xeCData,       //  4 literal data <![CDATA[{data}]]>

    xeCondSection, //  5 conditional section <![ IGNORE / INCLUDE [ markup ]]>

    xeDeclaration, //  6 xml declaration <?xml{declaration}?>

    xeStylesheet,  //  7 stylesheet <?xml-stylesheet{stylesheet}?>

    xeDocType,     //  8 doctype dtd declaration <!DOCTYPE{spec}>

    xeDtdElement,  //  9 dtd element <!ELEMENT >

    xeDtdAttList,  // 10 dtd attlist <!ATTLIST >

    xeDtdEntity,   // 11 dtd entity <!ENTITY >

    xeDtdNotation, // 12 dtd notation <!NOTATION >

    xeInstruction, // 13 processing instruction <?...?>

    xeWhiteSpace,  // 14 chardata with only whitespace

    xeQuotedText,  // 15 quoted text: "bla" or 'bla'

    xeEndTag,      // 16 </...> and signal function in binary xml

    xeError        // 17 some error or unknown
  );
  TsdElementTypes = set of TsdElementType;

 

  // End-Of-Line style
  TsdEolStyle = (
    esLF,        // write End-Of-Line as just LF (#$0A) like normalised xml
    esCRLF,      // write End-Of-Line as CR + LF (#$0D + #$0A), for the windows platform
    esCR         // write End-Of-Line as CR (#$0D), for the mac platform
  );

  // Note on TNativeXml.XmlFormat:
  // - xfCompact (default) to save the xml fully compliant and at smallest size
  // - xfReadable writes additional nonsignificant whitespace so the client can easily read the xml file with a standard editor.
  // - xfPreserve aims to preserve whitespace data just as it is parsed
  TXmlFormatType = (
    xfCompact,  // Save without any control chars except LF after declaration
    xfReadable, // Save in readable format with indents and end-of-lines
    xfPreserve  // Preserve whitespace whenever possible
  );

 

{
  NativeXmlNodes:
  Parsing and writing methods for XML nodes:
  Attribute, CDATA, CharData, Comment, Declaration, DocType, DTD-AttList, DTD-Element, DTD-Entity, DTD-Notation, Element, (processing) Instruction, QuotedText, StyleSheet and ancestors:

  TXmlNode
    TsdContainerNode
      TsdElement
        TsdDtdElement
          TsdDtdAttList
          TsdDtdEntity
          TsdDtdNotation
      TsdDocType
      TsdDeclaration
        TsdStyleSheet
    TsdCharData
      TsdQuotedText
      TsdProcessingInstruction
      TsdComment
        TsdCData
        TsdConditionalSection
    TsdAttribute
}