Thread: XML Parsing

I'm trying to pass into my C++ linux application an XML document, which is describing some algorithm, that should be executed by application. I need to sequentially parse this document and after reading some value from it, immediately execute appropriate action described by that value.

How should I parse that XML document and which XML parser should I choose? Actually, after reading some tutorials, I had already chosen IBM's xml4c2, but is it good decision?

What exactly does it mean "SAX" and what is it good for?

I don't know really. People have recommended tinyXML to me but I could never understand it properly

Why did they recommend you just TinyXML? Is there some good reason for it?

I don't really know it just seems to be popular. Probably because it is small and functional but there are other libraries out there to try.

http://www.grinninglizard.com/tinyxml/

sax is a method for parsing XML (Simple API for XML), it means basically that you parse the document using callbacks, such as onChar() for character data, onBeginElement and onEndElement. Using SAX you can process the document while parsing it and make it recover from XML errors if necessary. The other popular method is using a DOM parse (Document Object Model) where you give the parser an XML document and it give you back sort of a tree structure. The DOM type has larger memory requirements because it creates a data structure out of all the data. Typically when doing web services and such, you would use SAX, but when using an XML document as a file data structure or something like that, you would be more likely to use DOM.

In the past, I've used both expat and xerces, both open source.

So is there anybody out there, who have some experience with XML parsers and is willing to share it with me?

I would recommend using expat, it's small, fast and efficient. Xerces I've only used in java, I don't know if they have a C version of it, but you won't go wrong with expat, I've used it on the server side sometimes handling 400 connections per minute, each with XML to parse. They will all have the same interface if they support SAX, so they should be pretty much equally simple to use.