Re: [xml] libxml2 review in windows::developer

On Fri, Apr 11, 2003 at 12:49:27PM +0200, Peter Jacobi wrote:
Hi Daniel,

I wanted only to report this as a good news/bad news factoid:

  Yeah, I wasn't negative about the posting more about the fact that
they apparently didn't tried to use libxml directly, i.e. they didn't
bothered looking at the native APIs

good news: libmxml is considered mainstream, even on Win32
bad news: rather weak comparative benchmark doesn't give real information.


  Not much to me because I don't know what they benchmarked.

Reading the file into a tree, walking the tree, thawing objects:
void read_file(char const *input_name, FILE *output)
  using xmlpp::DomParser;
  using xmlpp::Node;
  using xmlpp::Attribute;

  DomParser       parser(input_name);
  Node            *root     = parser.get_root_node();
  Node::NodeList  children  = root->children("entry");

  you see, for me it's very hard to map that to the associated calls
in libxml2, I can't even tell where the document is actually parsed,
nor what root->children("entry"); actually does, is taht a call to XPath ?
or a direct walk of the children nodes, or ...

  Node::NodeList::const_iterator  begin = children.begin();
  Node::NodeList::const_iterator  end   = children.end();

  for(int i = 0; begin != end; ++begin, ++i)
    Node const      *node   = *begin;
    Attribute const *year   = node->attribute("year");
    Attribute const *month  = node->attribute("month");
    Attribute const *day    = node->attribute("day");
    Attribute const *who    = node->attribute("who");
    Entry           entry(year->value(),

  I can think of ways to significantly speed up attribute value 
lookup in the general case too..
  ATM I'm working on features, maybe I will have time to do some
speedup work before the summer..

Daniel Veillard      | Red Hat Network
veillard redhat com  | libxml GNOME XML XSLT toolkit | Rpmfind RPM search engine

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]