[Mono-docs-list] Monodoc Assembler Performance patch

Ben Maurer bmaurer@users.sourceforge.net
19 Jun 2003 22:14:05 -0400


On Thu, 2003-06-19 at 22:07, Joshua Tauberer wrote:
> Ben Maurer wrote:
> > ripped out all the XPath and XmlDocument stuff from the guts of
> > ecma-provider and replaced it with cool, clean XmlTextReader.
> 
> (I prefer XPath over switch statements)
So do I! But when parsing such large amounts of data, some times that is
not a choice. It ended up that we had to parse the XPath strings each
time they were used. To give you an idea of what this entailed, just
parsing the XPath was taking over 40 MB of memory (!)

As the documentation grows, the cost of parsing will grow as well. The
cost of doing an XmlReader implementation is (basically) O(n). The same
is not true for XPath/XmlDocument