[Xmlbench-common] Wondering about libxml2-2.6 vs. 2.4 results

[Daniel V. <vei...@re...>]

  I'm puzzled about the report of it being slower since I spent
quite a bit of time making optimization which were significant
between 2.5.x and 2.6.0.
  I started to look at the README and Makefile.in, seems you use
includes based on /usr/include and similar for libraries, so I'm
wondering if you used prepackaged versions for 2.6.5 and a manually
compiled version for 2.4 . I would like to reproduce your result,
what I can see from the driver files are:
   - you use xmlSAXUserParseMemory version 1 parser instead of the
     version 2 introduced with 2.6.0
   - you use xmlParseMemory and hence also version 1 parser instead
     of xmlReadMemory introduced with 2.6.0
I'm wondering if releaseXML() is called for each parse or only once
the test is finished, it's a library shutdown, not a parser shutdown
so it should really be called once in case of repeated parsing.

 I'm only starting to look at the code,

Daniel

-- 
Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
vei...@re...  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/