Re: [xml] ctxt->token patches

On Fri, Jun 28, 2002 at 04:49:35PM +0200, Peter Jacobi wrote:
Content-Description: Mail message body
Hi Daniel,

  Hum, can you tellwhere this occured ? 

It seems benefical to test for ctxt->token in front
of a large list of other "else if" conditionals.

I've attached a step-2 patch which removes all remaining 
ctxt->token uses in parser.c and parserInternals.c (except
the initializing to 0). htmlparser.c and docbookparser.c will
be treated in the next patch.

  Okay, applied and commited too !

This second patch eats about 60% of the performance gains
of the first, but I fully agree that's nonsense to leave voodoo
statements in the code.

  Well, there must be a good reason why the compiler
was generating better code when having extra NOP operations
to do. Problem is that it's not associated to the semantic of the
language :-\

To proceed further in performance enhancement I need to
do better profiling and perhaps look for more intrusive
changes in the parser.

  Yup, and again thanks a lot for your work on this !!!


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]