Using PI's will cause the XML to be have to read twice currently (unless your using CVS axkit).
Once to snoop the XML (so that you can extract the PI's and build the pipeline), the second
to process it. Also the extract_styles stuff isn't going to be the quickest process (though
the results are cached).
It's just some processing you can avoid.
Mind you, I used to disable that cache because I found it's caching to be annoying in a development
environment (you can't tell when a .htaccess file has been updated, and so the cache doesn't expire
correctly.
Also on a large site PI's are annoying as they override the httpd.conf rules, we used to ultimately
build one httpd.conf with all the rules in, and disable .htaccess files on the live server as well.
I'm not saying it'll be a problem, just when your trying to squeeze every last ounce out some boxes
you look at and consider everything.
On 16 Jun 2004, at 05:59, S. Woodside wrote:
are you sure about that? on linux x86 I use PIs for everything and have no performance problems.
simon
On Jun 15, 2004, at 7:06 AM, Mike Chamberlain wrote:
Don't use PI's, or any of the AxAdd directives which require to snoop the xml. (ie look at the first node for node name, namespace or dtd).
-- http://simonwoodside.com
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
