On Fri, 8 Feb 2002, David Crossley wrote: > Christian Schmitt wrote: > > Vadim Gritsenko wrote: > > > > Christian Schmitt wrote: > > > > > > > > Hi all, > > > > after further checking into our "challenge" here, > > > > that's what we've found out: > > > > > > > > The changes to sitemap.xsl checked in recently checking the > > > > pipeline rules, is what brings sitemap generation to a crawl. > > > > > > That's exactly what I was thinking after this email :) > > > > Vadim, > > what seems to be the real killer is this construct: > > > > <xsl:for-each select=".//*[namespace-uri()=$ns > > and (local-name()='generate' or local-name()='aggregate' or >local-name()='transform') > > and not(ancestor::map:handle-errors)]"> > > <xsl:if test="not(following::*[namespace-uri()=$ns > > and (local-name()='serialize' or local-name()='call') > > and not(ancestor::map:handle-errors)]/ancestor::map:pipeline[. = >$pipeline])"> > > <xsl:call-template name="error"> > > <xsl:with-param name="message">Sitemap pipeline with >map:generate|aggregate|transform must have map:serialize.</xsl:with-param> > > </xsl:call-template> > > </xsl:if> > > </xsl:for-each> > > > > The <xsl:if test ......> </xsl:if> takes about one second for each > > element found. > > > > Now, not being the XSL guru, I have no idea how to make this faster > > (besides commenting it out ;-) ). > > > > Cheers, > > Christian > > On my machine the initial Cocoon startup never gets past the > first generation of the sitemap. It has been sitting for 12 hours > on the log message "Beginning sitemap regeneration". > According to "top" the "java" process is at 90% CPU and at > 20% Memory. So it is not only slow, it is broken.
I've used a recent cvs checkout without any problems. Giacomo --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
