Comments on: Generative XPath is released http://uucode.com/blog/2007/06/10/generative-xpath-is-released/ advocating olpa's open source developments Fri, 29 Aug 2008 19:09:47 +0000 http://wordpress.org/?v=2.2.3 By: olpa http://uucode.com/blog/2007/06/10/generative-xpath-is-released/#comment-8282 olpa Sun, 10 Jun 2007 16:43:50 +0000 http://uucode.com/blog/2007/06/10/generative-xpath-is-released/#comment-8282 At the moment, not so good. The style of the compiled code is "good for optimizer", but the optimizator isn't written (yet). The worst, unhonest case. Generative XPath on top of a few layers vs native libxml2 XPath: 30 times slower. Bad case: "//a[b]" vs manually written (<a href="http://community.livejournal.com/ru_scheme/19262.html" rel="nofollow">a variant of this</a>): 20 times slower. Normal: vs SXPath: 2 times slower. But I expect that after even trivial optimizations, I'll beat libxml2. > http://xsieve.sourceforge.net/gxpath/ is dead Under construction. At the moment, not so good. The style of the compiled code is “good for optimizer”, but the optimizator isn’t written (yet).

The worst, unhonest case. Generative XPath on top of a few layers vs native libxml2 XPath: 30 times slower.

Bad case: “//a[b]” vs manually written (a variant of this): 20 times slower.

Normal: vs SXPath: 2 times slower.

But I expect that after even trivial optimizations, I’ll beat libxml2.

> http://xsieve.sourceforge.net/gxpath/ is dead
Under construction.

]]> By: Bacek http://uucode.com/blog/2007/06/10/generative-xpath-is-released/#comment-8281 Bacek Sun, 10 Jun 2007 15:56:22 +0000 http://uucode.com/blog/2007/06/10/generative-xpath-is-released/#comment-8281 (Hey, http://xsieve.sourceforge.net/gxpath/ is dead.) Do you have any benchmarks for your XPath implementation? (Hey, http://xsieve.sourceforge.net/gxpath/ is dead.)

Do you have any benchmarks for your XPath implementation?

]]>