Skip to Content.
Sympa Menu

xom-interest - Re: [XOM-interest] Attributes performance patch

xom-interest AT lists.ibiblio.org

Subject: XOM API for Processing XML with Java

List archive

Chronological Thread  
  • From: Wolfgang Hoschek <whoschek AT lbl.gov>
  • To: Elliotte Harold <elharo AT metalab.unc.edu>
  • Cc: xom-interest <xom-interest AT lists.ibiblio.org>
  • Subject: Re: [XOM-interest] Attributes performance patch
  • Date: Mon, 22 Nov 2004 13:06:51 -0800

There are two sets of patches, one completely turns of the checks, the other set faithfully keeps the checks, but makes them faster (better algorith, same behaviour). I guess you will want the second set. How about checking things out first before rejecting proposals?
Thanks,
Wolfgang.

On Nov 22, 2004, at 1:02 PM, Elliotte Harold wrote:

Wolfgang Hoschek wrote:

The other thing is that some days ago I submitted performance patches for URI and PCDATA checks in Verifier. They slighly restructure the algorithms, reducing memory allocation, some unnecessary calculations, and making the methods shorter and easier to inline.
These are actually much more important than the Attributes patch. Safe some 30% is some cases. Did you have a look at those? You can get them at http://dsd.lbl.gov/~hoschek/xom-patches

I'm not going to turn off the checks. That's just not going to happen. Correctness is more important than speed. You also seem to be assuming that if the Text is stored as a String rather than a byte array, it's impossible to have bad surrogate pairs. Sadly, that isn't true. Java Strings can be malformed UTF-16. :-(

I do plan to try some optimizations for the URI checking that may reduce a lot of redundant checks. That may solve your URI issues without eliminating the verifications.

--
Elliotte Rusty Harold elharo AT metalab.unc.edu
XML in a Nutshell 3rd Edition Just Published!
http://www.cafeconleche.org/books/xian3/
http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim





Archive powered by MHonArc 2.6.24.

Top of Page