Nicola Ken Barozzi wrote: >Cocoon in docs build is slow. :-( >On the same documentation set >(PIII 1Ghz, 512 Mb, W$2000SP1, Jdk Sun 1.3.1_01): > >Anakia: 4 sec >Cocoon of two weeks ago: 50 sec >Cocoon of two weeks ago precompiled sitemap: 16 sec >Cocoon HEAD: 45 sec >Cocoon HEAD with cocoon.xconf tweaks: 38sec >Cocoon HEAD precompiled sitemap: 18 sec >Cocoon HEAD+TreeProcessor: 2 sec (only index.html compiled) > Could you elaborate on the last line : do you mean Cocoon with TreeProcessor is the fastest, or that is fails after the first page ? Depending on this, I will be happy or sad ;)
>I'm struggling to keep other developers on the project use Cocoon, but it's >dead hard, and I can't blame them. >It seems that link crawling and setup times are the main culprit. >In what direction should I go to get faster performance? > For setup time, the main point is pool size : set all Poolable components with a pool-min="1" in cocoon.xconf >Also, do you think that docs generation can be a good performance >measurement of Cocoon speed? > No, because Cocoon has an important warm up time that pays on the long run, but is dramatic for one-time command line processing. >I would like to make the docs generation a Cocoon speed index: SpeCoon, >using my machine as a reference. ;-) > On this subject, I proposed today to build a JMeter (an Apache project ;) test suite that could be used to compare Cocoon versions. I don't know JMeter, so I'm looking for volunteers... >It would be nice if Gump ran the docs generation nightly (does it?) and >logged how much it took. >We could see daily how performance was better or worse. > >Comments? > See inline ! Sylvain -- Sylvain Wallez Anyware Technologies - http://www.anyware-tech.com --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
