PDF "Double Dipping" -- FOP bug?

[Paul Lee]

I know the subject line is pretty stupid, but hell, it’s been a long day…   I have started to use Cocoon to generate HTML and PDF output from a database.  It works great but I noticed that whenever I generate a PDF file the same query always get executed twice, while the equivalent HTML generation uses only one request.  This is especially noticeable as one of my queries take a minute long.  Cocoon also seem to reprocess the XML whenever I ask for the PDF transformation – this is evident from the “double spike” in the processing power usage (I am running Windows 2000 and you can see the two “hills” prominently in the Task Manager)   I am wondering if I am doing anything wrong, but I noticed that it seems to be doing the same thing even to the hello.pdf example on a fresh, plain installation of Tomcat/Cocoon.   This leads me to wonder: is this a bug in FOP?  Has anyone seen/solve this?   Thanks in advance!   Paul   p.s. Tomcat 4.0.4, Cocoon 2.0.3, Windows 2000 & Windows XP, JDK1.4 & 1.3.1