--- Comment #3 from Andreas L. Delmelle <>  2009-04-02 
12:58:21 PST ---

Just one small remark: the 'benchmark' seems to be taken from a single run, via
the command-line (I presume?). Plain fact is that the number of classes has
increased significantly between the two versions, so that makes the influence
of JVM warmup time all the greater. If you run FOP repeatedly from the
command-line, you may want to take the time to set up a small wrapper that
keeps the app sleeping in the background to avoid the VM startup.

Redo the test by running the same file in a Java program, looped to repeat 11
times. Compare times 2 to 10, and that will give you a more realistic picture
of the actual difference.

Configure bugmail:
------- You are receiving this mail because: -------
You are the assignee for the bug.

Reply via email to