Vincent Hennebert writes:
 > I fully agree. Good design should not be sacrificed for efficiency.
 > Anyway only the profiling of a FOP run would give us proper indications
 > of what is actually eating time and memory, and where we should start
 > optimizing.


If you apply the the recent patches here:
   http://issues.apache.org/bugzilla/show_bug.cgi?id=41656
and here:
   http://issues.apache.org/bugzilla/show_bug.cgi?id=41044
then process a big silly 40MB plus fop table and wait for it to
bomb out with an OOM you wind up with a set of instance counts
as shown:

134407 instances of class org.apache.fop.fo.FONode[]
106438 instances of class char[]
100833 instances of class org.apache.fop.fo.FOText
100805 instances of class org.apache.fop.fo.flow.Block
100803 instances of class org.apache.fop.fo.flow.TableCell
33601 instances of class org.apache.fop.fo.flow.TableRow
5618 instances of class java.lang.String
5047 instances of class java.lang.Integer
4316 instances of class java.util.HashMap$Entry
2620 instances of class org.apache.fop.fo.FObj$FObjIterator
1475 instances of class java.lang.Class

Richard

Reply via email to