Hi all, I've replaced most of the JDK 1.0 containers by 1.2 containers in the maintenance branch, ready to commit. The PDF produced from FOP examples compares ok with PDF produced before the conversion. I might have screwed up other renderers, in particular the MIF renderer which required quite a bit more than a simple S&R. I might also have produced additional MT problems, even though I set FOPImageFactory to synchronized.
As a side note, I made a script which intercepts GC notifications produced with -verbose:gc and computes total and peak memory. If this can be trusted, using 1.2 containers causes FOP to use 2% less total as well as peak memory. Run time performance is basically the same, apparently Java synchronization doesn't have much effect in a single-threaded environment. If there are no vetoes, I'll commit the change on 2002-08-02, 21:00 CEDST. My plans for the near and not so near future: - vacancy: I'm offline until 2002-08-11 - tracking down the zero width space problem - eleminating the instance data of FONode and as much as possible data of FObj - create the "children" array on demand - resolve doc issues and commit the docs (take deep breath) - implement immutable, reused property bundles and leaner property parsing - integrate image handling and the PDF renderer from head - TR14, spec conformant whitespace handling, linefeed treatment, wrapping and overflow="clip" - rearranging the Area inheritance tree (spaces don't have a height, which is bad), eleminate as much as possible unnecessarily inherited data for inline areas - correct lineheight computation and alignment. Should be enough work... J.Pietschmann --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
