Hey Fabien-
I must have missed your message when you posted to the list. I actually just
happened to find it a few days ago while looking through the archives.
Thanks for the tips. Unfortunately I still experienced problems after
implementing your suggestions. I was not getting the NPEs mentioned in the bug
for sun.font.GlyphLayout, so I didn't attempt to fix that. I'm running IBM's
Java 1.4.2 on AIX, so perhaps that makes the difference. (I did, however,
experience NPE problems with sun.awt.font.AdvanceCache.get. I was able to work
around those.)
Even after making org.apache.fop.render.pcl.PCLRenderer.renderSpace a
synchronized method I am still getting incorrect spacing at times and the wrong
typeface at other times.
Thanks for the investigation though!
Dave
>>>>>>>>>>>>>>>>>>>>>>>>
Hello David,
I have faced the same problem and made multiple investigations. Here are the 3
steps that seem to fix the multithreading issues with FOP / PCL :
(1) Fix a bug in sun.font.GlyphLayout, see
bugs.sun.com/bugdatabase/view_bug.do?bug_id=6367148
bugs.sun.com/bugdatabase/view_bug.do?bug_id=6367148 (using the endorsed
mechanism).
(2) The first FOP warnings appear in the fop layout manager. To avoid this, I
use the fop intermediate format (see
http://xmlgraphics.apache.org/fop/0.94/intermediate.html
http://xmlgraphics.apache.org/fop/0.94/intermediate.html ). So the generation
sequence is : XSLFO => FOP AREA TREE => FOP PCL.
(3) Then some texts have a wrong place on the generated page. Setting a lock
mechanism in org.apache.fop.render.pcl.PCLRenderer class on the renderSpace()
method have fixed the problem. I don't know exactly why, but it seems that some
resources used by this method have thread-unsafe operations.
Pfieeewww =^D ! Perhaps someone will have the courage to make further
investigations. No matter what, FOP is really a great tool =) !
Fabien
David Gerdt wrote:
According to the website documentation on multithreading:
"There is also a known issue with fonts being jumbled between threads when
using the Java2D/AWT renderer (which is used by the -awt and -print output
options). In general, you cannot safely run multiple threads through the AWT
renderer."
Three questions: 1. Am I correct in my cursory observation that this is also
true for the PCL renderer?
2. If it is the case that the PCL renderer also suffers, can someone elaborate
on the cause of these problems? A high level, very basic understanding is all
I'm curious for. I'm definitely inexperienced with thread related programming,
but from what I gather, the AWT java classes are thread safe, correct? Am I
making an illogical jump from reading "AWT renderer" in the FOP docs and
assuming that the problem has something to do with the AWT classes in Java?
3. Any general pointers on how to use multithreading profitably with the PCL
renderer? Right now, the app I wrote (standalone socket server) is
multithreaded but synchronizes access to the Transformer class which seems to
solve my problem from #1 but basically makes it a single thread program again.
Thanks for any insight.
-- View this message in context:
http://www.nabble.com/Multithreading-and-PCL-tp17658051p18870089.html Sent from
the FOP - Users mailing list archive at Nabble.com.
--------------------------------------------------------------------- To
unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL
PROTECTED]
