On Saturday 27 April 2002 04:45 pm, Stefano Mazzocchi wrote: > Now, the interesting thing is that it seems that Xalan is running it's > own thread. I don't recall seeing thread-forking code for Xalan spin-off > in Cocoon, so I wonder: is this a Xalan thing?
yes. > But the *really* weird thing is that I also get this: > > java.lang.RuntimeException: java.net.SocketException: Connection aborted > by peer: socket write error > at > org.apache.xalan.transformer.TransformerImpl.run(TransformerImpl.java:3231) > at java.lang.Thread.run(Thread.java:484) > > on STDERR! Try turning on "incremetal processing", you'll get those wrapped in RuntimeException's instead :) > I suspect this is a Xalan bug, but I'm not really that sure so I post it > on bost lists. I entered a bug for some of this, it appears that it is a somewhat-known problem due to the comments in the xalan source. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=8033 > Functionally, everything is all right (even if, I suspect that some > files are sometimes reloaded from the browser cache!), but these > stacktraces really fill up the log files quickly and this is very > annoying. I agree, it is very annoying to have such stacktraces fill the logs. This is a behavior that started somewhere in the 2.2 series for Xalan. IIRC, 2.0 didn't have the same behavior I'd like to know what the expected behavior from Xalan is so Cocoon can be modified to expect it accordingly. -pete -- peter royal -> [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
