We have been using
Xalan for 2+ years, and have been caching the Transformer object and sharing it
among users. In reading the Trax specs it appears that we should be caching the
Templates object, and getting a new Transformer for each thread instance. When
we tried this it used twice as much memory to do the transformation, so this is
not a good option for us.
We have never had a
thread problem to date in over two years, but I am wondering if this is really
an unsafe practice with Xalan, or if we can just leave it as-is and continue to
share the Transformer.
Thanks,
Cory
