svn java bindings fail to build in UTF-8 filesystems

Problem:I'm using svn under linux-ppc, and it fails to build java
bindings due to a couple of files having e-acute encoded in ISO-8859-1
in them. The compiler fails with a sun.io.MalformedInputException.

Solution:
- using é in javadocs (they will turn into HTML after all.
- using "e" instead
- using the (recommended) \uNNNN convention for non-ascii chars in files

In both cases, the purpose is to avoid dependency of the source files on
file encoding, given that there is no way this information can propagate
to a tarball, and thus the information is lossy. An alternative would be
to force a given source encoding in the build scripts, but it looks like
a kludge to me.

Regards

-- 
VP and Chair, Apache Portals (http://portals.apache.org)
Apache Software Foundation