Hi, I don't think it's possible to control the line endings produced from a transform. The usual answer is that it doesn't matter what line endings are produced, as when they are read in by any conformant xml parser they will be normalized to &_#xa; (underscore for converting mailers).
So I guess the answer is to look at what is reading the xml back in. You say in your post that you are generating java source code from the transform - that sounds pretty cool (and pretty difficult!) can you post a sample? cheers andrew === Hi, I'm generating Java source code from mathml using xslt. I also output in Unicode format for including Greek symbols (often in the original mathml and translates nicely). But I have the problem that the LF CR output does not get accepted by the jdk1.4 compiler(javac). I believe javac expects Unicode with only LF. Does anyone know how to crontrol LF CR output? --Roger ===
