On Wednesday 26 November 2003 13:46, Stephan Kulow wrote: > Attached is test case and patch. I check explicitly for ID_P, > even though I'm quite sure that the same happens with e.g. > ID_H1. Well, then the patch can only be wrong.