Hello out there, a while ago Walter Obermiller reported a bug on this mailing list concerning duplicated parts in generated indexes. I answered: > Walter told that under certain, not reproducible circumstances, parts > of a generated index seem to be duplicated at the beginning and/or > end. I'm very interested if somebody had similar problems (probably > under the Linux version given above?). Now I have found this stupid bug (RTFM to myself). The problem is that the generated files are not truncated to length zero in all cases. Thus, it can happen that the new stream of data is shorter than the old one resulting in some trailing garbage at the end of the file. Since I'm currently very busy I do not package a completely new version but putting a file on the ftp-Server that describes how to patch the sources. Cheers, Roger -- ====================================================================== Roger Kehr kehr@iti.informatik.th-darmstadt.de Computer Science Department Technical University of Darmstadt