Ole Tange wrote: > > Is there any other thing that can influence on the RAM usage? E.g. version > of glibc, version of kernel, version of filesystem?
To lay some foundation for the environment in which you are seeing this problem the following information would be useful. Can you determine the version of your libc? And other useful information. (I know you said you compiled the latest 4.5.9. Very good!) Debian: dpkg -l libc6 Redhat: rpm -q glibc uname -a cp --version | grep ^cp I can't see how but the type of the filesystem may have an effect there. Might as well check. What does 'df -T' say about the filesystem that you are copying from and to. Just the filetype is needed from that information. df -T . Since you are seeing this problem but Jim is not I suspect it is a difference based upon the differences in the way ./configure builds your binaries. An #ifdef there that takes the code into a different path. But perhaps not. In which case the failure is a data dependent. Do you see anything particularly interesting about the input tree? Filenames really long? Really short? Lots of them in one directory? Files already hardlinked many times? Something unusual? Thanks Bob _______________________________________________ Bug-fileutils mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-fileutils