On Thursday 11 May 2006 19:56, Joachim Goeschel wrote:
I am still pretty new to Linux and might not figure how to handle symbolic links correctly.
Basic format ln -s /path/object /path/link 'ln' = link '-s' specifies symbolic (not a 'hard' link) The problem is matlab expects to find libc.so.6 but it isn't there... the newer (frequently backwards-compatible) library has a newer version number than matlab knows about. Listing from my 10.0 /lib: -rwxr-xr-x 1 root root 1405291 2005-09-09 13:36 libc-2.3.5.so* lrwxrwxrwx 1 root root 11 2006-03-12 20:29 libcap.so -> libcap.so.1* lrwxrwxrwx 1 root root 14 2006-03-12 20:29 libcap.so.1 -> libcap.so.1.92* -rwxr-xr-x 1 root root 34670 2005-09-09 11:59 libcap.so.1.92* -rwxr-xr-x 1 root root 191505 2005-09-09 13:36 libcidn-2.3.5.so* lrwxrwxrwx 1 root root 16 2006-03-12 20:28 libcidn.so.1 -> libcidn-2.3.5.so* lrwxrwxrwx 1 root root 17 2006-03-12 20:31 libcom_err.so.2 -> libcom_err.so.2.1* -rwxr-xr-x 1 root root 6536 2005-09-09 11:58 libcom_err.so.2.1* -rwxr-xr-x 1 root root 47724 2005-09-09 13:36 libcrypt-2.3.5.so* lrwxrwxrwx 1 root root 17 2006-03-12 20:28 libcrypt.so.1 -> libcrypt-2.3.5.so* lrwxrwxrwx 1 root root 13 2006-03-12 20:28 libc.so.6 -> libc-2.3.5.so* See all the links carrying older library names pointing to later versions of the same libraries? That's what Carlos and I are suggesting as a reasonable test... or, at least I am. ;-)
Furthermore, if I remove the lines
LD_ASSUME_KERNEL=2.4.21 export LD_ASSUME_KERNEL
If you haven't seen any documentation recommending this change, why would you implement it? regards, Carl