I upgraded glibc this night (via autorpm) to the latest from the errata - glibc-2.0.7-29.2. Then tcsh (tcsh-6.07.09-1) didn't work anymore, it core-dumped. We are running RedHat 5.2, Linux 2.0.36 on a i686. I am able to reproduce the problem on several other RedHat 5.2-servers (with different configurations/software installed), so I suppose the problem is in glibc and not in local configuration.
Sorry, forgot the gdb-stuff: [magnio@nalle]:~> gdb tcsh core (no debugging symbols found)... Core was generated by `tcsh'. Program terminated with signal 11, Segmentation fault. Reading symbols from /lib/libnsl.so.1...done. Reading symbols from /lib/libtermcap.so.2...done. Reading symbols from /lib/libcrypt.so.1...done. Reading symbols from /lib/libc.so.6...done. Reading symbols from /lib/ld-linux.so.2...done. #0 0x40063434 in __open_catalog (catalog=0x8098008, with_path=1) at open_catalog.c:91 open_catalog.c:91: No such file or directory. (gdb)
I also saw this bug on my Redhat 5.2 installation. I thought that perhaps something was wrong with Redhat's (old) tcsh RPM, so I went and compiled the latest tcsh (6.09.00) from source, and sure enough, the bug remains. But because I compiled with debugging, I have a more detailed backtrace: #0 0x40094434 in __open_catalog (catalog=0x8093008, with_path=1) at open_catalog.c:91 #1 0x40094157 in catgets (catalog_desc=0x8093008, set=2, message=1, string=0x807b49b "Syntax Error") at catgets.c:137 #2 0x804ffa4 in errinit () at sh.err.c:215 #3 0x80567a5 in nlsinit () at sh.func.c:2396 #4 0x8049da1 in main (argc=1, argv=0xbffff668) at sh.c:206 So, something is wrong in the new library's catgets(), in line 137. I hope that gives you enough info to catch and fix this bug (hopefully quickly - some of my users want to kill me for distroying their working tcsh for what they see as not-important security concerns).
*** Bug 17201 has been marked as a duplicate of this bug. ***
For anyone, it should be fixed in glibc 2.0.7-29.3, this weekend the fixed rpms are available from ftp://ultra.linux.cz/private/glibc/ and updated errata will be issued hopefully Monday morning EDT.
the .3 rpm does appear to fix it. the upgrade went fine on one of machines. The other I had tried to downgrade to the 5.2 release glibc, and rpm failed to extract libc-2.0.7.so using cpio properly... but after figuring out how to get the 5.2 rescue disk series to load a scsi driver all is well. Thanks for the quick turnaround. I've got to say this was the nastiest issue with a RedHat system involving official RedHat stuff (it took down a lot more than tcsh).
*** Bug 17233 has been marked as a duplicate of this bug. ***
*** Bug 17281 has been marked as a duplicate of this bug. ***
This nails Perl. In particular the glob() function and <*> directory file handles. This is entirely ass!
Should be fixed by RHSA-2000:057-04
*** Bug 17192 has been marked as a duplicate of this bug. ***
*** Bug 17249 has been marked as a duplicate of this bug. ***