Description of Problem: I am having a similar problem compared to http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=60042 After doing all the updates plus probably twiddling things here locally, I now have pine crashing after doing any directory listings or trying to attach a file. The strace will be included in next ticket. rt_sigprocmask(SIG_BLOCK, [ALRM], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], [ALRM], 8) = 0 write(1, "\33[4;1HCc : \33[5;1H\33[7mAttchm"..., 69) = 69 write(1, "\33[5;11H", 7) = 7 select(1, [0], NULL, [0], {150, 0}) = 1 (in [0], left {147, 626913}) read(0, "\24", 1) = 1 time(NULL) = 1035485281 write(1, "\33[22;1H\33[K", 10) = 10 write(1, "\33[22;20H\33[7m[ Building file list"..., 69) = 69 stat64("/home/smoogen", {st_mode=S_IFDIR|0700, st_size=4096, ...}) = 0 open("/home/smoogen", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY) = 5 fstat64(5, {st_mode=S_IFDIR|0700, st_size=4096, ...}) = 0 fcntl64(5, F_SETFD, FD_CLOEXEC) = 0 getdents64(0x5, 0x833ea40, 0x1000, 0) = 3272 getdents64(0x5, 0x833ea40, 0x1000, 0) = 0 close(5) = 0 writev(2, [{"pine", 4}, {": ", 2}, {"relocation error", 16}, {": ", 2}, {"pine", 4}, {": ", 2}, {"undefined symbol: lstat", 23}, {"", 0}, {"", 0}, {"\n", 1}], 10) = 54 _exit(127) = ? Version-Release number of selected component (if applicable): smoogen:{smoogen}$ rpm -q pine glibc pine-4.44-7 glibc-2.2.5-40 things that are NOT standard kerberos -> upgraded to 1.2.6 for acluster reasons krbafs -> backported 8.0 cups -> backported 8.0 Omni -> backported 8.0 ghostscript -> "" openssh -> "" pam_krb5 -> "" webalizer -> ""
Created attachment 81966 [details] Complete strace -o fff1 pine sjs
You've indicated you're using software components that are not supplied by Red Hat, but which do link into pine. That is not a supported configuration for any reason. I wont consider this a legitimate bug report unless it can be reproduced on a clean installation of Red Hat Linux 7.3 with all Red Hat released erratum applied, and no externally upgraded libararies or other software which can affect the behaviour of the application. Closing bug NOTABUG.
You are absolutely correct. What caused this was a 8.0 kerberos recompiled back to 7.3. If I had been a bit more aware when I did it, I would have looked over the statglue patch that Nalin disabled during the changeover to 8.0. I am guessing that the statglue patch was added in 7.0? to fix problems between glibc, stat64, and the kerberos libraries version. Looking through the kerberos changelog glibc was melded to workaround this, but because of issues in kerberos it needed to be maintained throughout the 7.x series. I missed that changelog entry when I backcompiled the 8.0 version. So I am thinking that anyone who opens bugs with a missing stat has probably taken kerberos out of rawhide or something similar. My apologies for not catching my own foolishness.