anaconda[318]: segfault at 0000000000000060 rip 0000002a95e03052 rsp 0000007fbfffd8b0 error 4 Or, on another box: anaconda[332]: segfault at 00000000003f9065 rip 0000002a95e02fec rsp 0000007fbfffd8d0 error 4 According to tty 3: moving (I1) to step installpackages setting file_context_path to /etc/selinux/targeted/contexts/files/file_contexts
Created attachment 106050 [details] rpm diffs between -13 and -20 Here's the rpm changes between -13 and -20, just for reference. poptI.c and psm.c look obviously safe. Which leaves rpmts-py.c (#137116) and lib/formats.c (#135389)
Note - the rpmIncreaseVerbosity needs to go away.
Erm, bug 135859 is the lib/formats.c bug fixed, just for reference.
Happens on i386 too.
And, your backtrace is.... Program received signal SIGSEGV, Segmentation fault 0xb7eb7dec in pthread_mutex_lock () from /lib/libpthread.so.0 (gdb) bt #0 0xb7eb7dec in pthread_mutex_lock () from /lib/libpthread.so.0 #1 0xb7eba94e in flockfile () from /lib/libpthread.so.0 #2 0xb7dc2d26 in fputs () from /lib/libc.so.6 #3 0xb795a523 in rpmlogSetFile () from /usr/lib/librpmio-4.3.so ...
... annnd, in fact, removing the rpmIncreaseVerbosity calls gets rid of the segfault. Will add a patch to the FC3 branch in package cvs; Jeff, you probably want to hit CVS with the fix.
Created attachment 106056 [details] the patch
4.3.2-21 built.
iAdded to CVS. Thanks for the patch, apologies for the bother.