Red Hat Bugzilla – Bug 14679
rpm-4.0-0.59 dumps core when trying to uninstall Xaw3d
Last modified: 2008-05-01 11:37:57 EDT
For quite a while now (since I upgraded to the new rpm format actually),
stuck with Xaw3d in by rpm database, and there is no way to take it off...
a couple of cvs snapshots (for about two weeks I'd say), and the release
keeps growing without solving my problem.
If I run gdn on an unstripped version of rpm, I get:
[accot@obane rpm-4.0]# gdb ./rpm
GNU gdb 5.0
Copyright 2000 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you
welcome to change it and/or distribute copies of it under certain
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show warranty" for details.
This GDB was configured as "i386-redhat-linux"...
(gdb) r -e Xaw3d --nodeps
Starting program: /export/home/src/redhat/BUILD/rpm-4.0/./rpm -e Xaw3d
Program received signal SIGSEGV, Segmentation fault.
0x811b20f in memcpy (dstpp=0x81b0b48, srcpp=0x0, len=2) at
61 ../sysdeps/generic/memcpy.c: Aucun fichier ou ripertoire de ce
#0 0x811b20f in memcpy (dstpp=0x81b0b48, srcpp=0x0, len=2) at
#1 0x806fa9c in rpmRunTransactions ()
#2 0x8067ca6 in rpmErase ()
#3 0x804b018 in main ()
#4 0x8101267 in __libc_start_main (main=0x80499a0 <main>, argc=4,
ubp_av=0xbffff5e4, init=0x80480b4 <_init>, fini=0x815aa80 <_fini>,
stack_end=0xbffff5dc) at ../sysdeps/generic/libc-start.c:111
If I add a --justdb or --noscripts flag, things are similar. A database
rebuild does not solve
the problem either, as well as an upgrade to a new version of Xaw3d.
Pretty frustrating. I
did not see a similar behavior for any other package; only Xaw3d-1.5-2.
rawhide-release-20000720, by the way.
Hope the gdb output helps. Please let me know if I can do further tests.
This is a problem with the Xaw3d header in your database. Send me
(firstname.lastname@example.org) a copy
of your database
tar czvf /tmp/rpmdb.tar.gz rpm
so I can take a look. I'll send you back a clean copy of the database.
this isn't directly a problem in rpm, I'm closing this bug.