I wonder whether this is a mis-installation issue on my Gentoo laptop. I see the library currently in play is libdb-4.8 while the debug points to db3.h. Is there a way to check which major library version is available during runtime? # ldd /usr/sbin/kdc linux-gate.so.1 => (0xb7702000) libkdc.so.2 => /usr/lib/libkdc.so.2 (0xb76b4000) libhdb.so.9 => /usr/lib/libhdb.so.9 (0xb769b000) libkrb5.so.26 => /usr/lib/libkrb5.so.26 (0xb763c000) libpthread.so.0 => /lib/libpthread.so.0 (0xb7623000) libroken.so.18 => /usr/lib/libroken.so.18 (0xb760f000) libc.so.6 => /lib/libc.so.6 (0xb74c9000) libheimntlm.so.0 => /usr/lib/libheimntlm.so.0 (0xb74c3000) libsqlite3.so.0 => /usr/lib/libsqlite3.so.0 (0xb743c000) libicui18n.so.44 => /usr/lib/libicui18n.so.44 (0xb7277000) libicuuc.so.44 => /usr/lib/libicuuc.so.44 (0xb714a000) libwind.so.0 => /usr/lib/libwind.so.0 (0xb7121000) libdl.so.2 => /lib/libdl.so.2 (0xb711d000) libcrypto.so.1.0.0 => /usr/lib/libcrypto.so.1.0.0 (0xb6fc5000) libasn1.so.8 => /usr/lib/libasn1.so.8 (0xb6f42000) libcom_err.so.2 => /lib/libcom_err.so.2 (0xb6f3e000) libcrypt.so.1 => /lib/libcrypt.so.1 (0xb6f0b000) libdb-4.8.so => /usr/lib/libdb-4.8.so (0xb6db1000) libresolv.so.2 => /lib/libresolv.so.2 (0xb6d9c000) /lib/ld-linux.so.2 (0xb7703000) libstdc++.so.6 => /usr/lib/gcc/i686-pc-linux-gnu/4.4.4/libstdc++.so.6 (0xb6caf000) libm.so.6 => /lib/libm.so.6 (0xb6c89000) libgcc_s.so.1 => /usr/lib/gcc/i686-pc-linux-gnu/4.4.4/libgcc_s.so.1 (0xb6c6c000) libicudata.so.44 => /usr/lib/libicudata.so.44 (0xb5e2e000) libz.so.1 => /lib/libz.so.1 (0xb5e19000) # # emerge --unmerge sys-libs/db * This action can remove important packages! In order to be safer, use * `emerge -pv --depclean <atom>` to check for reverse dependencies before * removing packages. sys-libs/db selected: 1.85-r3 3.2.9_p2 4.2.52_p5-r1 4.3.29_p1-r1 4.5.20_p2-r1 4.6.21_p4 4.7.25_p4 4.8.30 protected: none omitted: none >>> 'Selected' packages are slated for removal. >>> 'Protected' and 'omitted' packages will not be removed. >>> Waiting 5 seconds before starting... >>> (Control-C to abort)... >>> Unmerging in: 5 4 ^C Exiting on signal 2 # From configure's output I would like to emphasize: checking db4/db.h usability... no checking db4/db.h presence... no checking for db4/db.h... no checking db3/db.h usability... yes checking db3/db.h presence... yes checking for db3/db.h... yes checking db.h usability... yes checking db.h presence... yes checking for db.h... yes checking db_185.h usability... yes checking db_185.h presence... yes checking for db_185.h... yes checking for db_create... yes, in -ldb checking for dbm_firstkey... yes, in -ldb # emerge -pv heimdal [ebuild R ] app-crypt/heimdal-1.3.3-r1 USE="X berkdb ssl threads -afs -hdb-ldap -ipv6 -otp -pkinit -test" 0 kB
Forgot to say that the crash happens in the backrgound during $ su - Password: $ # ldd /bin/su linux-gate.so.1 => (0xb7895000) libpam.so.0 => /lib/libpam.so.0 (0xb7857000) libpam_misc.so.0 => /lib/libpam_misc.so.0 (0xb7853000) libc.so.6 => /lib/libc.so.6 (0xb770d000) libdl.so.2 => /lib/libdl.so.2 (0xb7709000) /lib/ld-linux.so.2 (0xb7896000) # The kdc is contacted due to PAM support in su. # equery belongs su * Searching for su ... app-crypt/kth-krb-1.3_rc1 (/usr/athena/bin/su) sys-apps/shadow-4.1.4.2-r5 (/etc/pam.d/su) sys-apps/shadow-4.1.4.2-r5 (/bin/su) # BTW, where is su(1) from heimdal distribution?
Created attachment 243483 [details] build.log
Created attachment 243485 [details] config.log
(In reply to comment #1) > Forgot to say that the crash happens in the backrgound during > > $ su - > Password: > $ I hand-typed this. Importantly, I should have written `#' at the end instead of the `$' sign as I really do get root-shell as expected. My aim was to show there is no message returned to the user. The crash itself is logged through syslog in my case but that is not important: kdc[2306]: segfault at 0 ip (null) sp bfdb195c error 4 in kdc[8048000+6000]
Created attachment 243487 [details] kdc-stacktrace.txt
I get same output configure regarding the berkeley db version tests, so I can exclude that the db5 compatibility patch from bug #322709 is at fault.
Looks like linux version of dll hell. depclean/prune sys-libs/db and build heimdal again. You might want to revdep-rebuild for other applications as well. Reopen if problem still persists. Also, please attach full build logs, not just snippets.
# emerge --prune --verbose sys-libs/db Calculating dependencies... done! sys-libs/db-4.5.20_p2-r1 pulled in by: dev-lang/python-2.5.4-r4 sys-libs/db-4.7.25_p4 pulled in by: dev-lang/python-2.6.5-r3 sys-libs/db-4.8.30 pulled in by: app-arch/rpm-4.4.6-r7 app-crypt/heimdal-1.3.3-r1 app-office/openoffice-3.2.1 dev-lang/perl-5.12.1-r1 dev-lang/php-5.3.3-r1 dev-lang/ruby-1.8.7_p299 dev-libs/apr-util-0.9.12 dev-libs/apr-util-1.3.9-r1 dev-libs/cyrus-sasl-2.1.23-r1 dev-libs/redland-1.0.10-r2 dev-python/bsddb3-5.0.0 dev-vcs/subversion-1.6.12 gnome-extra/evolution-data-server-2.30.2 net-analyzer/dsniff-2.4_beta1-r4 net-proxy/squid-3.1.6-r1 perl-core/DB_File-1.820 sys-apps/iproute2-2.6.35-r1 sys-cluster/pvfs2-2.7.1 sys-libs/gdbm-1.8.3-r4 sys-libs/pam-1.1.1-r2 sys-libs/db selected: 4.3.29_p1-r1 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 4.6.21_p4 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 4.2.52_p5-r1 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 3.2.9_p2 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 1.85-r3 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 >>> 'Selected' packages are slated for removal. >>> 'Protected' and 'omitted' packages will not be removed. >>> Waiting 5 seconds before starting... >>> (Control-C to abort)... >>> Unmerging in: 5 4 ^Z [1]+ Stopped emerge --prune --verbose sys-libs/db Before I do this I will re-attach the complete build.log. I thought that the configure output is enough, sorry.
(In reply to comment #7) > Also, please attach full build logs, not just snippets. Complete build.log from the broken install host is at http://195.113.57.32/~mmokrejs/tmp/build.log
If you cannot attach logs (bacause of size limit for example), it is better to compress and attach rather than hosting them elsewhere. Just as a FYI for the future.
# emerge --prune --verbose sys-libs/db Calculating dependencies... done! sys-libs/db-4.5.20_p2-r1 pulled in by: dev-lang/python-2.5.4-r4 sys-libs/db-4.7.25_p4 pulled in by: dev-lang/python-2.6.5-r3 sys-libs/db-4.8.30 pulled in by: app-arch/rpm-4.4.6-r7 app-crypt/heimdal-1.3.3-r1 app-office/openoffice-3.2.1 dev-lang/perl-5.12.1-r1 dev-lang/php-5.3.3-r1 dev-lang/ruby-1.8.7_p299 dev-libs/apr-util-0.9.12 dev-libs/apr-util-1.3.9-r1 dev-libs/cyrus-sasl-2.1.23-r1 dev-libs/redland-1.0.10-r2 dev-python/bsddb3-5.0.0 dev-vcs/subversion-1.6.12 gnome-extra/evolution-data-server-2.30.2 net-analyzer/dsniff-2.4_beta1-r4 net-proxy/squid-3.1.6-r1 perl-core/DB_File-1.820 sys-apps/iproute2-2.6.35-r1 sys-cluster/pvfs2-2.7.1 sys-libs/gdbm-1.8.3-r4 sys-libs/pam-1.1.1-r2 sys-libs/db selected: 4.3.29_p1-r1 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 4.6.21_p4 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 4.2.52_p5-r1 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 3.2.9_p2 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 sys-libs/db selected: 1.85-r3 protected: none omitted: 4.5.20_p2-r1 4.7.25_p4 4.8.30 >>> 'Selected' packages are slated for removal. >>> 'Protected' and 'omitted' packages will not be removed. >>> Waiting 5 seconds before starting... >>> (Control-C to abort)... >>> Unmerging in: 5 4 3 2 1 >>> Unmerging sys-libs/db-4.3.29_p1-r1... No package files given... Grabbing a set. <<< obj /usr/share/db-4.3/package.env <<< obj /usr/share/db-4.3/lib/db-4.3.jar <<< obj /usr/sbin/berkeley_db43_svc <<< obj /usr/lib/libdb_tcl-4.3.so <<< obj /usr/lib/libdb_tcl-4.3.la <<< obj /usr/lib/libdb_tcl-4.3.a <<< obj /usr/lib/libdb_java-4.3.so <<< obj /usr/lib/libdb_java-4.3.la <<< obj /usr/lib/libdb_java-4.3.a <<< obj /usr/lib/libdb_cxx-4.3.so <<< obj /usr/lib/libdb_cxx-4.3.la <<< obj /usr/lib/libdb_cxx-4.3.a <<< obj /usr/lib/libdb-4.3.so <<< obj /usr/lib/libdb-4.3.la <<< obj /usr/lib/libdb-4.3.a <<< obj /usr/include/db4.3/db_cxx.h <<< obj /usr/include/db4.3/db_185.h <<< obj /usr/include/db4.3/db.h <<< obj /usr/bin/db4.3_verify <<< obj /usr/bin/db4.3_upgrade <<< obj /usr/bin/db4.3_stat <<< obj /usr/bin/db4.3_recover <<< obj /usr/bin/db4.3_printlog <<< obj /usr/bin/db4.3_load <<< obj /usr/bin/db4.3_dump <<< obj /usr/bin/db4.3_deadlock <<< obj /usr/bin/db4.3_checkpoint <<< obj /usr/bin/db4.3_archive <<< dir /usr/share/db-4.3/lib <<< dir /usr/share/db-4.3 --- !empty dir /usr/share --- !empty dir /usr/sbin --- !empty dir /usr/lib <<< dir /usr/include/db4.3 --- !empty dir /usr/include --- !empty dir /usr/bin --- !empty dir /usr * Creating db.h symlinks to db4.8 >>> Regenerating /etc/ld.so.cache... >>> Unmerging sys-libs/db-4.6.21_p4... No package files given... Grabbing a set. <<< obj /usr/share/db-4.6/package.env <<< obj /usr/share/db-4.6/lib/db-4.6.jar <<< obj /usr/sbin/berkeley_db46_svc <<< obj /usr/lib/libdb_tcl-4.6.so <<< obj /usr/lib/libdb_tcl-4.6.la <<< obj /usr/lib/libdb_tcl-4.6.a <<< obj /usr/lib/libdb_java-4.6.so <<< obj /usr/lib/libdb_java-4.6.la <<< obj /usr/lib/libdb_java-4.6.a <<< obj /usr/lib/libdb_cxx-4.6.so <<< obj /usr/lib/libdb_cxx-4.6.la <<< obj /usr/lib/libdb_cxx-4.6.a <<< obj /usr/lib/libdb-4.6.so <<< obj /usr/lib/libdb-4.6.la <<< obj /usr/lib/libdb-4.6.a <<< obj /usr/include/db4.6/db_cxx.h <<< obj /usr/include/db4.6/db_185.h <<< obj /usr/include/db4.6/db.h <<< obj /usr/bin/db4.6_verify <<< obj /usr/bin/db4.6_upgrade <<< obj /usr/bin/db4.6_stat <<< obj /usr/bin/db4.6_recover <<< obj /usr/bin/db4.6_printlog <<< obj /usr/bin/db4.6_load <<< obj /usr/bin/db4.6_hotbackup <<< obj /usr/bin/db4.6_dump <<< obj /usr/bin/db4.6_deadlock <<< obj /usr/bin/db4.6_codegen <<< obj /usr/bin/db4.6_checkpoint <<< obj /usr/bin/db4.6_archive <<< dir /usr/share/db-4.6/lib <<< dir /usr/share/db-4.6 --- !empty dir /usr/share --- !empty dir /usr/sbin --- !empty dir /usr/lib <<< dir /usr/include/db4.6 --- !empty dir /usr/include --- !empty dir /usr/bin --- !empty dir /usr * Creating db.h symlinks to db4.8 >>> Regenerating /etc/ld.so.cache... >>> Unmerging sys-libs/db-4.2.52_p5-r1... No package files given... Grabbing a set. <<< obj /usr/share/db-4.2/package.env <<< obj /usr/share/db-4.2/lib/db-4.2.jar <<< obj /usr/sbin/berkeley_db42_svc <<< obj /usr/lib/libdb_tcl-4.2.so <<< obj /usr/lib/libdb_tcl-4.2.la <<< obj /usr/lib/libdb_tcl-4.2.a <<< obj /usr/lib/libdb_java-4.2.so <<< obj /usr/lib/libdb_java-4.2.la <<< obj /usr/lib/libdb_java-4.2.a <<< obj /usr/lib/libdb_cxx-4.2.so <<< obj /usr/lib/libdb_cxx-4.2.la <<< obj /usr/lib/libdb_cxx-4.2.a <<< obj /usr/lib/libdb-4.2.so <<< obj /usr/lib/libdb-4.2.la <<< obj /usr/lib/libdb-4.2.a <<< obj /usr/include/db4.2/db_cxx.h <<< obj /usr/include/db4.2/db_185.h <<< obj /usr/include/db4.2/db.h <<< obj /usr/bin/db4.2_verify <<< obj /usr/bin/db4.2_upgrade <<< obj /usr/bin/db4.2_stat <<< obj /usr/bin/db4.2_recover <<< obj /usr/bin/db4.2_printlog <<< obj /usr/bin/db4.2_load <<< obj /usr/bin/db4.2_dump <<< obj /usr/bin/db4.2_deadlock <<< obj /usr/bin/db4.2_checkpoint <<< obj /usr/bin/db4.2_archive <<< dir /usr/share/db-4.2/lib <<< dir /usr/share/db-4.2 --- !empty dir /usr/share --- !empty dir /usr/sbin --- !empty dir /usr/lib <<< dir /usr/include/db4.2 --- !empty dir /usr/include --- !empty dir /usr/bin --- !empty dir /usr * Creating db.h symlinks to db4.8 >>> Regenerating /etc/ld.so.cache... >>> Unmerging sys-libs/db-3.2.9_p2... No package files given... Grabbing a set. <<< obj /usr/share/doc/db-3.2.9_p2/README.bz2 <<< obj /usr/lib/libdb_cxx-3.2.so <<< obj /usr/lib/libdb_cxx-3.2.la <<< obj /usr/lib/libdb_cxx-3.2.a <<< obj /usr/lib/libdb-3.2.so <<< obj /usr/lib/libdb-3.2.la <<< obj /usr/lib/libdb-3.2.a <<< obj /usr/include/db3/db_cxx.h <<< obj /usr/include/db3/db_185.h <<< obj /usr/include/db3/db.h <<< obj /usr/bin/db3_verify <<< obj /usr/bin/db3_upgrade <<< obj /usr/bin/db3_stat <<< obj /usr/bin/db3_recover <<< obj /usr/bin/db3_printlog <<< obj /usr/bin/db3_load <<< obj /usr/bin/db3_dump185 <<< obj /usr/bin/db3_dump <<< obj /usr/bin/db3_deadlock <<< obj /usr/bin/db3_checkpoint <<< obj /usr/bin/db3_archive <<< dir /usr/share/doc/db-3.2.9_p2 --- !empty dir /usr/share/doc --- !empty dir /usr/share --- !empty dir /usr/lib <<< dir /usr/include/db3 --- !empty dir /usr/include --- !empty dir /usr/bin --- !empty dir /usr * Creating db.h symlinks to db4.8 >>> Regenerating /etc/ld.so.cache... >>> Unmerging sys-libs/db-1.85-r3... No package files given... Grabbing a set. <<< obj /usr/share/doc/db-1.85-r3/ps/recno.3.ps.bz2 <<< obj /usr/share/doc/db-1.85-r3/ps/mpool.3.ps.bz2 <<< obj /usr/share/doc/db-1.85-r3/ps/libtp.usenix.ps.bz2 <<< obj /usr/share/doc/db-1.85-r3/ps/hash.usenix.ps.bz2 <<< obj /usr/share/doc/db-1.85-r3/ps/hash.3.ps.bz2 <<< obj /usr/share/doc/db-1.85-r3/ps/dbopen.3.ps.bz2 <<< obj /usr/share/doc/db-1.85-r3/ps/btree.3.ps.bz2 <<< obj /usr/share/doc/db-1.85-r3/changelog.bz2 <<< obj /usr/share/doc/db-1.85-r3/README.hash.bz2 <<< obj /usr/share/doc/db-1.85-r3/README.bz2 <<< sym /usr/lib/libndbm.so <<< sym /usr/lib/libndbm.a <<< obj /usr/lib/libdb1.so.2 <<< sym /usr/lib/libdb1.so <<< obj /usr/lib/libdb1.a <<< sym /usr/lib/libdb.so.2 <<< sym /usr/include/ndbm.h <<< obj /usr/include/db1/ndbm.h <<< obj /usr/include/db1/mpool.h <<< obj /usr/include/db1/db.h <<< obj /usr/bin/db1_dump185 <<< dir /usr/share/doc/db-1.85-r3/ps <<< dir /usr/share/doc/db-1.85-r3 --- !empty dir /usr/share/doc --- !empty dir /usr/share --- !empty dir /usr/lib <<< dir /usr/include/db1 --- !empty dir /usr/include --- !empty dir /usr/bin --- !empty dir /usr >>> Regenerating /etc/ld.so.cache...
After unmerging all sys-libs/db except 4.5.20_p2-r1, 4.7.25_p4, 4.8.30 I re-compiled heimdal and now kdc does not crash anymore. What I do not understand why is that if ldd claimed in original report that the same db library 4.8 was being used (as is showing now as well, see below). checking db4/db.h usability... no checking db4/db.h presence... no checking for db4/db.h... no checking db3/db.h usability... no checking db3/db.h presence... no checking for db3/db.h... no checking db.h usability... yes checking db.h presence... yes checking for db.h... yes checking db_185.h usability... yes checking db_185.h presence... yes checking for db_185.h... yes checking for db_create... yes, in -ldb checking for dbm_firstkey... yes, in -ldb Another puzzling fact is this out from configure - showing that db-4.8 is not used during compilation and it falls back to an old, 1.85 which had different database drivers (some were removed I think in version 2 and also in 3). I think the 2.85 are coming from those newer db version but are isntalled as the fallback support as additional stuff. # equery files sys-libs/db-4.8.30 * Searching for db4.8.30 in sys-libs ... * Contents of sys-libs/db-4.8.30: [cut] /usr/include/db4.8/db.h /usr/include/db4.8/db_185.h /usr/include/db4.8/db_cxx.h [cut] /usr/lib/libdb-4.8.a /usr/lib/libdb-4.8.la /usr/lib/libdb-4.8.so [cut] In brief, heimdal requires 1.85 support enabled through USE=db185^H^H^H^H^H^H. It seems luckily --enable-compat185 is hardwired in db-4.5.20_p2-r1 and db-4.8.30 ebuilds (I haven't bothered to check others). I think there used to be a USE flags for that in the very past. I will try to break my system again by installation of some of the db-3.* series.
(In reply to comment #12) > After unmerging all sys-libs/db except 4.5.20_p2-r1, 4.7.25_p4, 4.8.30 I > re-compiled heimdal and now kdc does not crash anymore. Good. > I will try to break my system again by installation of some of the db-3.* > series. In general, you should avoid the temptation to mess with your db files. If you do and break your system, you get to keep the pieces. Dont forget to run revdep-rebuild. Closing.
I installed db-3 along the 4.x series, restarted existing kdc binary which albeit claiming by ldd to be linked against libdb-4.8.so was actually compiled against the db1.85 compat header (see comment #12). Then I tried su(1) to a root shell, and kdc did not crash. Then I recompiled heimdal and let it's configure be fooled by Gentoo FHS locations of the db files. After restarting kdc is crashes again. I suspect configure does not cope properly with Gentoo layout of the db files. See the original build.log and config.log for details. Below are the files installed by the db-3 package: # ls -lat /usr/lib/libdb* lrwxrwxrwx 1 root root 11 Aug 24 21:48 /usr/lib/libdb-1.so -> libdb1.so.2 lrwxrwxrwx 1 root root 12 Aug 24 21:48 /usr/lib/libdb-3.so -> libdb-3.2.so lrwxrwxrwx 1 root root 11 Aug 24 21:48 /usr/lib/libdb.a -> libdb-4.8.a lrwxrwxrwx 1 root root 12 Aug 24 21:48 /usr/lib/libdb.so -> libdb-4.8.so lrwxrwxrwx 1 root root 11 Aug 24 21:48 /usr/lib/libdb.so.2 -> libdb1.so.2 lrwxrwxrwx 1 root root 12 Aug 24 21:48 /usr/lib/libdb.so.3 -> libdb-3.2.so lrwxrwxrwx 1 root root 11 Aug 24 21:48 /usr/lib/libdb1.so -> libdb1.so.2 lrwxrwxrwx 1 root root 16 Aug 24 21:48 /usr/lib/libdb_cxx-3.so -> libdb_cxx-3.2.so lrwxrwxrwx 1 root root 15 Aug 24 21:48 /usr/lib/libdb_cxx.a -> libdb_cxx-4.8.a lrwxrwxrwx 1 root root 16 Aug 24 21:48 /usr/lib/libdb_cxx.so -> libdb_cxx-4.8.so lrwxrwxrwx 1 root root 16 Aug 24 21:48 /usr/lib/libdb_cxx.so.3 -> libdb_cxx-3.2.so lrwxrwxrwx 1 root root 16 Aug 24 21:48 /usr/lib/libdb_java.a -> libdb_java-4.8.a lrwxrwxrwx 1 root root 17 Aug 24 21:48 /usr/lib/libdb_java.so -> libdb_java-4.8.so lrwxrwxrwx 1 root root 15 Aug 24 21:48 /usr/lib/libdb_stl.a -> libdb_stl-4.8.a lrwxrwxrwx 1 root root 16 Aug 24 21:48 /usr/lib/libdb_stl.so -> libdb_stl-4.8.so lrwxrwxrwx 1 root root 15 Aug 24 21:48 /usr/lib/libdb_tcl.a -> libdb_tcl-4.8.a lrwxrwxrwx 1 root root 16 Aug 24 21:48 /usr/lib/libdb_tcl.so -> libdb_tcl-4.8.so I wonder why all the symlinks were re-created or their time was modified by db-3 ebuild. # ls -lat /usr/include/db3/ total 124 drwxr-xr-x 423 root root 36864 Aug 25 01:32 .. drwxr-xr-x 2 root root 4096 Aug 24 21:48 . -r--r--r-- 1 root root 51043 Aug 24 21:40 db.h -r--r--r-- 1 root root 5657 Aug 24 21:40 db_185.h -r--r--r-- 1 root root 19381 Aug 24 21:39 db_cxx.h # ls -lat /usr/include/db4.8 total 652 drwxr-xr-x 423 root root 36864 Aug 25 01:32 .. drwxr-xr-x 2 root root 4096 Aug 18 00:31 . -r--r--r-- 1 root root 107756 Aug 18 00:18 db.h -r--r--r-- 1 root root 5960 Aug 18 00:18 db_185.h -r--r--r-- 1 root root 18695 Aug 18 00:18 dbstl_common.h -r--r--r-- 1 root root 44005 Aug 18 00:18 db_cxx.h -r--r--r-- 1 root root 16156 Apr 12 22:25 dbstl_base_iterator.h -r--r--r-- 1 root root 18699 Apr 12 22:25 dbstl_container.h -r--r--r-- 1 root root 34627 Apr 12 22:25 dbstl_dbc.h -r--r--r-- 1 root root 18206 Apr 12 22:25 dbstl_dbt.h -r--r--r-- 1 root root 23473 Apr 12 22:25 dbstl_element_ref.h -r--r--r-- 1 root root 6613 Apr 12 22:25 dbstl_exception.h -r--r--r-- 1 root root 1934 Apr 12 22:25 dbstl_inner_utility.h -r--r--r-- 1 root root 106690 Apr 12 22:25 dbstl_map.h -r--r--r-- 1 root root 12814 Apr 12 22:25 dbstl_resource_manager.h -r--r--r-- 1 root root 49254 Apr 12 22:25 dbstl_set.h -r--r--r-- 1 root root 15452 Apr 12 22:25 dbstl_utility.h -r--r--r-- 1 root root 96868 Apr 12 22:25 dbstl_vector.h # ls -lat /usr/include/db_185.h lrwxrwxrwx 1 root root 14 Aug 24 21:48 /usr/include/db_185.h -> db4.8/db_185.h # ls -lat /usr/include/db.h lrwxrwxrwx 1 root root 10 Aug 24 21:48 /usr/include/db.h -> db4.8/db.h # As you could see above, the header symlinks pointing to 4.8 were touched by 3.x ebuild as well. I do not say this is a bug but still wonder why is that happening at all. To recapitulate, on the "broken" host I have installed and if db-4.x are not to be installed on the same host then they should block each other: 1.85-r3 3.2.9_p2 4.5.20_p2-r1 4.7.25_p4 4.8.30
Please do # cd /usr/include # ln -s db4 db4.8 # USE=berkdb emerge =heimdal-1.3.3-r1 and see if it still segfaults. Thanks.
(In reply to comment #15) > Please do > > # cd /usr/include > # ln -s db4 db4.8 I did "ln -s db4.8 db4" instead. > # USE=berkdb emerge =heimdal-1.3.3-r1 checking db4/db.h usability... yes checking db4/db.h presence... yes checking for db4/db.h... yes checking db3/db.h usability... yes checking db3/db.h presence... yes checking for db3/db.h... yes checking db.h usability... yes checking db.h presence... yes checking for db.h... yes checking db_185.h usability... yes checking db_185.h presence... yes checking for db_185.h... yes checking for db_create... yes, in -ldb checking for dbm_firstkey... yes, in -ldb Yes, kdc does not crash after restart and su(1) anymore.
Created attachment 245473 [details] heimdal-1.3.3-r2.ebuild Please check if this version works for you wrt db files.
(In reply to comment #17) > Created an attachment (id=245473) [details] > heimdal-1.3.3-r2.ebuild > > Please check if this version works for you wrt db files. > I did not delete the symlink I created manually on your request in /usr/include: lrwxrwxrwx 1 root root 5 Aug 27 11:57 db4 -> db4.8 checking db4.8/db.h usability... yes checking db4.8/db.h presence... yes checking for db4.8/db.h... yes checking db3/db.h usability... yes checking db3/db.h presence... yes checking for db3/db.h... yes checking db.h usability... yes checking db.h presence... yes checking for db.h... yes checking db_185.h usability... yes checking db_185.h presence... yes checking for db_185.h... yes checking for db_create... yes, in -ldb checking for dbm_firstkey... yes, in -ldb I re-started kdc and it does not die when user uses su(1). I then removed the manually introduced symlink, recompiled, restarted kdc: [cut] * Applying heimdal_db5.patch ... [ ok ] * Applying heimdal_testsuite.patch ... [ ok ] * Applying heimdal_testsuite_extra.patch ... [ ok ] * Applying heimdal_disable-check-iprop.patch ... [ ok ] * Applying heimdal_openssl-1.patch ... [ ok ] >>> Source prepared. >>> Configuring source in /var/tmp/portage/app-crypt/heimdal-1.3.3-r2/work/heimdal-1.3.3 ... * econf: updating heimdal-1.3.3/config.guess with /usr/share/gnuconfig/config.guess * econf: updating heimdal-1.3.3/config.sub with /usr/share/gnuconfig/config.sub ./configure --prefix=/usr --build=i686-pc-linux-gnu --host=i686-pc-linux-gnu --mandir=/usr/share/man --infodir=/usr/share/info --datadir=/usr/share --sysconfdir=/etc --localstatedir=/var/lib --enable-kcm --disable-osfc2 --enable-shared --with-libintl=/usr --with-readline=/usr --with-sqlite3=/usr --libexecdir=/usr/sbin --disable-afs-support --enable-berkeley-db --disable-otp --disable-kx509 --disable-pk-init --enable-pthread-support --without-openldap --without-ipv6 --with-openssl=/usr --with-x [cut] checking db4.8/db.h usability... yes checking db4.8/db.h presence... yes checking for db4.8/db.h... yes checking db3/db.h usability... yes checking db3/db.h presence... yes checking for db3/db.h... yes checking db.h usability... yes checking db.h presence... yes checking for db.h... yes checking db_185.h usability... yes checking db_185.h presence... yes checking for db_185.h... yes checking for db_create... yes, in -ldb checking for dbm_firstkey... yes, in -ldb kdc works as well. Thanks for your efforts.
Thank you for testing. I will submit a proper patch upstream and revize the ebuild with the next bump.
(In reply to comment #19) > Thank you for testing. I will submit a proper patch upstream and revize the > ebuild with the next bump. > Would you mind providing us with 2.4 ebuild? ;-) Thanks!
(In reply to comment #20) > (In reply to comment #19) > > Thank you for testing. I will submit a proper patch upstream and revize the > > ebuild with the next bump. > > > > Would you mind providing us with 2.4 ebuild? ;-) Thanks! 1.4 of course. ;)
1.4 is a problematic release. 1.4.1 will be out soon and will be added to the tree. For the record, this bug has been properly fixed by the following commit: http://github.com/heimdal/heimdal/commit/a1c14b231996ebd72de69df1de472f08e82c2288
(In reply to comment #22) > 1.4 is a problematic release. 1.4.1 will be out soon and will be added to the > tree. > > For the record, this bug has been properly fixed by the following commit: > http://github.com/heimdal/heimdal/commit/a1c14b231996ebd72de69df1de472f08e82c2288 > Eray, would you mind committing the 1.3.3-r2 ebuild fixing the issue? How long is this bug open?
(In reply to comment #23) > Eray, would you mind committing the 1.3.3-r2 ebuild fixing the issue? How long > is this bug open? No. I am not planning to rev bump because of this bug. I will see what I can do for a 1.4.1_pre release.
*heimdal-1.4.1_pre20110216 (16 Feb 2011) 16 Feb 2011; Eray Aslan <eras@gentoo.org> +heimdal-1.4.1_pre20110216.ebuild, +files/heimdal-kadmind.confd, +files/heimdal-kadmind.initd-r1, +files/heimdal-kcm.confd, +files/heimdal-kcm.initd-r1, +files/heimdal-kdc.confd, +files/heimdal-kdc.initd-r1, +files/heimdal-kpasswdd.confd, +files/heimdal-kpasswdd.initd-r1, +files/heimdal_link_order.patch: Version bump. Dropped m68k bug #324097. Dropped s390 and sh bug #355253. Double blocker to mit-krb5 bug #296610. Better support for sys-libs/db bug #333341. Added confd options bug #339340.
Thanks, works for me.
