I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 07/16/2009 10:32 AM, Rick L. Vinyard, Jr. wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
To extract sources from an SRPM: rpm2cpio <src.rpm> | cpio --extract (Do this in its own directory)
To enable the old checksum (for building RHEL packages): rpmbuild -bs --define _source_filedigest_algorithm=1 <spec>
This will recreate the SRPM using an MD5 sum instead of a SHA1 sum.
You can then use that in mock.
- -- Stephen Gallagher RHCE 804006346421761
Looking to carve out IT costs? www.redhat.com/carveoutcosts/
Stephen Gallagher wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 07/16/2009 10:32 AM, Rick L. Vinyard, Jr. wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
To extract sources from an SRPM: rpm2cpio <src.rpm> | cpio --extract (Do this in its own directory)
To enable the old checksum (for building RHEL packages): rpmbuild -bs --define _source_filedigest_algorithm=1 <spec>
This will recreate the SRPM using an MD5 sum instead of a SHA1 sum.
You can then use that in mock.
Excellent. Thanks.
On Thu, Jul 16, 2009 at 9:32 AM, Rick L. Vinyard, Jr. rvinyard@cs.nmsu.eduwrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
-- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
IIRC, Fedora 11 uses SHA-1 instead of MD5 for checksums, which makes new rpms not verifyable on pre-F11 computers.
Its part of the Stronger Hashes feature [1].
[1] http://fedoraproject.org/wiki/Features/StrongerHashes
Regards,
On Thu, Jul 16, 2009 at 16:32, Rick L. Vinyard, Jr.rvinyard@cs.nmsu.edu wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
I did the following two days ago (all on a F11 host) : - rebuild the F11 srpm in a F10 mock chroot - rebuild the F10 srpm in a EL5 mock chroot
F11 and F10 rpm versions are compatible, and so are F10 and EL5.
Totally suboptimal, but it works, and I didn't have time to search more thoroughly for a better solution :)
----------
Mathieu Bridon (bochecha)
On Thursday 16 July 2009 09:51:32 am Mathieu Bridon (bochecha) wrote:
On Thu, Jul 16, 2009 at 16:32, Rick L. Vinyard, Jr.rvinyard@cs.nmsu.edu
wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
I did the following two days ago (all on a F11 host) :
- rebuild the F11 srpm in a F10 mock chroot
- rebuild the F10 srpm in a EL5 mock chroot
F11 and F10 rpm versions are compatible, and so are F10 and EL5.
Totally suboptimal, but it works, and I didn't have time to search more thoroughly for a better solution :)
just for this type of thing i added a new script to fedora-packager rpmbuild- md5 it creates rpms/srpms with the old style hashes. its a simple bash script that is a wrapper around rpmbuild with the defines for old style hashes. the fedora-packager with it is in updates-testing.
Dennis
On 07/16/2009 12:42 PM, Dennis Gilmore wrote:
On Thursday 16 July 2009 09:51:32 am Mathieu Bridon (bochecha) wrote:
On Thu, Jul 16, 2009 at 16:32, Rick L. Vinyard, Jr.rvinyard@cs.nmsu.edu
wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
I did the following two days ago (all on a F11 host) :
- rebuild the F11 srpm in a F10 mock chroot
- rebuild the F10 srpm in a EL5 mock chroot
F11 and F10 rpm versions are compatible, and so are F10 and EL5.
Totally suboptimal, but it works, and I didn't have time to search more thoroughly for a better solution :)
just for this type of thing i added a new script to fedora-packager rpmbuild- md5 it creates rpms/srpms with the old style hashes. its a simple bash script that is a wrapper around rpmbuild with the defines for old style hashes. the fedora-packager with it is in updates-testing.
Would it be possible to update the mock configs with these settings as well?
-Doug
On Thursday 16 July 2009 12:07:42 pm Doug Warner wrote:
On 07/16/2009 12:42 PM, Dennis Gilmore wrote:
On Thursday 16 July 2009 09:51:32 am Mathieu Bridon (bochecha) wrote:
On Thu, Jul 16, 2009 at 16:32, Rick L. Vinyard, Jr.rvinyard@cs.nmsu.edu
wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
I did the following two days ago (all on a F11 host) :
- rebuild the F11 srpm in a F10 mock chroot
- rebuild the F10 srpm in a EL5 mock chroot
F11 and F10 rpm versions are compatible, and so are F10 and EL5.
Totally suboptimal, but it works, and I didn't have time to search more thoroughly for a better solution :)
just for this type of thing i added a new script to fedora-packager rpmbuild- md5 it creates rpms/srpms with the old style hashes. its a simple bash script that is a wrapper around rpmbuild with the defines for old style hashes. the fedora-packager with it is in updates-testing.
Would it be possible to update the mock configs with these settings as well?
-Doug
It is nothing to do with what is in the mock configs. the issue is that if you create a srpm on F-11 or rawhide then it uses internal hashing that is not understood by anything previous. F-10 uses md5 but understands the new style also.
inside the mock chroot they use what is the default for that release.
Dennis
On 07/16/2009 01:16 PM, Dennis Gilmore wrote:
It is nothing to do with what is in the mock configs. the issue is that if you create a srpm on F-11 or rawhide then it uses internal hashing that is not understood by anything previous. F-10 uses md5 but understands the new style also.
inside the mock chroot they use what is the default for that release.
Great, thanks for the explanation. I apparently didn't read enough of the thread to understand.
-Doug
Dennis Gilmore wrote:
On Thursday 16 July 2009 09:51:32 am Mathieu Bridon (bochecha) wrote:
On Thu, Jul 16, 2009 at 16:32, Rick L. Vinyard, Jr.rvinyard@cs.nmsu.edu
wrote:
I've built the following package on Fedora 10 and 11 in mock without
any
problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum
mismatch
Any suggestions?
I did the following two days ago (all on a F11 host) :
- rebuild the F11 srpm in a F10 mock chroot
- rebuild the F10 srpm in a EL5 mock chroot
F11 and F10 rpm versions are compatible, and so are F10 and EL5.
Totally suboptimal, but it works, and I didn't have time to search more thoroughly for a better solution :)
just for this type of thing i added a new script to fedora-packager rpmbuild- md5 it creates rpms/srpms with the old style hashes. its a simple bash script that is a wrapper around rpmbuild with the defines for old style hashes. the fedora-packager with it is in updates-testing.
Even better. Thanks.
Dennis Gilmore wrote:
On Thursday 16 July 2009 09:51:32 am Mathieu Bridon (bochecha) wrote:
On Thu, Jul 16, 2009 at 16:32, Rick L. Vinyard, Jr.rvinyard@cs.nmsu.edu
wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
I did the following two days ago (all on a F11 host) :
- rebuild the F11 srpm in a F10 mock chroot
- rebuild the F10 srpm in a EL5 mock chroot
F11 and F10 rpm versions are compatible, and so are F10 and EL5.
Totally suboptimal, but it works, and I didn't have time to search more thoroughly for a better solution :)
just for this type of thing i added a new script to fedora-packager rpmbuild- md5 it creates rpms/srpms with the old style hashes. its a simple bash script that is a wrapper around rpmbuild with the defines for old style hashes. the fedora-packager with it is in updates-testing.
the real problem is that you can't rebuild fedora's src.rpm on epel. the real good thing would be add a script which will be rebuild the new src.rpms.
Farkas Levente wrote:
Dennis Gilmore wrote:
On Thursday 16 July 2009 09:51:32 am Mathieu Bridon (bochecha) wrote:
On Thu, Jul 16, 2009 at 16:32, Rick L. Vinyard, Jr.rvinyard@cs.nmsu.edu
wrote:
I've built the following package on Fedora 10 and 11 in mock without any problems. http://miskatonic.cs.nmsu.edu/pub/libMini-9.0.3-2.fc11.src.rpm
However, when I try and build for EPEL 5 using mock on a F11 machine I get the following in root.log:
DEBUG util.py:256: error: unpacking of archive failed on file /builddir/build/SOURCES/MINI-9.0.3.zip;4a5f3532: cpio: MD5 sum mismatch
Any suggestions?
I did the following two days ago (all on a F11 host) :
- rebuild the F11 srpm in a F10 mock chroot
- rebuild the F10 srpm in a EL5 mock chroot
F11 and F10 rpm versions are compatible, and so are F10 and EL5.
Totally suboptimal, but it works, and I didn't have time to search more thoroughly for a better solution :)
just for this type of thing i added a new script to fedora-packager rpmbuild- md5 it creates rpms/srpms with the old style hashes. its a simple bash script that is a wrapper around rpmbuild with the defines for old style hashes. the fedora-packager with it is in updates-testing.
the real problem is that you can't rebuild fedora's src.rpm on epel. the real good thing would be add a script which will be rebuild the new src.rpms.
It wouldn't be that hard... just unpack and repack with MD5. The only problem I see is that a corrupted file would be repacked with a valid checksum.