On Fri, 2010-01-29 at 15:45 +0000, Rawhide Report wrote:
[snip]
Looks like whatever generates this report is reordering items in the changelog relative to the %changelog in the specfile, but in some surprising ways; is this a known issue?
Normally the items in a %changelog are in order most-recent through least-recent.
But in this report, see e.g. how 2.2-9 appears, then 2.2-10:
audacious-2.2-10.fc13
- Thu Jan 28 2010 Michael Schwendt mschwendt@fedoraproject.org - 2.2-9
- Let set_tuple_cb() work on a copied tuple (the metadata updates flood is too racy IMO).
- Fix tuple_copy().
- Thu Jan 28 2010 Michael Schwendt mschwendt@fedoraproject.org - 2.2-10
- Fix tuple_copy() further (it was completely broken as the mowgli dict wasn't copied at all).
Likewise 2.0.78-15 then 2.0.78-16:
policycoreutils-2.0.78-16.fc13
- Thu Jan 28 2010 Dan Walsh dwalsh@redhat.com 2.0.78-15
- Add use_resolve to sepolgen
- Thu Jan 28 2010 Dan Walsh dwalsh@redhat.com 2.0.78-16
- Cleanup spec file
Resolves: 555835
Up to that point, it looks like a reversal.
Here's where it gets weird: 0.6.10-1 then 0.6.10-2 then 0.6.9-4 then 0.6.9-2 seems an arbitrary ordering:
python-setuptools-0.6.10-2.fc13
- Fri Jan 29 2010 Toshio Kuratomi toshio@fedoraproject.org - 0.6.10-1
- Update the python3 portions but disable for now.
- Update to 0.6.10
- Remove %pre scriptlet as the file has a different name than the old package's directory
- Fri Jan 29 2010 Toshio Kuratomi toshio@fedoraproject.org - 0.6.10-2
- Really disable the python3 portion
- Tue Jan 26 2010 Toshio Kuratomi toshio@fedoraproject.org - 0.6.9-4
- Fix install to make /usr/bin/easy_install the py2 version
- Don't need python3-tools since the library is now in the python3 package
- Few other changes to cleanup style
- Fri Jan 22 2010 David Malcolm dmalcolm@redhat.com - 0.6.9-2
- add python3 subpackage
[snip]
And here's another one with a non-monotonic ordering:
valgrind-3.5.0-14.fc12
- Wed Jan 20 2010 Jakub Jelinek jakub@redhat.com 3.5.0-13
- DW_OP_mod is unsigned modulus instead of signed
- fix up valgrind.pc (#551277)
- Wed Jan 20 2010 Jakub Jelinek jakub@redhat.com 3.5.0-14
- fix build against latest glibc headers
- Mon Dec 21 2009 Jakub Jelinek jakub@redhat.com 3.5.0-12
- don't require offset field to be set in adjtimex's ADJ_OFFSET_SS_READ mode (#545866)
- Wed Dec 02 2009 Jakub Jelinek jakub@redhat.com 3.5.0-10
- add handling of a bunch of recent syscalls and fix some other syscall wrappers (Dodji Seketeli)
- handle prelink created split of .bss into .dynbss and .bss and similarly for .sbss and .sdynbss (#539874)
Any ideas? Known problem?
Thanks Dave