# zypper -v dup
...
The following 3 packages are going to be downgraded:
python3-rpm 4.14.3-150400.41.4 -> 4.14.3-150300.46.1
rpm 4.14.3-150400.41.75 -> 4.14.3-150300.46.1
rpm-build 4.14.3-150400.41.75 -> 4.14.3-150300.46.1
4 packages to downgrade.
Overall download size: 1.7 MiB. Already cached: 0 B. After the operation, 49.0 B will be freed.
Continue? [y/n/v/...? shows all options] (y): n
That mismatch purposely and necessarily has been a tradition here. TDE historically hasn’t been built for Leap alphas or betas AFAI ever knew. On last refresh on this installation, zypper reported no such repo for 15.4. Besides, its the 15.3 Packman packages that the Trinity 15.3 packages require.
Also Repo 6 and 7 are for Leap 15.3, I would delete them.
This is the URL that might work: https://download.opensuse.org/repositories/home:/PunisherHD:/Trinity:/stable/15.4/
I didn’t know such repo existed, so thanks for that. I don’t remember seeing any announcement on either trinity mailing list that Trinity for openSUSE was ever provided by other than one single builder, who uploads to the TDE mirrors.
I’ve been using the latest Leap release version of Trinity and Packman for Leap betas going back much farther than I remember. Trinity is pretty well independent of Leap version, as long as it gets the Packman packages it needs.
Anyway, repos are not relevant to the backleveling issue. Except for (not installed) rpmkey-packman from Packman, Packman and Trinity provide no rpm.rpm packages. I did zypper al rpm, then did zypper dup --allow-vendor-change, which produced 10 downgrades, 15 vendor changes, 2 new packages & 23 packages to update, which left a total installed Trinity package count of 55. I followed that with zypper rl rpm, then tried zypper dup again, resulting in proposing to downgrade the same 3 packages as in OP.
The original question thus remains: Why have 15.4 repos replaced those three 15.4 rpm packages with the very same packages presently installed on 15.3?
Just curious. Your zypper lr seems to putput a very different set of columns then the zypper lr on my 15.3. Is that a 15.4 thing or did you manage to reconfigure what is produced by a plain zypper lr?
I am not sure if it is a good idea to confuse the less knowledgeble users here on the forums with a command that, when they use it, does not show what they see when you use it. But of course this should not stop you from customizing your own installations to your needs.
In the list as shown, I miss the Alias column. Aliases should be short, easy to use in a command (which I interpret as: no capitals and other “strange characters”) and yet easy to identify the repo with. And they the best IMHO to use in answers to a posted repo list with flaws like e.g. “Remove repo aap and disable repos noot and mies and vendor switch to repo packman”. Better then using the Numbers, because numbers will change during repo list maintenance.
I am aware of the fact that many Aliases are not as they should be (again IMHO). I have seen repos where the URL is used as an Alias :(. I do not blame the system managers for this, apparently some software (1-clicks, other installation tools?) are creating these monsters.
BTW, do I understand correctly that you want to change the default of which columns are shown when zypper lr is done wiithout options about which columns?
Please be aware of the problems that are created by changing defaults. People may use defaults because they like the default values as they are (that is probably why they are chosen as the default: because the designers hope they are the best/most used values). And they may have commands in scripts where the script depends on those defaults for further processing. Taking zypper lr as an example. I have a script (running on a regular base) that creates a repo list in HTML format in a web page (belonging to a set of pages describing a system, part of a web site that has those sets for all systems managed on the site). I use zypper lr -u -P in the script and thus this will not be influenced by an eventual change of the defaults. But alas if I would just use zypper lr. My script will fail.
Just an example of the wrath of many that one may call upon oneself when changing defautls.
The original question thus remains: Why have 15.4 repos replaced those three 15.4 rpm packages with the very same packages presently installed on 15.3?
# inxi -S
System:
Host: 00srv Kernel: 5.3.18-150300.59.60-default arch: x86_64 bits: 64
Console: pty pts/0 Distro: openSUSE Leap 15.3
# rpm -qa rpm
rpm-4.14.3-150300.46.1.x86_64
# rpm -q --changelog rpm | head -n8
* Thu Jan 20 2022 ...@suse.de
- Revert unwanted /usr/bin/python -> /usr/bin/python2 change we
got with the update to 4.14.3 [bsc#1194968]
new patch: no-python2.diff
* Wed Dec 22 2021 ...@suse.de
- Backport header check changes so that old rpms get no longer
rejected [bsc#1190824]
# inxi -S
System:
Host: p5bse Kernel: 5.14.21-150400.16-default arch: x86_64 bits: 64
Console: pty pts/0 Distro: openSUSE Leap 15.4 Beta
# rpm -qa rpm
rpm-4.14.3-150400.41.47.x86_64
# rpm -q --changelog | head -n43
* Wed Apr 21 2021 ...@suse.de
- set default package verification level to 'none' to be compatible
to rpm-4.14.1
new patch: verifylevel_none.diff
- make illegal obsoletes a warning
new patch: badobsoletewarn.diff
- backport header check security fixes from upstream [CVE-2021-3421]
[CVE-2021-20271] [CVE-2021-20266]
[bsc#1183543] [bsc#1183545]
new patch: headerchk3.diff
- fix potential access of freed mem in ndb's glue code [bnc#1179416]
new patch: ndbglue.diff
- permit secondary index open to fail for bdb_ro
new patch: bdbro_missingidx.diff
* Thu Apr 15 2021 ...@gmail.com
- update to rpm-4.14.3 [jsc#SLE-17817]
* add support for enforcing signature policy and payload
verification step to transactions
* add :humansi and :hmaniec query formatters for human readable output
* add query selectors –-whatobsoletes and –-whatconflicts
* add support for sorting caret (‘^’) higher than base version
- refreshed patches:
* checksepwarn.diff
* dbrointerruptable.diff
* finddebuginfo.diff
* ndb-backport.diff
* pythondistdeps.diff
* remove-brp-strips.diff
* whatrequires-doc.diff
- dropped patches:
* debugedit-bnc1076819.diff
* debugedit-riscv.patch
* disttag-macro.diff
* hardlinks.diff
* reproducible-debuginfo.patch
* rpm-4.14.1-initialize-verifyflags.diff
* safesymlinks.diff
* signbadregion.diff
* verifynodup.diff
* Mon Feb 22 2021 ...@suse.com
- Add rpm-4.14.1-initialize-verifyflags.diff [bsc#1180279]
# inxi -S
System:
Host: ab560 Kernel: 5.14.21-150400.16-default arch: x86_64 bits: 64
Console: pty pts/0 Distro: openSUSE Leap 15.4 Beta
# rpm -qa rpm
rpm-4.14.3-150400.41.75.x86_64
# rpm -q --changelog rpm | head -n15
* Wed Apr 21 2021 ...@suse.de
- set default package verification level to 'none' to be compatible
to rpm-4.14.1
new patch: verifylevel_none.diff
- make illegal obsoletes a warning
new patch: badobsoletewarn.diff
- backport header check security fixes from upstream [CVE-2021-3421]
[CVE-2021-20271] [CVE-2021-20266]
[bsc#1183543] [bsc#1183545]
new patch: headerchk3.diff
- fix potential access of freed mem in ndb's glue code [bnc#1179416]
new patch: ndbglue.diff
- permit secondary index open to fail for bdb_ro
new patch: bdbro_missingidx.diff
* Thu Apr 15 2021 ...@gmail.com
Last changes to python3-rpm were in September. rpm-build changes seem to match rpm exactly. I guess I don’t know how to look in the right place to find a reason why to go way back to the 15.3 versions in 15.4. Any ideas, maybe sent email to addresses in the changelogs? Or just wait and see what the next 15.4 build brings?
I’m only asking for reconsideration of defaults to be included in new installations. The main reason is the difficulty the wrapping that virtually always occurs creates when output including URIs is presented to a help forum. People asking for help are frustrated to start with. This frustration is extended by the additional wait imposed by a request for -d output after default output has already been presented, not to mention the frustration in attempting to evaluate wrapped output, output that not unusually is double wrapped. It’s a helper/helpee problem exacerbated by lengthy repo names and aliases that commonly add nothing to that inherent in the URIs.
If you bothered to actually read those changelogs beyond the first couple of lines you would have seen that 4.14.3-150300.46.1 is not “backlevel”, it is newer and includes all changes from other versions. To answer where previous versions came from one needs access to internal SUSE build infrastructure and I honestly do not care. 15.4 includes the latest available rpm package and that’s all that matters.
In case you have slept for the last year - you can no more reliably identify Leap release to which package belongs looking at package version. Leap directly includes binaries built for SLE and binaries are not rebuilt between SLE releases just for the sake of increasing release number.
I understand why you prefer the default being different (although I would prefer a different one, e.g. including Refresh, because we had several people with problems where it emerged that they had switched off refresh, and that was of course only seen from their zypper lr -d). But changing defaults is a pain in the … for many, has nothing to do with new installations or not. I will e.g. run the same scripts on each installation and steam will come out of my ears when it does not function on a new installation. But this warning against changing defaults is rather general and in fact off topic here.