2022-02-18T07:50:04  <adrianS> good morning ...
2022-02-18T07:50:14  <adrianS> bmwiedemann: any chance to grow the disk today? :)
2022-02-18T08:26:14  <bmwiedemann[m]> yes. I still have a meeting but should have some time later
2022-02-18T10:47:41  *** Water[m] is now known as Water7744[m]
2022-02-18T10:53:25  <bmwiedemann[m]> adrianS: resized => now there is 13611MB free
2022-02-18T10:55:35  <adrianS> bmwiedemann[m]: cool! thanks, let me test ...
2022-02-18T10:57:27  <adrianS> hm, currently directly 502 errors 
2022-02-18T10:57:37  <adrianS> not really starting to clone at all
2022-02-18T10:58:29  <adrianS> now it starts
2022-02-18T10:59:27  <adrianS> hm, still failing at 50-60% 
2022-02-18T11:04:42  * adrianS off for lunch
2022-02-18T11:31:51  * adrianS tests gitlab now ...
2022-02-18T11:44:21  <adrianS> works
2022-02-18T12:23:44  <LCP0968[m]> Conan Kudo (ニール・ゴンパ) would you have a look when you have the time?
2022-02-18T12:24:14  <Eighth_Doctor> with what?
2022-02-18T12:24:21  * Eighth_Doctor has no context
2022-02-18T12:24:48  <LCP0968[m]> code.o.o issue mentioned above
2022-02-18T12:25:28  <LCP0968[m]> set yourself a keyword for neal in element >:D
2022-02-18T12:25:43  <Eighth_Doctor> haha
2022-02-18T12:25:45  <Eighth_Doctor> there's already one for Conan Kudo
2022-02-18T12:26:12  <Eighth_Doctor> when I'm at-tagged on Discord or Matrix, I get pinged
2022-02-18T12:26:45  <LCP0968[m]> makes sense
2022-02-18T12:26:46  <Eighth_Doctor> same goes for IRC since my MXID is bound to my IRC nick
2022-02-18T12:26:53  <bmwiedemann[m]> There is also the problem that https://code.opensuse.org/package/fcitx5-chewing cannot be created (and maybe some others)
2022-02-18T12:27:05  <Eighth_Doctor> we're probably out of space, but I'll check
2022-02-18T12:27:18  <LCP0968[m]> no
2022-02-18T12:27:23  <LCP0968[m]> that's not it
2022-02-18T12:27:45  <LCP0968[m]> it seems like there is a resource shortage of some kind, but space was expanded already
2022-02-18T12:28:08  <bmwiedemann[m]> I maybe there was a problem earlier and some leftover blocks things now
2022-02-18T12:28:28  <Eighth_Doctor> damn it
2022-02-18T12:28:35  <Eighth_Doctor> all the journal logs were rotated and deleted already
2022-02-18T12:28:47  <Eighth_Doctor> wait, oh duh
2022-02-18T12:28:53  <Eighth_Doctor> not root 😅
2022-02-18T12:29:09  <LCP0968[m]> outstanding
2022-02-18T12:30:05  <Eighth_Doctor> oh for some reason redis died
2022-02-18T12:30:13  <Eighth_Doctor> that caused the queue to fail
2022-02-18T12:31:18  <bmwiedemann[m]> Adrian tried to push a 5GB repo earlier, when / only had 4GB available
2022-02-18T12:31:38  <Eighth_Doctor> hmm, I wonder if we're staging repos on /
2022-02-18T12:31:56  <Eighth_Doctor> we should be doing it on `/var/gitolite`
2022-02-18T12:32:04  <Eighth_Doctor> s/var/srv/
2022-02-18T12:32:09  <bmwiedemann[m]> watch df seemed to suggest that
2022-02-18T12:32:23  <Eighth_Doctor> that's mildly concerning
2022-02-18T12:32:34  <Eighth_Doctor> also, what the heck is Alp?
2022-02-18T12:32:42  <bmwiedemann[m]> I even tried to bind-mount /tmp onto /srv/gitolite, but it did not help
2022-02-18T12:32:51  <LCP0968[m]> 5gb repo
2022-02-18T12:33:24  <Eighth_Doctor> let me check pagure code on this
2022-02-18T12:33:34  <LCP0968[m]> alps is a brand of computer switches
2022-02-18T12:33:47  <Eighth_Doctor> O.o
2022-02-18T12:34:05  <LCP0968[m]> I guess they aren't computer exclusive
2022-02-18T12:34:11  <Eighth_Doctor> well, I wasn't expecting that
2022-02-18T12:34:21  <Eighth_Doctor> I half expected some weird codename related to suseEuler
2022-02-18T12:34:43  <LCP0968[m]> I don't think alp is related to alps
2022-02-18T12:37:00  <Eighth_Doctor> bmwiedemann: do you think you could identify a list of broken repos?
2022-02-18T12:37:11  <Eighth_Doctor> so I can delete them for you so you can reimport?
2022-02-18T12:39:29  <bmwiedemann[m]> yes. I'll write a check script, but it will take a while to get through 14915 fetches
2022-02-18T13:00:57  <adrianS> Eighth_Doctor: you can trigger it yourself by cloning via          git clone https://code.opensuse.org/adrianSuSE/Alp.git
2022-02-18T13:01:12  <adrianS> using gitlab.com for now, it seems to be able to deal with the repo
2022-02-18T13:01:32  <Eighth_Doctor> adrianS: yup, I'll take a look
2022-02-18T13:01:34  <adrianS> the clone aborts at ~ 50-65% 
2022-02-18T13:01:38  <Eighth_Doctor> I'm juggling a few things atm
2022-02-18T13:01:45  <adrianS> you need quite some bandwith :)
2022-02-18T13:02:05  <adrianS> no hurry, but if we should consider paguer in future it would be good to have it fixed
2022-02-18T13:02:16  <adrianS> atm all that is just playground ...
2022-02-18T13:02:34  <LCP0968[m]> I wonder what the performance of that is to the US
2022-02-18T13:02:54  <adrianS> hm, maybe you can use any VM in nuernberg to trigger it?
2022-02-18T13:03:01  <adrianS> or just tell me when I should run it ...
2022-02-18T13:03:27  <adrianS> we suspect that the git backend of pagure (written in python, right?) has a problem
2022-02-18T13:05:08  <LCP0968[m]> git backend is just gitolite, isn't it?
2022-02-18T13:05:25  <LCP0968[m]> it there some additional middleware?
2022-02-18T13:05:56  <Eighth_Doctor> git backend is pygit2
2022-02-18T13:06:00  <Eighth_Doctor> which is libgit2
2022-02-18T13:06:19  <Eighth_Doctor> the directory is poorly named now :)
2022-02-18T13:06:41  <LCP0968[m]> oh right
2022-02-18T13:07:14  <LCP0968[m]> that's true
2022-02-18T13:08:27  <bmwiedemann[m]> cloning over https goes through nginx and /srv/gitolite/.pagure_web.sock
2022-02-18T13:08:58  <Eighth_Doctor> cloning over ssh goes through pagure_worker and sshd
2022-02-18T13:09:15  <Eighth_Doctor> though pagure_worker only does an auth check
2022-02-18T13:16:50  <Eighth_Doctor> adrianS: is this your all-git experiment thing?
2022-02-18T13:20:28  <adrianS> yes, I turned devel:LEO into devel:ALP based on git basically
2022-02-18T13:20:46  <adrianS> but I will check if I can script somehow to put every package in a seperate git repo
2022-02-18T13:21:01  <adrianS> however, this is also not wanted by some ....
2022-02-18T13:21:07  <adrianS> lots of open discussion points ...
2022-02-18T13:21:51  <bmwiedemann[m]> https://github.com/bmwiedemann/openSUSE/blob/master/scripts/syncone2 is the splitter I used
2022-02-18T13:22:24  <bmwiedemann[m]> err. wrong link
2022-02-18T13:23:13  <bmwiedemann[m]> this one https://github.com/bmwiedemann/openSUSE/blob/master/scripts/splitone2
2022-02-18T13:23:49  <bmwiedemann[m]> The `git subtree split` turned out to be too slow for a repo of that size
2022-02-18T13:25:20  <adrianS> pague-new-package looks helpful
2022-02-18T13:27:26  <Eighth_Doctor> the per package repo thing makes permissions way easier too
2022-02-18T13:28:11  <Eighth_Doctor> monorepos are horrible for git in general :/
2022-02-18T13:28:18  <adrianS> the complain is that set of packages should be grouped
2022-02-18T13:28:50  <Eighth_Doctor> grouped by what?
2022-02-18T13:28:54  <adrianS> -> submit only entire devel projects to avoid extra testing needs
2022-02-18T13:29:09  <adrianS> since testing happened in devel project already
2022-02-18T13:29:12  <Eighth_Doctor> but that's not how things work even now?
2022-02-18T13:29:16  <adrianS> however, this is just one fraction
2022-02-18T13:29:20  <bmwiedemann[m]> `git subtree merge` can still create a monorepo from split package repos
2022-02-18T13:29:39  <adrianS> no one says it has to stay like it is now :)
2022-02-18T13:29:45  <Eighth_Doctor> hehe
2022-02-18T13:29:58  <Eighth_Doctor> having worked with both models, I have... opinions :)
2022-02-18T13:32:06  <bmwiedemann[m]> do you also have opinions on `git submodule` ?
2022-02-18T13:32:15  <bmwiedemann[m]> I have not used it much, yet but it seems to be designed to track refs to commits in separate repos
2022-02-18T13:32:23  <Eighth_Doctor> yeah
2022-02-18T13:33:00  <Eighth_Doctor> git-submodule is a good choice if you intend to make it clear nobody develops in that repo
2022-02-18T13:33:10  <Eighth_Doctor> you probably want automation to update submodule refs as things update, but it's good for that easy "overview" kind of thing
2022-02-18T13:34:04  <bmwiedemann[m]> there will certainly be scripts to manage things
2022-02-18T13:34:15  <adrianS> that was my proposal as well, but not entirely accepted ... because people hate submodules
2022-02-18T13:34:46  <bmwiedemann[m]> would be interesting to find out why.
2022-02-18T13:35:16  <bmwiedemann[m]> for me it is mostly that I know too little about it to make good use of it
2022-02-18T13:36:45  <Eighth_Doctor> people hate submodules because they're not designed to be interacted with directly
2022-02-18T13:36:52  <Eighth_Doctor> they're effectively read-only references
2022-02-18T13:37:04  <Eighth_Doctor> if you make changes in a submodule, you can't really commit and push right from there
2022-02-18T13:37:21  <Eighth_Doctor> subtree changes it so that the git trees are vendored in
2022-02-18T13:37:34  <Eighth_Doctor> so then you can modify them and not worry about syncing back with the upstream repo
2022-02-18T13:37:47  <Eighth_Doctor> git-subtree was invented for vendor monorepos
2022-02-18T13:37:55  <Eighth_Doctor> git-submodule was invented for tracking monorepos
2022-02-18T13:38:55  <Eighth_Doctor> one mindset thing that I don't like in openSUSE is that people are generally indifferent to landing things into the main distribution
2022-02-18T13:39:18  <Eighth_Doctor> and I think that comes from how devel projects work today
2022-02-18T13:39:35  <bmwiedemann[m]> I think, the only changes we would want to do in the collection repo would be to add+remove+update packages
2022-02-18T13:39:42  <Eighth_Doctor> yup
2022-02-18T13:39:49  <Eighth_Doctor> so that would make sense as a tracking monorepo
2022-02-18T13:40:15  <Eighth_Doctor> packages could be maintained in per package repos, and you could submit updates to the "factory" repo as submodule updates
2022-02-18T13:41:01  <adrianS> bmwiedemann[m]: I wonder if we could just adapt your upload scripts for now to apply the the RemoteAssets tags and cleanup the tar balls?
2022-02-18T13:41:16  <adrianS> bmwiedemann[m]: and I would just reference them via submodules for devel:ALP ?
2022-02-18T13:41:44  <bmwiedemann[m]> yes. And since there is only one source for packages, merge conflicts would not occur on the main repo, but in the individual package repos
2022-02-18T13:41:48  <Eighth_Doctor> adrianS: btw, do you have some script code for fetching RemoteAssets?
2022-02-18T13:42:21  <Eighth_Doctor> and out of curiosity, why `RemoteAsset` in the spec vs a separate file like RH/Fedora, OpenMandriva/ROSA, et al?
2022-02-18T13:42:26  <adrianS> download_assets .... part of the build script
2022-02-18T13:42:47  * Eighth_Doctor makes note to wire that into obsctl internally
2022-02-18T13:43:09  <adrianS> well, you do not want to touch two places on updating
2022-02-18T13:43:28  <adrianS> and duplicating information
2022-02-18T13:43:49  <Eighth_Doctor> yeah, I think the reason everyone else does it that way is because we don't want to do more spec parsing than we have to :)
2022-02-18T13:43:53  <Eighth_Doctor> wait, does download_assets support Dist-Git lookaside now?
2022-02-18T13:43:58  <adrianS> bmwiedemann[m]: https://code.opensuse.org/adrianSuSE/ALP-importer/blob/master/f/patch_spec_file
2022-02-18T13:44:10  <adrianS> bmwiedemann[m]: this script parses a spec file and adds the tags
2022-02-18T13:44:42  <adrianS> download_asset --clean       would remove the obsolete tar balls afterwards
2022-02-18T13:44:43  <Eighth_Doctor> lol, Novell
2022-02-18T13:45:59  <adrianS> bmwiedemann[m]: means of course we would need to push tar balls where no remote reference exists ...
2022-02-18T13:47:01  <bmwiedemann[m]> Conan Kudo: got the list of broken pagure repos: ca-certificates-mozilla-prebuilt fcitx5-anthy fcitx5-chewing fcitx5-hangul fcitx5-kkc fcitx5-libthai fcitx5-m17n fcitx5-sayura fcitx5-skk fcitx5-table-extra fcitx5-table-other fcitx5-unikey fcitx5-zhuyin hsetroot htcondor ip2unix isorelax llvm13 orthos-client perl-Mail-SpamAssassin-Plugin-dqs perl-MouseX-Getopt perl-XML-Generator-PerlData python3-pyside6
2022-02-18T13:47:01  <bmwiedemann[m]> python-azure-eventhub-checkpointstoreblob-aio python-ConfigArgParse python-mathics-pygments python-mpl-animators python-prettytable python-term-background python-weasyprint
2022-02-18T13:47:25  <bmwiedemann[m]> adrianS: I would prefer if there was a way to reference tarballs in some storage, such as my IPFS links do.
2022-02-18T13:49:53  <bmwiedemann[m]> as someone using such a repo, I always want to know "is this file the one that was meant to be there", so SHAsums and/or GPG sigs help there.
2022-02-18T13:55:14  <adrianS> there is already support for a sidestorage for fedpkg assets
2022-02-18T13:55:53  <adrianS> on the other side, the obs src server is also kind of a sidestorage for assets ...
2022-02-18T13:57:35  <bmwiedemann[m]> if there is a way to fetch a file by SHAsum from src server, that would help
2022-02-18T13:59:44  <Eighth_Doctor> sha256 or better ideally
2022-02-18T13:59:54  <Eighth_Doctor> fedora upgraded from md5 to sha512 some years ago
2022-02-18T14:00:03  <Eighth_Doctor> though old packages probably still use md5
2022-02-18T14:04:28  <Eighth_Doctor> bmwiedemann: done:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/9c2ff0cff2c6b15390c020ad99775eb8089efdb0)
2022-02-18T14:05:30  <Eighth_Doctor> adrianS: I wrote this a while ago to import Fedora packages quickly into OBS: https://pagure.io/obs-packaging-scripts/blob/master/f/distgit-obsimport.py :)
2022-02-18T14:06:41  <Eighth_Doctor> of course, that's because there was no fedpkg support in OBS :)
2022-02-18T14:07:20  <adrianS> sounds like the opposite way?
2022-02-18T14:08:57  <laki> hey, if I can ask the same question here, as it might be more appropriate
2022-02-18T14:09:27  <laki> where can I request access to a specific bug on bugzilla, or ask for it to be opened to public, if possible?
2022-02-18T14:09:36  <Eighth_Doctor> adrianS: it is, because I wrote it with the mindset that people don't do things in OBS and it'd fire when a commit landed
2022-02-18T14:09:48  <laki> this is mostly regarding quite old CVEs, so I can get info on who the original reporters were
2022-02-18T14:09:58  <Eighth_Doctor> and the openSUSE Build Service doesn't have a source service to run fedpkg-minimal to fetch sources
2022-02-18T14:11:24  <bmwiedemann[m]> Conan Kudo: I think one problem is that pagure serves the same thing under https://code.opensuse.org/package/python-WeasyPrint and https://code.opensuse.org/package/python-weasyprint though these were different packages in OBS
2022-02-18T14:11:44  <LCP0968[m]> laki, this feels like a factory question tbh, since it's more on the release management side of things
2022-02-18T14:12:02  <Eighth_Doctor> bmwiedemann[m]: that's a thing across most git forges, we're going to have a problem :(
2022-02-18T14:12:47  <Eighth_Doctor> I think we used to be case-sensitive here, and we stopped because it broke people
2022-02-18T14:13:24  <bmwiedemann[m]> @laki most security bugs should be public. Can you paste me some CVEs so I can look/ask
2022-02-18T14:13:54  <Eighth_Doctor> adrianS: I used my script for https://build.opensuse.org/project/show/home:Pharaoh_Atem:CS9EPELDev and https://build.opensuse.org/project/show/isv:Datto:Backports:PHP_Extensions
2022-02-18T14:14:54  <laki> bmwiedemann[m] : sure, here's one example -- CVE-2010-3087
2022-02-18T14:15:09  <laki> there were some more, though I don't know which specifically, but they were also libtiff bugs
2022-02-18T14:16:23  <bmwiedemann[m]> It says " Zarro Boogs found"
2022-02-18T14:16:35  <laki> ah, it seems to link to this
2022-02-18T14:16:37  <laki> https://bugzilla.suse.com/show_bug.cgi?id=624215
2022-02-18T14:17:14  <laki> I already got the info from another member, but I'm just wondering if I encounter more of these private bugs where could I ask for them to be disclosed
2022-02-18T14:25:08  <bmwiedemann[m]> I remember that there was an email on the Factory ML some year(s) ago on how to request bug-info to be provided, but my searching failed today
2022-02-18T14:30:31  <bmwiedemann[m]> laki: found it https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/MTLUY3MZRLN74AR55JYGFOE5BJ2IQQZU/
2022-02-18T14:34:13  <bmwiedemann[m]> adrianS: it seems, same as pagure, gitlab and github all ignore case of package names, which OBS did not - would that create trouble?
2022-02-18T14:36:42  <bmwiedemann[m]> Conan Kudo: but then, a git fetch for https://code.opensuse.org/package/python-weasyprint.git returns a 500, so the case-ignore seems to not work cleanly everywhere
2022-02-18T14:37:16  <Eighth_Doctor> ugh
2022-02-18T14:37:26  <Eighth_Doctor> it's probably related to how it exists on disk
2022-02-18T14:38:18  <Eighth_Doctor> bmwiedemann: could you file a ticket on the pagure issue tracker about that?
2022-02-18T14:38:40  <bmwiedemann[m]> you mean, we should use VFAT ? 
2022-02-18T14:38:56  <bmwiedemann[m]> will do
2022-02-18T14:40:17  <Eighth_Doctor> bmwiedemann[m]: oh god no
2022-02-18T14:46:40  <bmwiedemann[m]> Conan Kudo: https://code.opensuse.org/package/python-mathics-pygments is still broken for some reason. all others should be better now
2022-02-18T14:47:08  <Eighth_Doctor> because I messed up
2022-02-18T14:47:41  <bmwiedemann[m]> saw it: `'python-ConfigArgParse python-mathics-pygments'`
2022-02-18T14:47:47  <Eighth_Doctor> fuck
2022-02-18T14:47:53  <Eighth_Doctor> I just re-deleted all of them
2022-02-18T14:50:15  <Eighth_Doctor> blech I need to manually purge from disk now
2022-02-18T14:51:20  <bmwiedemann[m]> now most of them are broken and create fails
2022-02-18T14:51:43  <bmwiedemann[m]> I guess, you need a script for that to mess up less
2022-02-18T14:52:18  <bmwiedemann[m]> a `pagure-delete-pkg $PKG` that will do DB and FS updates.
2022-02-18T14:52:36  <Eighth_Doctor> yep
2022-02-18T14:52:43  <Eighth_Doctor> I need to make one of those
2022-02-18T14:53:06  <bmwiedemann[m]> so that packages are either there and working or they are deleted properly and can be recreated
2022-02-18T14:53:43  <Eighth_Doctor> yup
2022-02-18T14:54:02  <Eighth_Doctor> I'm going to whack everything again except python-ConfigArgParse and python-mathics-pygments
2022-02-18T14:54:08  <Eighth_Doctor> I assume those two are fine
2022-02-18T14:56:08  <bmwiedemann[m]> python-ConfigArgParse is broken, too for some reason
2022-02-18T15:00:13  <Eighth_Doctor> it's not on disk or the database
2022-02-18T15:00:28  <Eighth_Doctor> there's a lowercase one that's winning
2022-02-18T15:02:35  <bmwiedemann[m]> I have a feeling, this will occur more often
2022-02-18T15:05:16  <bmwiedemann[m]> can you ping me when the cleanup is done?
2022-02-18T15:07:03  <Eighth_Doctor> yeah
2022-02-18T15:07:12  <Eighth_Doctor> cleanup is done
2022-02-18T15:07:15  <Eighth_Doctor> try again
2022-02-18T15:53:45  <LCP0968[m]> I have been testing this channel and #opensuse-factory with the new bridges between matrix and irc, I may start rolling out the rest of the channels
2022-02-18T15:53:46  <LCP0968[m]> unless there are some objections/technical reasons not to
2022-02-18T15:55:13  <LCP0968[m]> I do suspect #opensuse-chat and #opensuse will be apprehensive to any change though, so that may end up being impossible there 😛
2022-02-18T16:08:20  <bmwiedemann[m]> Conan Kudo: still not good: {"error":"The requests repo package/ca-certificates-mozilla-prebuilt.git already
2022-02-18T16:08:20  <bmwiedemann[m]>  exists","error_code":"ENOCODE"} and `git push` fails with
2022-02-18T16:08:20  <bmwiedemann[m]> fatal: repository 'https://code.opensuse.org/package/ca-certificates-mozilla-prebuilt/' not found
2022-02-18T16:08:38  <Eighth_Doctor> blech
2022-02-18T16:09:10  <Eighth_Doctor> I'm in a meeting, Sasi Olin can also do these :)
2022-02-18T16:09:36  <bmwiedemann[m]> if there was a runbook, I could also
2022-02-18T16:25:17  <Eighth_Doctor> I should write scripts for this so people can do them
2022-02-18T16:26:36  <bmwiedemann[m]> 👍️
2022-02-18T16:29:02  <Eighth_Doctor> > <@bmwiedemann:opensuse.org> Conan Kudo: still not good: {"error":"The requests repo package/ca-certificates-mozilla-prebuilt.git already... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/75559034d13cf5b6562075d4e218419ae32b7464)
2022-02-18T16:30:36  <Eighth_Doctor>  * I don't see it here:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/c93354627f6d487708e5562a76ab7be9c9740c42)
2022-02-18T16:30:51  <Eighth_Doctor>  * I don't see it here:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/e06c21d7fc40107e681f8192c0ecfafc776820eb)
2022-02-18T16:30:53  <Eighth_Doctor>  * I don't see it here:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/3929469d91576ededa5765878f8eb64c943a7e8e)
2022-02-18T16:30:59  <Eighth_Doctor>  * I don't see it here:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/e7df72850f578026c4c46406f68f1bbd66691bd9)
2022-02-18T16:31:08  <bmwiedemann[m]> but that is what I still get from `curl --header 'Content-Type: application/json' --header 'Authorization: token XXX' --request POST --data '{"name":"ca-certificates-mozilla-prebuilt","description":"ca-certificates-mozilla-prebuilt", "namespace": "package", "wait": "true"}' https://code.opensuse.org/api/0/new`
2022-02-18T16:31:31  <Eighth_Doctor> restarting pagure services
2022-02-18T16:34:35  <bmwiedemann[m]> logs just show  - - [18/Feb/2022:16:33:30 +0000] "POST /api/0/new HTTP/1.0" 400 113 "-" "curl/7.66.0"
2022-02-18T16:37:19  <Eighth_Doctor> does the create call work now?
2022-02-18T16:37:56  <bmwiedemann[m]> strace showed it in `/srv/gitolite/repositories/requests/package/ca-certificates-mozilla-prebuilt.git`
2022-02-18T16:42:18  <Eighth_Doctor> oh requests...
2022-02-18T16:43:01  <Eighth_Doctor> holy crap
2022-02-18T16:43:07  <Eighth_Doctor> I forgot all about this
2022-02-18T16:44:06  <bmwiedemann[m]> I moved them to an `old` subdir, and now get another error
2022-02-18T16:44:55  <bmwiedemann[m]> maybe a create in pagure itself should even try to be atomic so either it all works or nothing happens
2022-02-18T17:23:51  <bmwiedemann[m]> The new `/usr/local/sbin/deletepackage` seems to do a decent job. I wonder to which git repo we should add that
2022-02-18T17:26:58  <LCP0968[m]> salt I would say
2022-02-18T17:33:20  <bmwiedemann[m]> now the remaining 4 failures are probably from case collisions: perl-MouseX-Getopt python-ConfigArgParse python-prettytable python-weasyprint
2022-02-18T17:51:24  <bmwiedemann[m]> manual delete through pagure web-UI helps there... only that it redirects to /user/bmwiedemann that throws a 502 (probably because I have too many repos)
2022-02-18T18:00:53  <bmwiedemann[m]> => https://pagure.io/pagure/issue/5279
2022-02-18T19:12:24  <bmwiedemann[m]> Conan Kudo:  is pagure's flask.request.url_root set to http://code.opensuse.org/ ? can we change it to https:// ? - because that is where the openid redirect goes and it might contain some token that we would not want to leak.
2022-02-18T19:13:05  <bmwiedemann[m]> or LCP#0968 ^^
2022-02-18T19:16:22  <LCP0968[m]> I don't think so
2022-02-18T19:16:28  <LCP0968[m]> this seems like the only place where the url is set for pagure https://code.opensuse.org/heroes/salt/blob/production/f/salt/profile/pagure/files/pagure.cfg#_76
2022-02-18T19:17:50  <bmwiedemann[m]> ah, it gets replaced from `self.app.config['PREFERRED_URL_SCHEME']`
2022-02-18T19:20:19  <bmwiedemann[m]> `/usr/lib/python3.6/site-packages/flask/app.py:        'PREFERRED_URL_SCHEME':                 'http',`
2022-02-18T19:26:00  <bmwiedemann[m]> LCP#0968: added the fix to MR 542
2022-02-18T19:31:48  <LCP0968[m]> I can't access the vpn atm on the account that my internet is down :D
2022-02-18T19:31:57  <LCP0968[m]> so I will merge it later I guess
2022-02-18T19:53:51  <Eighth_Doctor> bmwiedemann: it's fine with me to fix it
2022-02-18T19:54:16  <Eighth_Doctor> I can't access the VPN right now, as I'm doing work things on work laptop
2022-02-18T19:54:29  <Eighth_Doctor> but if nobody merges it by the time I get to it this evening, I'll do it
2022-02-18T19:55:15  <bmwiedemann[m]> thanks.
2022-02-18T20:37:36  <bmwiedemann[m]> PSA: if you login to code.opensuse.org now, id.o.o will ask for another confirmation, because I changed it to use https