rbasak | cpaelzer: o/ I notice you synced ssh-import-id last week. I just did a new "upstream" release and have been trying to follow smoser's release process, which involved maintaining a VCS branch. But since it's now synced, that branch is out of date and has a different history to the synced package. I'm not sure how to resolve this while also doing it exactly as smoser did. Any suggestions? | 05:39 |
---|---|---|
rbasak | Maybe it's time to ditch ssh-import-id "upstream"'s packaging branch now, since there's one in Salsa, and we're in sync. | 05:39 |
rbasak | Then if a delta is required, I can do it without VCS or with just git-ubuntu, on the assumption that it should stay relatively small. | 05:40 |
cpaelzer | rbasak: I ensured that the sync has all we had before (as usual) | 06:15 |
cpaelzer | rbasak: which content would you want to release as new version? | 06:15 |
cpaelzer | 5.11 from master I guess seeing you tagging it yesterday | 06:16 |
cpaelzer | Well, the tag is enough for upstream as a release I guess. | 06:18 |
cpaelzer | Is the problem about making a tarball that matches the old style? | 06:18 |
cpaelzer | or about getting a debian/ubuntu upload that is 5.10-1 + the new changes? | 06:18 |
rbasak | cpaelzer: the latter. I can easily construct an upload of course, but I want to fit it into ssh-import-id's existing process for Ubuntu VCS maintenance which I think means changing it | 06:37 |
rbasak | https://gist.github.com/smoser/3cd61fe6e08f0c0955ef949350897d15#file-readme-md is what I followed | 06:37 |
cpaelzer | rbasak: TBH for the pkg-upload I'd recommend rebaseing 5.10..5.11 onto pkg/ubuntu/devel | 06:48 |
cpaelzer | rbasak: that should content-wise be the same, match the orig tgz and has individual commits | 06:49 |
cpaelzer | what else would you want? | 06:49 |
cpaelzer | and also yes, given that Debian seems to now care about it you could propose just the same change to them as well | 06:49 |
cpaelzer | which would allow to re-sync later as usual | 06:49 |
cpaelzer | and afterwards you could throw a "align debian/*" commit onto master that unites Debian/Upstream packaging | 06:51 |
cpaelzer | but anyway I start to outline my preferred way, you could easily have a different one and that still would be fine | 06:51 |
rbasak | 5.10 and 5.11 are upstream tags though | 06:51 |
rbasak | Rebasing the upstream branch onto a packaging branch seems wrong, unless I misunderstand? | 06:52 |
lordievader | Good morning | 06:52 |
cpaelzer | rbasak: it is the same as if you'd import a new orig tarball, but retaining individual commits | 06:53 |
cpaelzer | rbasak: what I wanted to say above, this probably should not be over-thought. Do it the way that you like and nothing bad will happen. | 06:53 |
cpaelzer | good morning lordievader | 06:54 |
rbasak | I'm trying to sort it out so there's a straightforward process for the future | 06:54 |
rbasak | Of course I can easily hack together something for this time | 06:54 |
cpaelzer | ah now I see your point, more for future than this upload ... | 06:54 |
rbasak | Right | 06:54 |
rbasak | I'd like the process to be quick and mindless | 06:55 |
rbasak | Right now it really isn't :-/ | 06:55 |
rbasak | I was hoping to just put Scott's gist into VCS and follow it exactly, so I wouldn't have to change anything, but AFAICT that won't work. | 06:55 |
rbasak | Let's see what Scott thinks | 06:56 |
rbasak | Right now I think my suggestion is to stop using a packaging branch at all, stop maintaining a debian/ directory in upstream master, and rely on Salsa and git-ubuntu for the packaging and Ubuntu delta. | 06:56 |
cpaelzer | that would work for me and make this "less special" | 06:57 |
cpaelzer | and maybe leave a Readme explaining that we'd want to avoid to maintain debian/* in two or three places and therefore this is how it now works | 06:57 |
rbasak | We're agreed then I think. Let's wait for Scott. On the Readme, my plan is to take Scott's gist and check it in to master, and I can add an explanation there. | 06:58 |
Tuor | Hi, I trying to find documentation on the unattended_upgrades configuration. I know where I have to change options, but the man doesn't explain the options and doesn't list the available options. | 11:31 |
RoyK | Tuor: https://wiki.debian.org/UnattendedUpgrades, perhaps? | 12:01 |
Tuor | RoyK: sort of. But it's not complete. I'm more looking to know all possible options that can be configured and what they do. | 12:05 |
Tuor | In the default config comming with the package, there are comments for a lot of options. Is the default config listing every available option? | 12:05 |
RoyK | I guess the hint is to RTFS - sorry don't know a better way | 12:06 |
Tuor | Which manual?! Where is the manual? I'm asking for the manual and you say I should read it. Really? | 12:08 |
lordievader | Tuor: I think you want `man unattended-upgrade` and `man 5 apt.conf` | 12:50 |
lordievader | Wait, apt.conf may not apply. | 12:52 |
lordievader | Not sure if there are options beyond those in the default `/etc/apt/apt.conf.d/50unattended-upgrades` config. | 12:56 |
RoyK | Tuor: RTFS as in Source | 13:12 |
RoyK | odc: it's just perl after all - should be readable | 13:13 |
odc | true | 13:14 |
ogra | isnt that actually mutually exclusive ? | 13:14 |
odc | I hate how people say perl isn't readable and seem to have no problem with haskell, rust, c++, etc. | 13:15 |
ogra | well, i have tons of perl scripts i wrote myself 10y ago that will take me a day to undersand again 🙂 | 13:15 |
odc | ah yes, that's a common problem ^^ | 13:17 |
* RoyK rather likes perl | 13:19 | |
TJ- | rather like knitting ... knit one, pearl one, drop one, curl one | 13:29 |
ogra | lol | 13:33 |
=== ddstreet_away is now known as ddstreet | ||
Tuor | Hmm, first time I'll read some perl, lets see how easy it is to read. :D | 14:05 |
Tuor | perl? looks like python to me, but I have no idea how perl looks.. | 14:06 |
ogra | my perl always looks like shell ... but so does my python as well 😛 | 14:12 |
Tuor | github.com:mvo5/unattended-upgrades.git has a readme :) | 14:27 |
Ussat | Well, foilks, Ubuntu as a server should get a TON more popular: https://blog.centos.org/2020/12/future-is-centos-stream/ | 14:57 |
quadrathoch2 | damn, oracle will have a huge uptake | 15:00 |
Ussat | Ya....I mean I am a mixed shop, RH and Ubuntu...but ya | 15:03 |
Tuor | has the apt package ovirt-guest-agent been named differently in focal or just dropped? | 15:17 |
quadrathoch2 | Tuor can't find it with packages.ubuntu.com :/ | 15:22 |
luna_ | \o/ | 15:54 |
Tuor | it seams qemu-guest-agent replaces ovirt-guest-agent... | 16:01 |
Tuor | at leat the description with apt show qemu-guest-agent looks similar to what apt show ovirt-guest-agent used to look. | 16:02 |
quadrathoch2 | sounds to me that wouldn't be the case, as ovirt is not qemu Tuor | 16:07 |
cpaelzer | Tuor: they are different things - ovirt-guest-agent talks with oVirt manager; qemu-guest-agent exposes other things (like freeze/thaw hooks, shutdown, ...) | 16:14 |
Tuor | hmm sad. | 16:15 |
Tuor | My Red Hat Virtualization is happy with qemu-guest-agent with ubuntu 20.04 and was not happy with the package with the same name under 18.04, with 18.04 I had to install ovirt-guest-agent. | 16:30 |
=== ijohnson is now known as ijohnson|lunch | ||
=== denningsrogue4 is now known as denningsrogue | ||
shubjero | coreycb: Do you know when we might expect the updated openstack-dashboard packages for CVE-2020-29565 for http://ubuntu-cloud.archive.canonical.com/ubuntu bionic-updates/train ? | 19:55 |
ubot3 | An issue was discovered in OpenStack Horizon before 15.3.2, 16.x before 16.2.1, 17.x and 18.x before 18.3.3, 18.4.x, and 18.5.x. There is a lack of validation of the "next" parameter, which would allow someone to supply a malicious URL in Horizon that can cause an automatic redirect to the provided malicious URL. <https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2020-29565> | 19:55 |
coreycb | shubjero: typically the ubuntu security team handles CVE fixes but I can see if we can get that moving along to help them out | 20:14 |
coreycb | shubjero: it looks like we need to get the ubuntu package patched for queens, stein, and train. ussuri+ already include it. | 20:15 |
=== ijohnson|lunch is now known as ijohnson | ||
=== vlm_ is now known as vlm | ||
coreycb | shubjero: we have that horizon change in progress | 23:04 |
=== vlm_ is now known as vlm |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!