Wizzup | Any clue when fix for this https://bugs.devuan.org/cgi/bugreport.cgi?bug=353 will hit beowulf? | 12:52 |
---|---|---|
Wizzup | base-files=10.3+devuan3.2 | 12:52 |
Wizzup | is not available yet | 12:52 |
LeePen | It has been held up by a possible kernel build failure. Although ATM I think that may be something else. | 12:53 |
LeePen | You can certainly download it manually and install from beowulf-proposed: | 12:53 |
LeePen | https://pkgmaster.devuan.org/devuan/dists/beowulf-proposed/main/ | 12:54 |
Wizzup | thank you. | 12:55 |
LeePen | The actual file is https://pkgmaster.devuan.org/devuan/pool/main/b/base-files/base-files_10.3%2Bdevuan3.2_all.deb | 12:55 |
LeePen | If you do, it would be good to hear if you have no problems. | 12:56 |
LeePen | Thanks. | 12:56 |
Wizzup | I have over 150 packges to rebuild, and not having this causes a lot of headaches :)( | 12:56 |
Wizzup | files that get built without -O2, without debug symbols, useless dbgsym packages, etc. | 12:56 |
LeePen | That should give it a good test! | 12:57 |
Wizzup | it runs in jenkins, though, so having it avail in normal repo would make life easier :p | 12:57 |
LeePen | Well you could add beowulf-proposed to the sources? | 12:57 |
Wizzup | yeah, I guess | 12:58 |
Wizzup | parazyd: ^^ | 12:58 |
gnu_srs1 | LeePen: Upgraded from Buster to Beowulf+unstable with the latest udev, eudev packages. No problems except with modules for kernel: 5.4.0-3-amd64 | 14:47 |
gnu_srs1 | However, after rebooting that kernel seems to be OK too. | 14:48 |
gnu_srs1 | dpkg -l | grep systemd <empty> :) ;) :D | 14:49 |
LeePen | Great/ Could you try an ascii->beowulf migration too? | 14:49 |
LeePen | ceres debootstrap seems fixed as well. | 14:50 |
LeePen | Thanks. | 14:50 |
gnu_srs1 | OK, but ascii->beowulf would be ascii->beowulf+unstable, since *udev packages are not yet built for Beowulf. | 14:51 |
LeePen | Hmmm. Might be too fiddly to be useful. | 14:56 |
LeePen | Been working on reportbug today. Fixed #54 and #134. | 16:01 |
LeePen | Also sorted having devuan's list of pseudopackages ratehr than debian's. | 16:02 |
LeePen | Looking at the pseudo packages, do we need one for mirrors onefang, Evilham? | 16:02 |
onefang | Other than the recent spate of "CC.deb.devuan.org doesn't work" reports about broken mirrors have been very rare. | 16:32 |
gnu_srs1 | BTW: Regarding ascii->beowulf I've tested several upgrades of eudev from 3.2.2-13 to 3.2.7-6 to 3.2.9-3 back and forth, no problems. | 16:33 |
gnu_srs1 | And udev packages are not in play here. | 16:33 |
onefang | Of those reports, if I recall correctly, tend to resolve to - That mirror hasn't updated yet, wait half an hour. Pilot error. Local network glitch. Something covered by apt-panoptican that I already knew about. Something covered by apt-panopticon that a mirror operator reported to me and we worked through it. | 16:35 |
onefang | In that last category, one resolved to a bug in apt-panopticon that I fixed. B-) | 16:36 |
LeePen | onefang: That's fine. I was just offering as there are a couple or mirror bugs on the BTS that I wanted to clean up. | 16:36 |
LeePen | Evilham: Are you happy #228 is fixed? | 16:37 |
onefang | Ah, I haven't seen those. Mirror bugs would get emailed to Evilham and myself automatically? | 16:37 |
LeePen | Only it we setup a mirror pseudopackage with you as maintainers. | 16:39 |
LeePen | HAve a look at #250 (useless!) | 16:39 |
onefang | OK, that works for me. Let's see what Evilham thinks. | 16:40 |
gnu_srs1 | I just upgraded eudev from 3.2.9-3 to 3.2.9-4 on a beowulf image. No problems :P | 16:40 |
LeePen | gnu_srs1: That's great work! | 16:40 |
gnu_srs1 | tks ;) | 16:41 |
onefang | Yep, useless and fits into the "CC.deb.devuan.org doesn't work" category. | 16:42 |
LeePen | Yep | 16:45 |
onefang | The BTS search system is also useless. Can't just do a "list all bugs that mention mirrors". Or even just list the lot. | 16:48 |
onefang | So which are the other mirror bugs you want cleaned up? | 16:48 |
onefang | Or, if you retroactively move them to the mirror pseudo package, will I get an email? | 16:50 |
LeePen | Actually it is closed already: #93. | 16:51 |
LeePen | Yes, for still open ones. | 16:51 |
LeePen | I am really just cleaning up and collecting together. | 16:52 |
onefang | Yep, that's been fixed for ages. | 16:53 |
LeePen | Thanks. | 16:53 |
onefang | https://pkgmaster.devuan.org/devuan_mirror_walkthrough.txt says to use deb.debian.org for that. | 16:58 |
onefang | Oh, the other category is - mirror problems detected by apt-panopticon that I haven't talked to mirror admins about yet, coz this unexpected server migration has soaked up all my time, and I was gonna start on that at the beginning of the year. | 17:04 |
gnu_srs1 | Just a simple question: When are we going to scrap *-proposed, and automatically merge unstable packages to testing after a number of days, like Debian? | 17:21 |
Evilham | LeePen: the pseudo-package could be useful | 17:38 |
LeePen | gnu_srs1: When beowulf is released. We can't do it now, because packages built in unstable can get the wrong binary dependencies to migrate in to beowulf. See https://pkgmaster.devuan.org/britney/excuses.html | 17:38 |
LeePen | Evilham onefang: OK. mirrors? devuan-mirrors? | 17:39 |
Evilham | package-mirrors | 17:39 |
LeePen | You get to choose ;) | 17:39 |
Evilham | or deb-mirrors | 17:39 |
onefang | There are package mirrors, and ISO mirrors. They are separate. | 17:40 |
onefang | Though some, like mine, do both. | 17:40 |
onefang | apt-panopticon only probes package mirrors, though I have a TODO to probe ISO mirrors as well. | 17:41 |
onefang | And there are sometimes bug reports for ISO mirrors. Would both be confusing to bug reporters? | 17:42 |
LeePen | Not necessarily. You can have a sentence to explain what they are for. | 17:42 |
onefang | I'll leave that up to Evilham to choose then. B-) | 17:43 |
Evilham | I'd have two separate pseudo-packages; either package-mirrors or deb-mirrors and files-mirrors or iso-mirrors, and I don't really think it that necessary for the latter, but for the package mirrors it does make sense | 17:44 |
onefang | "Package" and "ISO" is the language I usually use, FWIW. | 17:45 |
LeePen | OK. | 17:46 |
Evilham | LeePen: re reportbug, do you mean the control file in all devuan packages has to be changed? :-/ | 17:46 |
Evilham | how about a non-forked package? in that case would the report go to debian? | 17:46 |
LeePen | Unfortunately. | 17:46 |
LeePen | Yes, that is the advantage, so we only get the reports of forked packages. | 17:47 |
Evilham | hum, that's not necessarily great | 17:47 |
LeePen | Explain your thoughts. | 17:47 |
Evilham | I mean: it is likely that a report against e.g. nginx has a solution that should go through debian, but not necessarily something debian wants or should deal with on the very first instance | 17:48 |
Evilham | because the origin of the bug, could be in something that is devuan-specific | 17:48 |
Evilham | so, by having non-forked packages reach debian, we are potentially spamming them with unrelated bugs | 17:49 |
Evilham | *bug reports for non-forked packages | 17:49 |
Evilham | and that'd likely be detrimental long-term | 17:50 |
LeePen | OK. | 17:50 |
LeePen | Perhaps I was focusing too much on reducing our workload? | 17:51 |
Evilham | oh yes, I understand your logic perfectly, tbh I'm quite unsure :-D I see this as a potentially big downside, but also it is indeed likely that the issue does come from debian and should be fixed there | 17:52 |
LeePen | The problem I have seen working on the BTS is that there are lots of bugs there that are ignored because we don't have a forked package. | 17:53 |
LeePen | The package that they are reported against don't have a maintainer (for what that's worth). | 17:53 |
Evilham | yup, they also didn't have the visibility they have now thanks to you | 17:53 |
onefang | Some of the time the issue has to be passed all the way upstream to the packages developers. Might be helpful to skip a step. | 17:54 |
fsmithred | It's a good idea to make sure it's not a devuan-specific bug before bothering debian about it | 17:54 |
Evilham | it's a trade-off and I don't think Devuan should make that call unilaterally, since it does affect Debian's BTS and their maintainers | 17:54 |
LeePen | I am not sure who we would ask. | 17:55 |
LeePen | We could go stright to teh top and ask Sam Hartmann. He is a decent chap. | 17:55 |
Evilham | could this be sth where we can poke DPL privately and be like "hey, this is up, we don't think we should make the call on our own" | 17:56 |
onefang | What do the hundreds of other Debian based distros do? How annoying is that to Debian? | 17:56 |
Evilham | precisely :-D | 17:56 |
LeePen | OK. I'll ask him, I have had a few email conversations with him recently and he is very open and strightforward. | 17:56 |
Evilham | onefang: Devuan is not "just another Debian-based distro", most Debian-based distros repackage things heavily, our approach is to fork as few as needed | 17:57 |
onefang | Alas gone are the days when the then DPL lived near me, and we where members of the same user group. | 17:57 |
Evilham | AFAIK just re-using the .deb's as they are produced in Debian is actually quite unique | 17:57 |
LeePen | However, the Origin/Vendor dpkg system is built in to Debian, so in one sense they can't complain if distros use it! | 17:58 |
LeePen | Yes, most other recompile. | 17:58 |
LeePen | I'll ask him and see. | 17:58 |
LeePen | Thanks. | 17:58 |
Evilham | I think Origin is more of a "bugs from third-party reports shouldn't bother Debian" kind of thing | 17:59 |
Evilham | anyway, if you could ask DPL that'd be great, I think it's the best approach :-) | 17:59 |
LeePen | Will do. | 17:59 |
fsmithred | is there a way to make our bug tracker check for similar bugs in debian for packages we don't change? | 18:00 |
LeePen | Could you verify #228 is fixed? I certainly don't see the behaviour. It seems to be amprolla more than apt. | 18:00 |
LeePen | fsmithred: possibly. The whole soap debbugs thing is difficult anyway and the modules only work with Debian BTS ATM. | 18:02 |
fsmithred | or automatically tell the submitter to check for debian bugs | 18:02 |
onefang | I very rarely see anything other than short package descriptions in synaptic since I switched to Devuan. | 18:02 |
LeePen | fsmithred: That is easier, and easy for somebody to ignore as a line from reportbug goes past. | 18:04 |
fsmithred | do all submissions get an email response? | 18:04 |
LeePen | Yes, unless you add the X-Debbugs-No-Ack header. | 18:05 |
onefang | A very quick look through a random Packages file, and all I saw was short one sentence descriptions. | 18:05 |
fsmithred | I think in general, we need to reply with "more info needed" | 18:07 |
Evilham | LeePen: it isn't, e.g. apt search WebDAV doesn't return nginx-full, which should be returned | 18:07 |
Evilham | and apt info nginx-full only show sthe short description | 18:07 |
Evilham | I'll add a task about looking into translations in amprolla, but touching it is... not necessarily pleasant :-D | 18:09 |
Evilham | (a task --> my personal task list) | 18:09 |
LeePen | Wear gloves? | 18:09 |
LeePen | I'll reassign that to amprolla then. | 18:09 |
onefang | "apt-cache search keyer" as given in the bug report on ascii returned nothing, other search terms renurned short descriptions. The bug report was for jessie though. | 18:10 |
Evilham | onefang: up-to-date debian returns 3 packages :-p it's the same issue | 18:10 |
onefang | With short or long descriptions? It's the length of the description that is the bug. As I said, when looking at descriptions in synaptic on Devuan the first time, I noticed all those short descriptions, which I didn't see in Debian before I migrated. | 18:12 |
onefang | With shorter descriptions, there's less text to match the search term, so less matches. | 18:17 |
Evilham | those are the translations, yes | 18:23 |
Evilham | those are missing atm | 18:23 |
LeePen | Evilham onefang: Just setting up the new pseudopackages. Maintainer email address can only be 1 address. Who is going to take the lead? | 18:26 |
* onefang and Evilham point to each other. | 18:27 | |
LeePen | :-D | 18:27 |
Evilham | sigh mirrors@devuan.org? :-D | 18:27 |
LeePen | Does that exist and work? | 18:27 |
onefang | I was about to suggest that. | 18:28 |
Evilham | aye exists, works | 18:28 |
LeePen | And you see it? | 18:28 |
Evilham | often enough, yes | 18:28 |
LeePen | Excellent. Thanks. | 18:28 |
onefang | It gets delivered to a mailbox on pkgmaster that I monitor several times a day the hard way, and Evilham monitors every now and then the easy way. | 18:29 |
Evilham | hum? :-D the hard way? | 18:29 |
onefang | "ls Maildir/new" more or less. | 18:29 |
onefang | Well, I keep it open in MC and hit Ctrl-R every now and than. B-) | 18:30 |
onefang | Last email to it was sent by me, near Christmas, the one before that was spam. | 18:31 |
onefang | Most are it's subscription to devuan-mirrors@lists.dyne.org, which we both have our separate subs to anyway. | 18:32 |
onefang | At least we are in agreement. Use mirrors@devuan.org. B-) | 18:36 |
LeePen | Evilham: can I put you down as maintainer for amprolla? | 18:36 |
Evilham | I guess that makes sense, is amprolla a pseud-package? | 18:36 |
LeePen | fsmithred: can I put you down as maintainer for minimal-live? | 18:36 |
onefang | That's one I can avoid. B-) | 18:36 |
LeePen | Evilham: yes. | 18:36 |
fsmithred | LeePen, yes | 18:37 |
LeePen | Great, thanks all. | 18:37 |
Evilham | maybe someone points at self should make that a real package | 18:37 |
LeePen | That is the new pseudo packages set up. We need somebody to take responsibility for arm-sdk. hkraemer? | 19:00 |
LeePen | rrq: Will you take devuan-installer? | 19:01 |
onefang | rrq is probably asleep now, like a normal aussie, and unlike me, who isn't a normal aussie. B-) | 19:17 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!