| thecdnhermit | so i'm trying to change the keyboard shortcut ctrl+alt+<up-arrow> (open workspace selector screen) to ctrl+alt+KP0 but i can't find the original shortcut to change it | 00:40 |
|---|---|---|
| thecdnhermit | using cinnamon DE on devuan 5 | 00:40 |
| gnarface | i don't know the answer to that one but stick around, someone certainly does | 00:41 |
| thecdnhermit | oh crap i just found it LOL | 00:41 |
| gnarface | hooraaay! | 00:42 |
| thecdnhermit | in the "general" list | 00:42 |
| thecdnhermit | half the time that's how it works...get frustrated and ask, then find the answer immediately! | 00:43 |
| debdog | you're welcome :D | 00:44 |
| AlexLikeRock | good nigth | 04:04 |
| systemdlete | so back to building (or trying to) nut 2.8.2 for daedalus. First, I followed the steps at https://www.debian.org/doc/manuals/maint-guide/update.en.html#newupstream, with the 2.8.0 source (since that's what is in the repo), then I created a tarball of the 2.8.2 directory it created. Then I proceded to follow the steps at https://forums.debian.net/viewtopic.php?t=38976 (as appropriate, of course). And here is the output: | 08:06 |
| systemdlete | https://dpaste.com/A8CK8K8JT | 08:06 |
| systemdlete | I inspected the patch file it is crowing about, but I don't see anything wrong with it; besides, it is just a copy, really, from the 2.8.0, right? (Unless the scripts modified the patches) | 08:07 |
| systemdlete | I am doing all these steps as non-root user. | 08:08 |
| systemdlete | And I am doing these steps on a daedalus system. | 08:08 |
| systemdlete | Do I need to be root for this? (Other than for installation of the package eventually) | 08:09 |
| gnarface | systemdlete: make sure you have "fakeroot" installed just in case, as long as you do, no | 08:33 |
| systemdlete | fakeroot is installed | 08:42 |
| systemdlete | the script/patch it is crowing about: https://dpaste.com/5F4LFNHHE | 08:44 |
| systemdlete | and I have NOT touched any artifacts presented here or others generated by the dpkg scripts | 08:45 |
| systemdlete | I pray this problem is not even more collateral damage from porting around u-no-watt | 08:46 |
| gnarface | systemdlete: uh... paste.debian.net and i'll look at it, but if you can't tell what's wrong with it i probably can't either. did you check the count of difference in blank lines? | 10:23 |
| gnarface | if there's 4 blank lines of difference and the fuzz count is set to 3 it might fail while looking very much like the same patch at first glance | 10:24 |
| gnarface | (for example) | 10:24 |
| systemdlete | where do I find the fuzz count? | 10:25 |
| systemdlete | is that set by the dpkg build program? | 10:26 |
| gnarface | i'm not sure, but when it needs it to succeed the patch output usually says something like "succeeded with fuzz count X" or something like that | 10:27 |
| gnarface | i would have to assume you saw it somewhere in at least one or two of the other patches before... | 10:27 |
| gnarface | i think the default is 3, just not sure where it's set | 10:27 |
| systemdlete | ?? | 10:27 |
| systemdlete | It is stuck on the 1st patch | 10:27 |
| gnarface | oh | 10:28 |
| systemdlete | I think, given it is 0001... and it is the only one I see in the output | 10:28 |
| gnarface | Hunk #1 succeeded at 1 with fuzz 2 (offset -1 lines). | 10:29 |
| gnarface | looks like this | 10:29 |
| gnarface | i see them all the time building kernels from source packages | 10:29 |
| systemdlete | https://paste.debian.net/hidden/7756edea/ | 10:31 |
| systemdlete | https://paste.debian.net/hidden/fb5d2aca/ | 10:32 |
| systemdlete | buildpackage output, then the 0001 script | 10:32 |
| gnarface | ah | 10:33 |
| gnarface | dpkg-source: error: LC_ALL=C patch -t -F 0 -N -p1 -u -V never -E -b -B .pc/0001-reproductible... | 10:33 |
| gnarface | fuzz is set to 0 here | 10:33 |
| systemdlete | ok | 10:33 |
| gnarface | so if there's even a minor difference in whitespace it will fail | 10:33 |
| systemdlete | minor difference in... what's whitespace? | 10:34 |
| systemdlete | whose whitespace, which file(s) | 10:34 |
| systemdlete | before and after patch? | 10:34 |
| gnarface | difference in whitespace between the patch and the target file | 10:34 |
| gnarface | as in a different number of blank lines before and after the parts it's chaning | 10:35 |
| gnarface | changing* | 10:35 |
| systemdlete | and I can't figure out how to get a trace of what the buildpackage script is doing internally, so I don't have any idea which files it is applying to | 10:35 |
| gnarface | the output you pasted says it | 10:36 |
| gnarface | patching file docs/Makefile.am | 10:36 |
| gnarface | this one ^ | 10:36 |
| gnarface | and this one v | 10:36 |
| gnarface | patching file docs/man/Makefile.am | 10:36 |
| gnarface | those are the two fails | 10:36 |
| systemdlete | oh | 10:36 |
| gnarface | this one succeeds > patching file configure.ac | 10:36 |
| gnarface | look at lines 14-21 on that paste | 10:37 |
| systemdlete | most languages don't care too much about varying amounts of white space/blank lines | 10:37 |
| gnarface | it shows 3 files getting patched, one fail, one success, then one fail | 10:37 |
| gnarface | yes, but the patch binary is being called with -F 0 here | 10:38 |
| gnarface | so what matters is that patch is being told to care | 10:38 |
| systemdlete | right, but why does the build process care? | 10:38 |
| systemdlete | the logic eludes me | 10:38 |
| gnarface | well it bails out on the patch binary returning non-zeo | 10:38 |
| gnarface | zero* | 10:38 |
| gnarface | i think it's assumed that if any patches fail you'll want to fix them instead of trying to build it anyway | 10:39 |
| systemdlete | yes, I see that. I'm just trying to figure out why the designers of the build package tool were concerned about "fuzz" in the first place... if the patch script is accurate, that shouldn't be an issue | 10:39 |
| systemdlete | (I would think) | 10:40 |
| * gnarface shrugs | 10:40 | |
| gnarface | it should be easy to fix either way | 10:40 |
| systemdlete | but anyway, it seems I have to fix a bug in someone else's code | 10:40 |
| gnarface | looks that way, though sometimes sloppy build dependencies are the actual cause | 10:40 |
| gnarface | like maybe something assumed to be present or of a certain version and it's not | 10:41 |
| gnarface | if it's actually just whitespace it should be really easy to fix | 10:41 |
| gnarface | on the plus side, you've probably solved the mystery of why this version isn't in debian | 10:43 |
| systemdlete | ok, thanks, I'll dick around with this some more tomorrow. | 10:50 |
| m_m | Hi, I have a problem while debootstrapping devuan. I run on the latest liveCD on a virtual machine. debootstrap daedalus /mnt https://mirror.vgprp.io/devuan : Could not find these debs: ca-certificates | 11:50 |
| m_m | it seems that it cannot find any deb, I tried including openssh-server, and it returns "Could not find these debs: ca-certificates, openssh-server" | 11:51 |
| rrq | is that a vaild url ? | 11:59 |
| m_m | I tried with different mirrors | 12:03 |
| m_m | coming from pkgmaster.devuan.org | 12:03 |
| rrq | that host seems worse than some, but in any case the url should end with /merged | 12:03 |
| m_m | erf... ok that's because man debootstrap tells to use /debian at the end of the URI | 12:05 |
| m_m | thanks rrq ! | 12:05 |
| rrq | and you should use devuan's debootstrap (fork) | 12:07 |
| m_m | ok I'd like to but why isn't it the debootstrap command that comes with the devuan liveCD? | 12:09 |
| rrq | yes I would think it should be.. I don't know why/if it's not | 12:11 |
| m_m | I think it's the right one but the manpage is the one from debian | 12:11 |
| rrq | agree. It says "Devuan-Dev Team" is maintainer... maybe that detail is considered less urgent | 12:15 |
| m_m | heh, documentation is top one priority :p | 12:17 |
| m_m | in the best of the worlds | 12:17 |
| rrq | as you may know, "devuan" is not a corporation or company but a collective of volunteers | 12:17 |
| m_m | I know I was trolling | 12:17 |
| rrq | :) | 12:17 |
| m_m | and Devuan is amazing | 12:18 |
| m_m | it really gives me back hope when I see the state of linux distribution these days... I have much less problems using it than my friends using debian (or on the few machines left on which I use debian) | 12:19 |
| fsmithred | debootstrap is in the live isos. It's at /usr/sbin/debootstrap so if you used 'su' to become root instead of 'su -' you would get 'Command not found' | 12:25 |
| fsmithred | or call it with the full path | 12:26 |
| m_m | fsmithred: yeah that wasn't the point, we were talking about the "man debootstrap" | 12:28 |
| rrq | yes, the man page is not revised wrt devuan's fork, so the "/merged" url suffix is not ovious | 12:31 |
| rrq | b | 12:31 |
| m_m | currently installing :) | 12:31 |
| systemdlete | gnarface: the man page for patch seems to indicate that fuzz has nothing to do with whitespace but rather whether patch can find a matching context or not. But at any rate, it looks to me that the patch is for solaris usb nuances in the first place, so why this is even a step in building for debian is a mystery to me. Something in configure step must be generating the step. | 19:00 |
| bgstack15 | before I get too off-track, maybe somebody's experienced this too. When I try to use acetoneiso, I get this error: | 19:27 |
| bgstack15 | QIODevice::read (QFile, "/home/bgstack15/virtual-drives/1/readme.txt"): device not open | 19:27 |
| bgstack15 | I can't seem to find any useful info on the Internet about this error. Am I missing a driver, or something? I've reinstalled the apt package for acetoneiso. I can mount a .iso file manually with just `mount HPDJ2540.iso /mnt/foo`without any problems. | 19:28 |
| systemdlete | ahhhh. Finally making some progress. Had to repair 2 patch files and now I am getting past the hunking problems with patch step. Then I found I need the -b option to work around issues concerning "unexpected upstream changes" (found this by ddg search). It's building now... | 20:48 |
| systemdlete | wholly crap is it ever... | 20:49 |
| systemdlete | next step will be to test the thing and see if it (a) works and (b) corrects the disappearing devices issue, which is where this whole buggery started | 20:50 |
| systemdlete | thanks to gnarface and all the others who made suggestions; i've been trying to build this package for weeks. I think I am almost understanding how it works | 20:52 |
| systemdlete | (and that frightens me LOL) | 20:52 |
| cousin_luigi | So, when is Daedalus going to be on par with bookworm 12.8? | 21:24 |
| godbed | cousin_luigi: never! devuan is different, it 'misses out' on systemd. if'd be on par with debian, it'd be debian ;) | 21:31 |
| godbed | in other words: what do you mean? | 21:31 |
| systemdlete | godbed, I think they mean upgrades, not systemd | 21:32 |
| APic | 😉 | 21:34 |
| fsmithred | cousin_luigi, we will probably never have x.8 on any release. We're lucky if we get one point-release together. So if you want to have your devuan "on par" with debian, you have to do 'apt update && apt upgrade' after you install it. Then the package versions will all be the same as debian's. | 21:35 |
| fsmithred | Of course, as soon as some other package gets upgraded in the repo, the 12.8 isos will no longer have the latest software. | 21:36 |
| fsmithred | and again, you'll have to update and upgrade after installation to get all the latest. | 21:36 |
| gnarface | systemdlete: ah, my bad, i had misunderstood or forgotten the details of "fuzz" because i had mostly done this by observation with probably a lot less reading of man pages than i should have; the man page says it's "context" that the fuzz factor applies to, which are the optionally-included unchanged lines along side "lines to change" in the patch itself; what confused me is that usually in practice where i remember | 21:36 |
| gnarface | seeing these they were just whitespace | 21:36 |
| cousin_luigi | Let me rephrase that. | 21:37 |
| fsmithred | ok | 21:37 |
| cousin_luigi | godbed: fsmithred: Is Daedalus ever going to have /etc/debian_version updated to 12.8 from the current 12.7 ? | 21:37 |
| fsmithred | cousin_luigi, try the version of base-files that's in daedalus-proposed-updates | 21:39 |
| fsmithred | 12.4+deb12u8devuan1 | 21:39 |
| cousin_luigi | fsmithred: I see. In other words I only have to wait for base-files to pass muster. | 21:39 |
| fsmithred | we fork base-files, so yeah, we have to update it for the debian version. | 21:40 |
| fsmithred | hang on a sec. | 21:40 |
| gnarface | systemdlete: anyway, let us know if it ended up working | 21:42 |
| fsmithred | apt -t daedalus-proposed-updates install base-files | 21:42 |
| fsmithred | # cat /etc/debian_version | 21:42 |
| fsmithred | 12.8 | 21:42 |
| fsmithred | At some as yet undetermined time the devuan devs will agree to move proposed-updates into the main repo. | 21:44 |
| fsmithred | wheneverly | 21:44 |
| systemdlete | gnarface, thanks for your help. When I read the man page, things got clearer for me. And I don't expect you to be the knower-of-all-things unix/linux/debian or even devuan. | 21:45 |
| systemdlete | Now it gets through most of the build, but crashes at the end due to a missing file--though I wonder if it is just a renamed file. | 21:46 |
| systemdlete | tmp files | 21:46 |
| gnarface | systemdlete: well, i'm sure at one point i knew that detail but it slipped my mind because it's been a while since i had to actually fix a patch | 21:46 |
| gnarface | what's the file it's missing? | 21:47 |
| systemdlete | see https://paste.debian.net/1335348/ | 21:47 |
| gnarface | systemdlete: that's weird, digging for tmp files in the ./debian/ directory, i've never seen that | 21:47 |
| gnarface | i wonder if it's even something important or if you can just put an empty file there | 21:48 |
| systemdlete | there is, however, a file debian/tmp/usr/lib/tmpfiles.d/nut-common-tmpfiles.conf | 21:48 |
| systemdlete | not sure about that if this other file is the replacement | 21:48 |
| gnarface | ah, interesting... so there's a possibility a version difference in some tool changed to implicitly include the ".conf" part at some point after the dependencies for the last debian release for this were included in bookworm... | 21:49 |
| gnarface | (just a theory) | 21:49 |
| systemdlete | contents of this other file is https://paste.debian.net/1335351/ | 21:49 |
| gnarface | uh... looks important | 21:50 |
| systemdlete | my opinion also | 21:50 |
| gnarface | maybe check into the change to implicitly tack on ".conf" that might be documented somewhere... | 21:50 |
| gnarface | then you know you can just rename it | 21:51 |
| systemdlete | now I'm wondering if I had first stepped my process to build 2.8.1, then 2.8.2 if things might have gone smoother. Maybe those errors I ran into with the patches are not really so much errors as failure to go through the upgrades one by one | 21:51 |
| systemdlete | well, I can just try it. Doesn't matter, this is a dev/test attempt | 21:51 |
| systemdlete | gnarface, yes it does appear to be rename: https://github.com/networkupstools/nut/commit/72edde7c62bdf5e79d324eb043f8bad8612c4bdc | 22:14 |
| systemdlete | seems to be in the systemd path though. I guess our maintainers change that for us | 22:16 |
| systemdlete | but given it is a simple rename, it should be easy to create a workaround, as you said. | 22:17 |
| systemdlete | looks like this change was made almost 2 years ago, so that would be past 2.8.0, the current nut version in daedalus | 22:18 |
| systemdlete | (I think 2.8.0 is from about 2019 iirc) | 22:18 |
| gnarface | hmm, checking these nut package versions in the repo, at first glance the first few i checked don't appear to be devuan-forked | 22:21 |
| gnarface | so maybe part of the issue is that 2.8.1 and 2.8.2 mark a migration to systemd within the package | 22:22 |
| systemdlete | that would be just wunnerfull | 22:22 |
| * systemdlete spews hate on systemd | 22:23 | |
| gnarface | maybe one of our maintainers has seen enough such migrations that they can pitch in with some helpful insight though | 22:23 |
| systemdlete | here's an interesting data point to consider: I just installed nut 2.8.2 on a freebsd VM here. | 22:28 |
| systemdlete | afaik, systemd doesn't exist there, but maybe I am wrong. | 22:28 |
| systemdlete | or perhaps they modified the build for bsd's | 22:28 |
| gnarface | i think the idea of trying to use the automated tools to upgrade to 2.8.1 first then 2.8.2 is a worthy hypothesis to test | 22:29 |
| gnarface | i would not be surprised if they modified the build for BSD but gave the middle finger to systemd-free linux distros, but doesn't freebsd have some compatibility shim or something? | 22:29 |
| systemdlete | hmmm. Maybe I'll try that later today, but I need to go atm. | 22:30 |
| gnarface | i remember hearing there was talk about some bsd building a shim but i forget where that ended up | 22:30 |
| gnarface | anyone seeing some repos down? | 23:39 |
| gnarface | i seem to be having trouble getting a full "apt-get update" right now | 23:39 |
| gnarface | not sure which end it's at | 23:40 |
| gnarface | 147.78.194.22 didn't even respond to pings | 23:40 |
| gnarface | are they all expected to respond to pings? i know it isn't a requirement for http to work... | 23:41 |
| godbed | https://mishka.snork.ca/apt-panopticon/results/Report-web.html hehe, even this does not work here | 23:42 |
| godbed | this one does: http://veritas.devuan.org/apt-panopticon/results/Report-web.html | 23:42 |
| gnarface | oh wow, so like half of them are down | 23:43 |
| gnarface | i guess that checks out | 23:43 |
| gnarface | thanks | 23:43 |
| gnarface | i kept retrying, thinking that just one or two was down and i was having bad luck | 23:51 |
| gnarface | even the dns seems to be intermittently failing to resolve | 23:52 |
| gnarface | is this stuff under attack or is someone just doing a bunch of sweeping updates and configuration changes? | 23:52 |
| godbed | prolly them dang chinese haxors! | 23:53 |
| gnarface | even when i get a response from one of them, most the time it's transferring at somewhere less than 1b/s | 23:54 |
| gnarface | i finally managed to download what i was going for, but if someone doesn't already know what's up with this i think we'd better get someone looking into it... | 23:55 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!