blizzow | I have something a little weird going on in my fresh daedalus install on a thinkpad X1 Extreme 5th Gen. The USB-C ports don't work. | 02:10 |
---|---|---|
blizzow | They used to work in poopOS but seem to have stopped. I can't find any switch on the machine to enable or disable the USB-C ports. I haven't changed BIOS. | 02:11 |
blizzow | I see nothing in dmesg. | 02:11 |
rrq | what does "USB-C ports don't work" mean? | 02:14 |
ham5urg | I did an testing upgrade an fall into the same trap as many others. | 02:18 |
ham5urg | When I do a '/usr/lib/usrmerge/convert-usrmerge', I get an error like | 02:18 |
ham5urg | Both /lib/udev/rules.d/60-libsane1.rules and /usr/lib/udev/rules.d/60-libsane1.rules exist. | 02:18 |
ham5urg | Indeed, both directories have many identical files. | 02:19 |
ham5urg | How to process from here on? | 02:19 |
gnarface | blizzow: you sure it's not just a permissions issue? how'd you test? | 02:20 |
blizzow | gnarface, I have both usb a and c ports on the laptop. the a ports work fine. | 02:26 |
blizzow | The c ports, not so much. | 02:26 |
ham5urg | Should I remove all links from /lib/udev/rules.d/ ? | 02:26 |
ham5urg | And rerun /usr/lib/usrmerge/convert-usrmerge ? | 02:27 |
ham5urg | What a mess | 02:27 |
gnarface | ham5urg: stick around, i'm sure someone has advice | 02:28 |
ham5urg | Yes, I won't reboot soon :D | 02:29 |
ham5urg | Did an 'apt dist-upgrade' before knew about convert-usrmerge | 02:29 |
gnarface | blizzow: that didn't tell me how you tested it | 02:29 |
u-amarsh04 | I check the dates of the files concerned | 02:30 |
blizzow | I plugged a device (usb hd) into the c ports and it didn't work. I grabbed a different cable and plugged it into the a ports and it worked. | 02:30 |
blizzow | I grabbed an adapter/dock and plugged it into the c ports and did get this: | 02:31 |
blizzow | [ 6341.399472] ucsi_acpi USBC000:00: UCSI_GET_PDOS failed (-95) | 02:31 |
blizzow | The only thing I found was a message to a kernel mailing list saying a reboot while the usbc was plugged in seems to have it show up. I did not try that yet. | 02:32 |
blizzow | The other solution I saw was to use a different kernel. | 02:32 |
gnarface | blizzow: kernel or modules difference is all i can think of too | 02:33 |
gnarface | blizzow: looks like kernel 6.5 is in daedalus-backports, maybe try that | 02:34 |
blizzow | I might give that a shot. Or just wait out ceres until the usrmerge issues are fixed. | 02:34 |
rwp | ham5urg, I don't know anything about /usr/lib/usrmerge/convert-usrmerge but AFAIK those identical files are now declared to be accidentally forked and if they truly are identical then remove the /lib one. | 02:34 |
rwp | The plan would be to repeat that process until /lib and /bin are both completely empty. Then symlink them to the /usr versions. | 02:35 |
rwp | Caution for the main commands cp, mv, ln that you don't break them while the movement is in progress because of shared libraries. Use ldd $(command -v ln) to see what I am talking about. Those libs are in /lib and if they disappear then the command stops working. | 02:37 |
ham5urg | Easier to make a fresh install, isn't it? | 02:38 |
rwp | Note that klibc-utils with /usr/lib/klibc/bin/* includes most of the basic commands statically linked or use in initramfs images and that can be a good rescue if that bad case happens. | 02:38 |
u-amarsh04 | I had to use an Devuan live image on DVD to mount the hard disk and move files on the boot hard disk's /lib | 02:38 |
rwp | AFAIK there is no installer for Unstable/Testing and one usually installs Stable and then upgrades. But UsrMerge makes that a problem. | 02:39 |
u-amarsh04 | my older pc has been migrated and upgraded from a Debian install in June 2004, never re-installed since then | 02:39 |
ham5urg | u-amarsh04, did you documented your process? | 02:39 |
ham5urg | wau since 2004 | 02:39 |
rwp | Also there have been reports of Desktops Environments being broken on Testing due to UsrMerge right now too. | 02:39 |
u-amarsh04 | ham5urg, not on the /lib side, but on dev1galaxy.org under hardware and system configuration | 02:40 |
rwp | u-amarsh04 is the one who gave me the hint that /usr/lib/klibc/bin/* static binaries were available. That was a good hint! | 02:40 |
u-amarsh04 | be careful that the /usr/lib/klibc/bin/* binaries are more limited than the usual versions | 02:41 |
ham5urg | u-amarsh04, this https://dev1galaxy.org/viewtopic.php?id=6290 ? | 02:41 |
rwp | I think they previously used BusyBox but I guess they don't anymore. | 02:41 |
u-amarsh04 | ham5urg, yes | 02:42 |
rwp | Even after UsrMerge migration I would expect various breakage because for example cpio moved binaries from /bin to /usr/bin which is okay but then forgot to update the postinst and prerm scripts dealing with update-alternatives and are still using the old locations. | 02:43 |
u-amarsh04 | I stated in the thread mentioned above that package acl appeared to do the right thing as far as having symbolic links created | 02:44 |
ham5urg | Oh man, as I'm reading a question arises. Will this be the end for Devuan as a whole? | 02:45 |
onefang | That'll be "not a support question", best take it to #devuan-offtopic | 02:46 |
u-amarsh04 | I wish that the Debian people who decided on usrmerge tested everything themselves and documented a proper process for packages maintainers to avoid breakage | 02:46 |
golinux | ham5urg: I am going to wait until the dust settles before getting off of chimaera. It is the transitional period that is so fragile. | 03:12 |
fluffywolf | I've done two daedalus upgrades. neither went well, and tons of things ended up broken. lots of anti-features in new packages, too. | 03:24 |
fluffywolf | bookworm wasn't ready for release imho. | 03:24 |
debdog | my experience was: fresh installs work. some ubgrades went _really_ terribly. | 03:25 |
fluffywolf | daedalus is by far the worst debian/devuan upgrade experience I've had, and I've been using debian since... bo? I don't even remember now. | 03:28 |
fluffywolf | it was on my new shiny 486/33 with 4mb ram... | 03:28 |
Xenguy | golinux, Yeah, give me the trailing edge, or give me death | 03:33 |
Xenguy | The future with its usrmerge/systemd BS is just depressing | 03:33 |
fluffywolf | I upgraded a 16 year old debian install through to chimaera, through every release, with no real issues... but chimaera to daedalus fucking sucks. | 03:34 |
Xenguy | This June or so I'll be forced to upgrade to Chimaera | 03:34 |
Xenguy | Which is fine, except for wicd | 03:35 |
fluffywolf | upgrading to chimarea is pretty smooth, if I remember correctly. | 03:35 |
fluffywolf | wicd still works on chimaera. | 03:35 |
Xenguy | Ummm, hrm | 03:35 |
fluffywolf | it and everything it depends on has been removed from repos, but if installed it keeps working fine. | 03:36 |
Xenguy | Sheesh, looks like you're right, thank you | 03:36 |
Xenguy | https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=wicd&x=submit | 03:36 |
Xenguy | Why did I think that wicd disappeared after Beowulf ? | 03:37 |
Xenguy | That's great news for me, if so | 03:38 |
fluffywolf | hrmm, I thought it had been removed from the chimaera repos, but apparently I'm wrong. I know it's gone in daedalus. | 03:39 |
Xenguy | Yep | 03:39 |
fluffywolf | python is really harmful to the linux ecosystem, imho. | 03:39 |
u-amarsh04 | when upgrading some really old installation (pre-Devuan days) I found that I had to search for packages that had been removed but not purged and had files left behind: doing a search in aptitude for ~c found them | 03:39 |
Xenguy | wicd is a good example of that perhaps | 03:40 |
Xenguy | Software that has no staying power | 03:40 |
u-amarsh04 | perl upgrades can be "fun" also waiting for all the packages that are installed to catch up | 03:40 |
Xenguy | Good ol' perl | 03:41 |
Xenguy | Those were the days | 03:41 |
fluffywolf | u-amarsh04: but they keep working. unlike python, where they change the language and abi with every fucking single minor interpreter version bump, often in ways that prevent older software from continuing to work. | 03:41 |
fluffywolf | your software works with 3.7? too bad it won't run on 3.11! haha! fuck you! | 03:41 |
Xenguy | The disregard for backward compatibility seems cavalier, but there may be details I don't understand | 03:42 |
fluffywolf | you shouldn't be using older software. get with the program. | 03:42 |
fluffywolf | we tell you what software you should be using. | 03:43 |
Xenguy | Right, the new software that no longer exists because... | 03:43 |
Xenguy | Lather, rinse, repeat | 03:43 |
fluffywolf | poettering would fit in great with the python devs. | 03:43 |
Xenguy | ^^ How Rome must have felt when the 'barbarians' arrived... | 03:44 |
fluffywolf | debian packages usually depend on exact python interpreter versions, and of course you can't install multiple versions at once... | 03:45 |
fluffywolf | Package: borgbackup Depends: python3-pkg-resources, python3 (<< 3.12), python3 (>= 3.11~) | 03:46 |
fluffywolf | >11 and <12. | 03:46 |
fluffywolf | heh | 03:46 |
Xenguy | Weird | 03:46 |
fluffywolf | python causes 90% of packaging-related issues every time I upgrade anything, despite being used for such a small portion of software. | 03:46 |
fluffywolf | nothing weird about it... it won't work with 3.10 or 3.12. heh. | 03:47 |
fluffywolf | because python fucking breaks everything even with minor version bumps | 03:47 |
fluffywolf | in both directions. never forwards or backwards compatible. | 03:47 |
golinux | Xenguy: No wicd in chimaera . . . :( | 03:59 |
golinux | You can thank python3 | 04:00 |
Xenguy | That's what I thought golinux , but pkginfo reports otherwise, so not sure anymore | 04:00 |
golinux | That OK because I don't have wifi. It just connects. fsr helped me set it uplike that. | 04:02 |
rrq | Xenguy: afaict "wicd" in daedalus is a dummy package that pulls in some of the other guis to replace it | 04:23 |
Xenguy | rrq, pkginfo lists wicd in chimaera, which surprised me, as I thought beowulf was the last of wicd | 04:25 |
rrq | yes, there exists a "wicd" package, but it does not bring the old wicd software; it merely makes a dependency to bring in | 04:26 |
rrq | network-manager-gnome or network-manager or connman-gtk or cmst or cunnman-ui as substitute | 04:26 |
golinux | I'd forgotten about that . . . | 04:27 |
* rrq thinks it's like asking for potatoes and get cabbage | 04:28 | |
DelTomix | the cabbages they are on cabgmaster.devuan.org? :-) | 04:32 |
al1r4d | wicd.. ah old one =)) | 04:32 |
Xenguy | rrq, Aha, things are starting to make sense now, thank you | 04:40 |
Xenguy | For clarifying | 04:40 |
Xenguy | Did you hear that fluffywolf ? | 04:41 |
rwp | Speaking of upgrades... Is anyone successfully running irssi on Daedalus? It starts, starts to connect, but never actually successfully connects. irssi is a casualty of Daedalus for me. | 05:24 |
rwp | The last thing irssi reports is "[Libera] Waiting for CAP LS response..." which leads me to wonder if the problem is in the connection. But it works on Chimaera okay. | 05:25 |
rwp | I should build up a chimaera chroot so I can run the chimaera version on the exact same machine and verify that the only difference is chimaera-daedalus. | 05:26 |
brocashelm | not sure, but i've been able to get weechat to work | 05:29 |
brocashelm | i would then install the chimaera version and block the daedalus version if that's the case, or do a source installation from upstream | 05:31 |
rwp | I am debootstrap'ing a chimaera chroot on the laptop and will give that a test. | 05:49 |
rwp | I installed chimaera into a chroot, installed irssi, copied my ~/.irssi directory into it, and everything worked perfectly. A-B testing shows it works on Chimaera and (for me anyway) fails on Daedalus. | 05:58 |
rwp | I guess I do now have a way to run this, in the chroot. Like I am doing typing this right now. | 05:59 |
rwp | I guess the next thing is to see if the chimaera package installs on daedalus and try just the version change. | 05:59 |
rrq | rwp: I run irssi on daedalus, but with a znc proxy that goes to libera | 06:00 |
rwp | I guess I should also try it on a Debian Bookworm equivalent system and see if it fails there too. | 06:00 |
rrq | with=via | 06:00 |
rwp | I am usually (like now on my old laptop) running emacs-erc through znc through irssi so I could try that connection too. | 06:01 |
rrq | my daedalus is an upgrade from chimaera | 06:02 |
rwp | My Daedalus laptop is a pristine installation. | 06:03 |
rwp | I'll make a copy of the chimaera chroot and upgrade it to daedalus and see if irssi continues to work or not. | 06:03 |
rwp | My working chimaera upgraded to daedalus stops working. | 06:30 |
rwp | I had exactly the same result in upstream Debian for Bullseye (irssi working) and Bookworm (irssi not working). | 07:39 |
grayrider1 | QUESTION: ANYONE KNOW THE PASSWORDS FOR THE "TRIAL" RASPBERRY PI IMAGES? | 07:55 |
al1r4d | root:toor(?) | 08:14 |
grayrider1 | root toor doesn't work; devuan devuan not; pi raspberry not; blah blah | 08:29 |
grayrider1 | tried blank pwds, too ... no go | 08:29 |
cousin_luigi^ | grayrider1: Are you connecting via ssh? | 10:24 |
grml | With the current daedalus netinstall iso I get prompted for a missing regulatory.db. How do I get it? I have ethernet connected during install. | 11:35 |
rkta | grml was me, question remains. | 12:18 |
cousin_luigi | rkta: wireless-regdb ? | 12:26 |
rkta | cousin_luigi: The prompt mentions 'regulatory.db' | 12:27 |
cousin_luigi | rkta: What prompt? | 12:28 |
rkta | When installing, during configuring of the network, I get a prompt about that file missing. | 12:29 |
cousin_luigi | oh, uhmm | 12:30 |
cousin_luigi | Possibly not supplied by the iso or not installed in the correct order. Can you skip it? | 12:31 |
cousin_luigi | (since you use wired, but perhaps this is a bug worth reporting) | 12:31 |
rkta | I aborted the install and fixed my system in the mean time. Yes, was thinking about a bug. | 12:32 |
onefang | platform regulatory.0: firmware: failed to load regulatory.db (-2) | 12:33 |
onefang | platform regulatory.0: Direct firmware load for regulatory.db failed with error -2 | 12:33 |
onefang | cfg80211: failed to load regulatory.db | 12:33 |
onefang | That sort of "prompt"? | 12:34 |
rkta | No, in the TUI, I get a prompt, with buttons to abort and continue, about that missing file with the option to load it from some media. | 12:35 |
onefang | I get that every time I boot. I think that's the "This is what the WiFi laws in your country allow". WiFi works fine anyway. | 12:35 |
rkta | I'm talking about the installer. | 12:36 |
onefang | Did you try "continue"? | 12:36 |
rkta | Like I said, I managed to fix my system, hence I did not continue the install. | 12:37 |
rrq | rkta: did you take the prior "load components" step? | 12:38 |
rkta | rrq: I'd think yes, but I might be wrong... I can reboot later and try to reproduce. | 12:41 |
rrq | thanks | 12:42 |
rkta | rrq: You are right, I must have skipped this step. Sorry for the noise. | 12:51 |
rrq | thanks though :) | 12:51 |
rkta | btw, since when does one need a non-free-firmware repo? Wasn't firmware not included in non-free? Or did I mess up my sources.lst some time... | 12:56 |
brocashelm | rkta: for things like the intel, broadcom, realtek, etc. wifi drivers that are proprietary, as well as the intel-microcode, amd64-microcode, firmware-amd-graphics, etc. packages; debian moved those out of non-free, in case you'd need them | 12:57 |
brocashelm | as of bookworm/daedalus, you need to append "non-free-firmware" on your sources.list file | 12:58 |
brocashelm | example: deb http://deb.devuan.org/merged/ daedalus main contrib non-free non-free-firmware | 12:58 |
rkta | brocashelm: I found out after rebooting, when my system was borked. I guess I should have read the release/upgrade notes more carefully I guess. :D | 12:59 |
brocashelm | yeah, i can see the pros and cons of such a move; once that's out of the way, it's smooth sailing | 13:00 |
brocashelm | now, we are faced with excalibur (testing) and ceres (unstable) having the misfortunes of usrmerge changing where the kernel loads firmware from (i.e. instead of /lib, it's /usr/lib sans any symlink for backwards compatibility with legacy setups) | 13:01 |
rkta | I'm more and more disappointed with the linux world. | 13:04 |
djph | Debian's done things this way for 20 years ... | 13:12 |
djph | oh no nvm, misread, ignore me | 13:13 |
adam_free2air | haven't yet looked into this in depth, but perhaps overlayfs is worth looking into as a tool to help redress usrmerge issues? | 14:51 |
cousin_luigi | Joril: Pardon my stalking, but is it normal that with radicale, entries seem to go one way only? I can't push existing ones from my phone and I can't receive new ones from my pc. | 15:47 |
cousin_luigi | adam_free2air: Good grief, adding another layer of complexity to make thing simpler? | 15:47 |
cousin_luigi | things* | 15:47 |
Joril | cousin_luigi: Well that's strange, I push events from my phone and read them from PC all the time... I'm using CalDAV-sync on the phone and Thunderbird on the PC | 15:51 |
adam_free2air | cousin_luigi: there is rarely 'good' grief. we should think about dealing with this with a mininum amount of maintenance in the long term as well as a minimum amount of grief and complexity. other suggestions welcome. | 22:04 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!