libera/#devuan/ Tuesday, 2021-04-06

rwpOn the good side that seems to be a generic Linux kernel and driver problem and not something OS specific.00:00
fsmithreddebian wiki is a good place to check for wireless00:00
rwpfsmithred, https://paste.debian.net/plain/1192440 but I know you will ask what problems occurred due to this and literally it was several weeks now and I have forgotten the details.00:01
systemdletefsmithred: Looks like the firmware is from 2 years ago.00:01
systemdletemaybe there has been an update since then?00:01
systemdleteexperimental?00:02
fsmithreddid you have to install firmware for the wireless? There might be something newer in beowulf-backports. Also newer kernel. Someone mentioned 5.3 in that thread. I think bpo has 5.10 now.00:03
rwpsystemdlete, Which kernel are you running now? The very last message in that Mint forum suggested that it was fixed in the Linux 5.x series.  So maybe try a backport kernel?00:03
systemdleteI see there are bug reports (only a few) on the dev's site for the driver00:03
fsmithredrwp, oh THAT alternative.00:04
fsmithredyou upgraded to beowulf or to chimaera?00:04
systemdletestock kernel00:04
systemdleteinstalled beowulf, nothing else00:04
systemdlete4.19.0-1600:04
rwpYes.  A fresh pristine installation always works okay.  But I upgraded from Stretch to Beowulf that system and it blew chunks during the upgrade and needed me to kick it through to get to the end.00:04
fsmithredsorry, that was for rwp00:04
systemdleteoh00:05
* systemdlete is easily confused, if you cannot tell...00:05
fsmithredwhose desktop-base is installed? debian or devuan?00:05
rwpAnd at this moment it won't start X for the user.  It starts lightdm okay but at login on the graphical login it fails with an error.  I tried some things but need to dig deeper.00:05
fsmithredI think we forked xfce4-session00:06
fsmithredyeah: 4.12.1-6+devuan200:06
fsmithredand maybe some elogind stuff is wrong00:06
fsmithredTake a look at this. Ignore the mate packages: https://termbin.com/l68fe00:07
* systemdlete installs kernel from backports, crosses fingers...00:08
rwpThis system had Cinnamon installed and that would be preferred for this user but I would switch them to XFCE if that was needed.  This user is an 85 year old wife of my airplane mechanic just by way of environment.  She doesn't like changes.  I swapped laptops with her to keep her going.00:08
fsmithredbpo kernel works00:08
fsmithredso why not keep cinnamon?00:08
rwpI need to go back in and dig through why it was having upgrade problems.  But it error'd repeatedly on packages in that topic area (sorry, I've slept since then and forgotten details) and I had to manually walk it forward.00:09
rwpBut yes Cinnamon would be preferred for this user since it keeps things closer to what she is used to using.00:10
fsmithrednp. If it's not right in front of my, it doesn't exist.00:10
fsmithredand you have to upgrade it, not reinstall?00:10
rwpI can do a re-install fresh on this box.  But upgrades *should* work too.00:11
systemdleterwp:  I just had a similar experience with ascii->beowulf.  It was hopeless after reboot, so I said u-know-what-with-it installed beowulf.00:11
fsmithredyes, they should00:11
rwpMeanwhile I have migrated quite a few server systems from Stretch to Beowulf without any DE installed and all of those have gone great.00:12
systemdleteIt was easier for me to install beowulf from scratch than try to figure out what broke during the upgrade.00:12
systemdleterwp:  I guess ymmv00:12
rwpsystemdlete, Understood.  But you know I will try to fix things when I can help fix things and it is something that really _should_ be working.00:12
systemdletewell, fsmithred and I tried many different things to recover from that disaster, but ultimately I threw my hands up00:13
rwpAnd I hated to bring up that it was problematic if I hadn't done the work to see why it was problematic.  I only got as far as hitting those alternatives links and ran out of debug time on it.00:13
fsmithredmake sure you have devuan's lightdm installed00:13
systemdleterwp:  I agree.  And, generally, I DO make some attempt to figure out why things are not broken, sometimes moreso than other times.00:14
systemdleteIn this case, I just want my laptop back!  LOL00:14
systemdletegnarface, otoh, calls this sort of rundown "for science" only00:14
fsmithredone easy solution might be to remove desktop-base00:14
rwpI can see that I had lightdm installed but currently have slim installed as a test.00:15
fsmithredyou can set themes manually00:15
* systemdlete will be quiet now and let you guys tackle your problem...00:15
fsmithredslim doesn't do session management well00:15
fsmithredyou might have consolekit and elogind stuff mixed up, too00:15
fsmithredor a part missing00:15
rwpI have a longer history with lightdm but I could test other things in order to try to debug it.00:15
rwpRight.  Don't stress about it at this moment.  Other than just another vote that something was wonky in that general area.  I can dig into it.  And will do so when I get the time for it.00:16
rwpI dont' have the time right now.  And I mitigated the problem for me by swapping one laptop for a different one that had a pristine installation on it.  I have the original problem here available for me to wack on when I can do so.00:17
fsmithredsee the lightdm stuff in Release Notes00:18
fsmithredcheck /etc/pam.d/lightdm-greeter00:18
fsmithredsession   optional pam_elogind.so00:19
rwpWhat do you suggest for gnome-themes-extra ?  It is not currently installed.00:19
rwpWhat should I check for in /etc/pam.d/lightdm-greeter ?? It loads "session optional pam_systemd.so" which I usually remove to avoid the errors logged by it not being present.00:20
fsmithredsession   optional pam_elogind.so00:21
fsmithredadd that line00:21
fsmithredor make sure you have the devuan version of lightdm00:21
fsmithredthe alternatives links you posted look correct00:22
fsmithredkitchen time00:24
fsmithredbbl00:24
crashoverrideI am beginning to be allergic to computers, I think00:25
zeroability[m]Is there a line forming for that?00:26
rwpfsmithred, More details here https://paste.debian.net/plain/119244100:27
rwpfsmithred, Again, let me dig into things a little deeper.  I just don't have the time at the moment but should be able to get something dug out of it.00:27
systemdletefsmithred:  So I installed and booted the 5.x kernel from backports.  Now, it is a little bit more reliable maybe.  I can reach the laptop but only intermittently.00:39
systemdleteThis may work well enough for my purposes for now.  I just need to be able to do a backup once a day; the backup tool (bareos) does retries every few hours so the backup will eventually complete... assuming that it succeeds at least once in 7 tries.00:40
systemdleteAnd even if the backup fails on one particular day, it may not matter much... I am only using this laptop as a tablet for reading.00:41
systemdleteThe newer kernel does seem to make a difference, even if it is not consistent.00:41
systemdleteI'll continue working with the openwrt folks too to see if they have any ideas.00:42
rwpsystemdlete, If the problem is the WiFi hardware + driver then a different workaround might be to use a USB WiFi dongle.00:43
rwpI have a small'ish button one which works well and does not stick out very much.00:43
rwpI consider that a workaround but...  It can work very well.00:44
systemdleteYeah, I have a G here that works.  But that means some degradation in peroformance of course00:44
rwpAnother thing I have done with the old Broadcom chips that had no free drivers was to swap out the internal WiFi card.  When the BIOS firmware allowed that and did not complain!00:44
rwpI hate it that some vendors lock the firmware to halt if we decide to change out the WiFi card.  It should not care!00:44
systemdlete"swap out?"  This is a SoC; not sure I can do that.00:44
systemdleterwp:  Sorry.  Should have mentioned.  This is a 2-in-1 tablet, not a real laptop00:45
rwpIn that case you can't.  But often the WiFi is on a mini-PCIe slot internally.  That's the way most of my laptops are configured.00:45
rwpOh!  Yes.  Sorry.00:45
systemdleteThere does not seem to be any public entrance to the internal workings of this POS00:45
systemdleteLet's say I buy a tiny dongle (that would be sweet).  Is there a way to disable the internal qualcomm chip?00:48
systemdlete(I mean aside from removing realtek fw which might be needed for the new device)00:48
fsmithredyou don't need to disable it. Just configure stuff for wlan1 or whatever the new interface is called01:06
systemdletefsmithred:  I would want the default route to go through the new dongle.01:20
systemdlete?01:20
fsmithredhow are you configuring your interface? wicd? config file?01:21
fsmithredeither way, bring up the one you want to use and it will be used01:22
systemdleteWhen I said "disable" I just meant to ensure that it doesn't get used.  I've been using wicd which has an icon I can easily view to determine its state.  I would actually prefer to use the interfaces file, but I'd need a substitute for the display.01:45
fsmithredset the default wireless interface in wicd preferences or in the manager-settings config file in /etc/wicd01:48
fsmithredor in /etc/network/interfaces01:49
xrogaanquestion about runit: can it support user daemon?02:06
xrogaanso far I know, sysv doesn't and it's been an annoyance of mine.02:06
fsmithredwhat is user daemon?02:06
xrogaanright, so, software that runs in the background but get either killed when the user logoff, or don't get restarted needlessly.02:07
xrogaanFor instance, my mpd get started every time I login through slim, but it never gets stopped.02:08
fsmithredyou would probably have to write your own runscripts02:09
gnarfacehmm, doesn't lightdm do that automatically if you tell it to save your session?02:09
gnarfacei was pretty sure there was a way to get it to remember what was running and then run that again the next time you logged in02:10
masonxrogaan: man 5 crontab and look at "@reboot"02:10
gnarfacei remember because i had a problem with it doing that to something i didn't like02:10
masongnarface: session managers do that02:10
xrogaangnarface: that's not strickly what I need. xfce can do that too.02:10
fsmithredhttps://salsa.debian.org/runit-team/runscript-collection02:11
xrogaanmason: sure, but how do you stop the thing once the user logs off?02:14
fsmithredput it in the desktop's startup apps02:16
masonxrogaan: Ah, that sounds like user services. What's a practical example?02:17
adhocthere used to be a dot file that bash (or maybe another shell) would run when the user logs out02:20
adhoclike .profile or .bashrc02:20
adhocnot sure that is useful if you open up random terminals regularly02:20
gnarfaceit's .logout, and it still works in bash02:20
gnarfaceyou could use that to kill off stuff you'd backgrounded if you stored the PID somewhere02:21
gnarfaceit wouldn't be a super difficult scripting process to do so02:21
gnarfacebut i would assum ethe built-in xfce feature has it covered for most use cases02:21
adhocright, used that alot to clean up processes after your dialup link had dropped,02:21
adhocnot so sure that would be handy in X land...02:22
gnarfaceoh, it's possible that some session managers might not call it properly.  i did noticed lately lightdm doesn't call .bash_profile properly either02:22
gnarfaceor .login or whatever02:22
gnarfacethat's a new thing; previously people coding these things cared about preserving the shell environment's behavior02:23
gnarfacenow they don't even seem to remember it's there02:23
adhocyes02:24
adhocthey use systemd for that now02:24
* sadsnork should have seen that coming.02:24
adhocif shell scripting is too hard, then what chance are they going to have of maintaining a sensible shell environment02:25
adhoceverything is in .ini type files02:25
* adhoc is going through that hell in a work project02:25
xrogaanmason: yeah, user service.02:44
xrogaanalso, yeah, it's quite difficult to change the PATH of the xfce session02:49
xrogaanThere is no real practical example per say. I'd just like for the system to not start new processes or stop processes running as I create a new session.02:51
xrogaanI usually have to kill pulseaudio because it doesn't end itself, since it's not depending on the graphical session.02:52
adhocis that a background service?02:54
adhocif not, why not ?02:54
adhocI've had plenty of issues with pulseaudio02:54
xrogaanI'm looking at https://github.com/rootkiwi/an2linuxserver03:01
xrogaanIt could run as a user service. But then I'll need something to take care of its state.03:02
orcus-delightdm no sourcing .profile /etc/profile and friends is related to the different defaults they are using - ubuntu has different ones + it does deliver some missing support files ...  it's the same reason why ~/bin and ~/.local/bin do not get added to the PATH anymore  (at least at debian - didn't check for it at devuan) ...03:02
orcus-dethey == debian03:03
gnarfaceyea but the lines to add ~/bin to the user's path are still commented-out in .bash_profile... which doesn't run if you're using lightdm03:06
orcus-dethe xsession doesn't care about a bash profile03:06
gnarfacethe latter is an example of differing defaults, but the former breaks expected behavior in a bad way and the only reason they have for not doing it that way was "woops, didn't know it was supposed to do that.  oh well, i'm not doing it over.  suffer."03:06
orcus-dehttps://www.orcus.de/main_workarounds/lmde3_fix_profile.htm03:07
orcus-dedid never ever test it with devuan ... I'm just hanging around here most of the time ...03:07
gnarfaceand then in the mean time they bold on all the same functionality over03:07
orcus-deI gave up a bit on reasoning with people that it's a shame to break functionality out of nowhere.. for almost zero reason03:09
gnarfacewell, there wasn't zero reason, the reason was they didn't even know the work was already done before they were halfway to reinventing the wheel03:11
gnarfacebut that's almost defensible... not backpedaling to restore the break though, that's just pure ugly pride and nothing more03:11
gnarfacei dunno, it's not even software i actually use myself, but this type of thing causes me support issues03:12
gnarfaceand when it's support issues with something that has worked since debian sarge, i think the "woops didn't know the work was done already!" defense is pretty thin03:13
orcus-deI asked mint-devs (done all the work to debug it and finding the reason being different compile time defaults ++ grabbing the required files) to add it again to lmde (the mint debian version) ... which ended in NOPE ... as "we are just doing it like debian does" ... with all the nasty side-effects (but ok lmde is more the fallback for mint in case canonical would get worse than it already is) ... at that point you can either give up and turn away from03:18
orcus-dea distro or try your best to keep people going in some way03:18
xrogaanforgot how I got xfce to look into ~/bin03:19
xrogaanprobably ~/.xsessionrc03:19
systemdlete2I'm using wicd, and it was working fine before about an hour or two ago.  All I did was change the ip address, which is set to be a fixed address on the laptop.  It is correctly assigning the address and I can see, briefly, when the route is set.06:43
systemdlete2But the connection does not last.  When it tries to connect, I see a message in the wicd window saying "openwrt verifying access point association," then finally it says "openwrt could not contact the wireless access point"06:44
systemdlete2This was not the behavior previously.06:44
systemdlete2(nothing but headaches with wireless lately...)06:44
adhocare you using dhcp ?06:44
adhocif so, is the dhcpd putting an IP for the host or are you setting the IP in wicd ?06:47
systemdlete2wicd is setting the IP address06:47
systemdlete2wicd is set to static address06:48
systemdlete2there is NO entry whatsoever in /etc/network/interfaces for the wireless06:48
systemdlete2wpa_supplicant IS running.06:48
systemdlete2(just telling you my environment; please excuse my tone)06:48
systemdlete2adhoc:  Do you know what those 2 messages mean specifically?06:49
systemdlete2I wish wicd had an option to show a trace of what it is doing/thinking...06:49
systemdlete2sorry, missed your 2nd question.  The dhcp server DOES have an entry for the laptop, but as I said, wicd is configured to set the address the same.06:56
systemdlete2adhoc: I also tried rebooting.  In fact, I even did a COLD boot.  Neither helped.06:56
adhocwhat is the name of your access points SSID ?07:06
adhocSome times when you set a static IP it ignores the DHCP data entirely07:06
adhocI had the problem in Debian 8 and 907:06
adhocIf you put an entry in your /etc/dhcp/dhcpd.conf with the MAC address for you laptop, you would have a better chance of debugging this.07:07
adhocIt is possible (and I'm speculating) that the DHCP client is not getting a new lease and thefore the interface times out and is dropped.07:08
adhocI'm not sure what the DHCP server is, or its behaviour.07:08
adhocI hit something fairly similar last week setting up some new PXE booted machines.07:09
adhocWhen I hard set the IPs, the DHCP leases expired, as you need the lease in the first place to get the boot image, system config and kernel image.07:10
adhocAssuming you can edit the config for the DHCP server, I would set as little as you needed to for the client.07:10
adhocthen add more config is as you need it.07:10
systemdlete2These are not PXE clients.07:15
systemdlete2They have their own /boot with their own images and initrd's07:15
systemdlete2The server uses dnsmasq, but again, it is not serving dhcp in this case.  wicd is set to ignore dhcp with a static address.07:17
adhocok, similar problems may exist.07:17
systemdlete2How would the SSID impact any of this?07:17
adhocwhat is the "openwrt" in relation to the message you posted above?07:18
systemdlete2It's the software running on the router.07:18
systemdlete2just fyi, my android phone connects to the router just fine.  So it is not a router misconfig.07:19
adhocok, step back, are you running DHCP at all on the router?07:19
systemdlete2Also, the laptop WAS working a few hours ago, connecting to the router without a problem.07:19
systemdlete2not "dhcp" per se, but dnsmasq, which includes a dns and dhcp server combined.07:19
adhocright07:19
systemdlete2it is functionally equivalent, by and large.07:19
adhoccan you look at its logs ?07:20
systemdlete2I have.  No hints, sadly...07:20
systemdlete2Right now I am recharging the laptop.  Its battery only lasts a few hours.07:20
systemdlete2I can probably run the dnsmasq from the router console and get more info that way07:21
systemdlete2I've done that in the past.  But that won't tell me what is going on on the laptop.07:21
systemdlete2adhoc: One thing I definitely can report is07:22
systemdlete2that the wireless "chip" on the laptop is flakey.07:22
adhocis there a reason you need to run the fixed IP on the laptop on the network?07:22
systemdlete2why?07:22
adhocwhy not let the router hand out the IP, DNS, Route, etc?07:23
systemdlete2yes.  In order to do backups, which are run from a lan-side system.07:23
systemdlete2For the phone, it does do exactly that.07:23
adhoccan you IP for the specific MAC on the router?07:23
systemdlete2There is a mac filter for each of the devices.07:24
systemdlete2I've tried disabling it, but no difference.07:24
systemdlete2and, in fact, I can see the laptop connecting on the router... for about 3 seconds.07:24
systemdlete2If it were a problem with the filter, I would not see the laptop connection from the router, even for a moment.07:25
systemdlete2I don't think it is dnsmasq or the filter.07:25
systemdlete2Again, this was all working just hours ago.07:25
systemdlete2Something else is going wrong.  And this is a brand new devuan beowulf install on the laptop.07:26
adhocyou mention the "chip" us flakey?07:26
adhocdo you have any tools to look at the available SSIDs on the laptop and look at the relative signal strenghts ?07:26
systemdlete2Well, let me tell you what was going on, when I had it connecting at all...07:26
systemdlete2the strength is good07:26
adhocuntil it isn't ?07:26
systemdlete2usually, the router shows up in wicd as the 1st 2nd or 3rd07:27
systemdlete2well, let me explain07:27
adhocwhich chipset are you running ?07:27
systemdlete2I was able to connect from the laptop to the router, but trying to, say, ping the laptop was intermittent.  Sometimes it worked, sometimes not.  Even without changing a thing in the laptop or router config.07:27
gnarfacea lot of this stuff seems conflicting and similar to other problems but not quite07:28
gnarfacehard to say what is missing07:28
systemdlete2qca937707:28
adhocgnarface: indeed.07:28
systemdlete2hi gnarface.  how ar you doing?07:28
systemdlete2*are07:28
gnarfaceit almost sounded like an issue with forgetting you'd enabled MAC address randomization on the client and MAC whitelisting on the router at the same time07:28
gnarfacebut after that last line it sounds more like there are either two devices sharing one IP or two routers sharing one SSID07:29
systemdlete2no.  I haven't done anything to randomize the mac address... unless that is a default.07:29
gnarfaceis it a pinephone?  some of the older pinephones had a bug briefly where MAC address randomization got enabled by default and was not possible to disable07:29
systemdlete2There is only one router configured for wireless.07:29
adhochave not run into the qca9377 before.07:29
gnarfacethough they fixed that on the pinephones shortly afterwards07:29
systemdlete2it is an alcatel running an old version of android (about 6 or so)07:30
systemdlete2but the phone doesn't have this issue.07:30
adhocthe client is an alcatel ?07:30
systemdlete2it gets a connection to the router and continues smoothly.07:30
systemdlete2the phone is an alcatel, which is the device that is working with the router07:30
adhocok, so what is the client device?07:31
systemdlete2it's the laptop that is having problems.07:31
systemdlete2with the same router.07:31
systemdlete2It's flakey.07:31
adhocis the laptop running devuan ?07:32
systemdlete2I even ran tcpdump on both sides of the connection earlier.  A few hours ago I was able to get a solid, steady connection.  I could send packets from the laptop to the Internet (ping, etc), but I could not receive packets.07:32
systemdlete2adhoc: Of course.  I would not be asking here if it weren't.07:32
systemdlete2but once in a while, some packets did make it through to the laptop from the openwrt side07:33
systemdlete2weird.07:33
systemdlete2I'd run a test, take a breath for 30 seconds maybe, or a minute, whatever.07:33
systemdlete2Try again, and it would change behavior.  One moment working, rest, then not, rest etc07:33
adhocwhat is the wifi device on the laptop; internal or USB ?07:34
systemdlete2it's internal.  The laptop is really a 2-in-1 SoC.  It has a separable keyboard07:35
systemdlete2again, wifi device is a qca937707:35
adhocso its an atheros 10k ?07:36
systemdlete2This is one of the issues also.  I'm looking for a dongle I can use to compare.07:36
systemdlete2right07:36
systemdlete2ath10k07:36
systemdlete2I have the fw for it installed, too.07:36
systemdlete2hmmm.  I'm thinking maybe I should run memtest86 on it and see if it is having intermittent memory issues.07:37
systemdlete2At least I could eliminate that as a problem.07:37
systemdlete2anyway, the laptop is almost juiced out now.  So I will return to this again later.07:37
systemdlete2thanks for your help.07:38
gateway2000I'm getting some tofu (like this  â€œ) on the desktop for special characters like ",", etc, but not in my browser. I did the live install and selected en_US.utf8 or whatever it is as the locale10:21
gateway2000This is with the xfce default desktop for beowulf. Anybody have a clue how to straighten this out?10:22
gnarfacegateway2000: you just need to install some fonts that contain those glyphs10:43
gnarfacefirefox hides the problem by bundling its own now i think10:43
gateway2000Idk because the characters in question are " and - , but only in some situations10:43
gateway2000I think maybe it's a locale thing but idk much about it10:44
gnarfaceif it's just in stuff you're typing then it could be your keyboard mapping too10:44
gateway2000I saw the behavior in the display module of the xfce settings application and in a message in hexchat10:45
gnarfaceprobably the font then10:45
gateway2000Thanks, I'll see what I can see...10:45
gnarfacemake sure you did actually use the right locale10:46
gnarfaceen_US and en_GB are very similar but differ on characters like that10:46
onefangCould be the web site using those characters downloads it's own (or a Google font) font.  That's common.10:46
gnarfaceyea it could be sideloading fonts through a number of unsafe methods but you can verify cases like that by trying the characters in the URL bar10:47
gnarfaceif you get color emoji's in the URL bar it's using it's built-in glyphs10:47
gnarfaceits*10:48
krzychhi, how to mount lvm wolumen whith rw, not read-only in rescue mode10:48
gnarfacei don't use lvm or i'd tell you sorry10:48
gnarfacegateway2000: try this to get a partial listing of available font packages: apt-cache search ^ttf\-\|^fonts\-10:49
krzychgnarface: ok, maybe someone uses and knows :)10:51
gnarfacekrzych: hang out, it's a slow channel but someone knows10:51
gateway2000I tried rebooting into the live usb session and the issue did not persist. I know it's not an issue of what fonts are on the system. I'm thinking maybe somewhere along the line using the refracta installer something's dropped11:24
gateway2000locale output is the same for live and for installed session. I also tried regenerating the locales on the installed session just in case11:24
gateway2000If I can't figure anything out I'm going to try reinstalling from the full desktop iso11:25
systemdleteDoes anyone here know of a reason why devuan might not respond to an ARP request over wireless?   I've narrowed down the problem to this and it is reproducible, but intermittent.11:38
crashoverrideI'm pretty sure this is firmware related, not OS related.11:38
systemdleteIf I start a connection over the wireless link between devuan beowulf and my openwrt router, about 75-90% of the time roughtly it fails.11:38
crashoverrideI'd say "swap the WLAN adapter" with another laptop.11:39
systemdleteI agree crashoverride.  But I thought I should check first to see if anyone knows of a previously reported issue.11:39
systemdleteIt's a SoC and it's a tablet.  No servicable parts here.11:39
crashoverrideif problem happens with the other laptop too (has to be a non-windows driver) and does not happen with your current's, you know.11:39
crashoverrideah well11:39
crashoverridesorry then11:39
systemdleteNo, that's good really.11:40
crashoverridehow so?11:40
systemdleteI strongly suspect it is the firmware also11:40
crashoverridethen put android on it11:40
crashoverridesee if it has a problem11:40
crashoverrideif not11:40
crashoverrideextract the blob used as a firmware for it11:40
systemdleteYou know I have actually thought of just that...11:40
crashoverridethen put debian back11:41
crashoverridedevuan sorry11:41
systemdletewell, I have multiple partitions I can use11:41
crashoverrideand use that firmware to see11:41
crashoverrideeven better11:41
systemdleteso you are saying to extract android's qca9377 blob?11:41
systemdleteor something else?11:41
crashoverrideextract all the blobs11:42
crashoverridejust in case11:42
systemdleteSo, it sounds like your theory rests on the notion that android's collection of proprietary blobs are different from linux's11:42
systemdletebut I thought that android was a linux variant?11:43
systemdleteshared code, something like that.  I haven't followed that story too closely.11:43
crashoverridesystemdlete: android is a COMMERCIALLY supported linux variant.11:44
crashoverrideand linux is a KERNEL.11:44
systemdleteInteresting idea though.  But frankly, I'd like to see what a different wireless dongle might do11:44
crashoverrideit would be good for people to finally start understanding the meaning of those very terms.11:44
systemdletesorry, you are 100% right11:44
systemdletemy bad.  very bad.11:44
crashoverridenot your fault tho11:44
systemdletelet me rephrase:  I thought that android used a modified linux kernel and its drivers.11:45
systemdlete(I know everyone does that, don't they?)11:45
crashoverrideit uses a patched kernel11:45
systemdletepatched ~ modified11:45
crashoverrideand the drivers are usually only available to google, usually as blobs only.11:45
crashoverrideso...11:45
systemdlete"to google" you mean chrome os?11:46
crashoverrideno, android releases11:46
crashoverrideeven tho, that's valid for the pixel only11:46
systemdleteoh, ok.  I see now.11:46
crashoverridefor other brands, it goes straight to $vender.11:46
crashoverride$vendor*11:46
crashoverridedon't quote me on that, I'm no AOSP expert.11:47
crashoverridebut anyway, getting the blobs out of a supported android might be a good way to start.11:47
systemdleteI think my next move is to get myself another wireless dongle that works on linux., hopefully a tiny one that doesn't actually dongle11:48
systemdleteDon't I need to find the blob for the qca9377 specifically though?11:48
systemdleteIt's a qualcomm atheros.  Are blobs for those available to android?  idk.11:49
crashoverrideYou can surely find a xda-developers.com link for that11:50
systemdleteyeah.11:50
systemdletetbh, I've never been 100% content with this tablet.11:51
systemdleteI bought it thinking the reviews got praise, so maybe it is decent.  Of course, first I had to remove the malware it comes with.11:51
systemdleteOh sorry.11:51
systemdleteNot "malware" -- I meant the OS that was on it.  sorry.11:51
crashoverrideit was clear the first time.11:52
crashoverrideyou're not the only one dealing with those kind of problems11:52
systemdleteOh I am sure others have struggles too, esp with so many proprietary impediments out there11:52
crashoverrideand honestly, going by reviews when buying a phone/tablet as a computer enthusiast (or worse, IT professional) is like checking the call of duty chat before buying a firearm.11:55
* ShorTie snickers11:56
systemdleteWell, I do more than that.  I also check which chipset it has and other technical specs on it, bands, features.11:57
Wonkacheck lineageos wiki...11:58
Wonkamaybe even checkout the git repo containing the data files for said wiki, grep through them for features you want...11:58
systemdleteAnd when I do check reviews, I look at the worst reviews first.  If there are like 50 people who complain about, say, a disk drive burning out the first week, I'll avoid that product.11:59
systemdletethanks Wonka, will do.11:59
Wonkamy next one might be either a FairPhone 3+ or some OnePlus device11:59
crashoverrideI honestly am at loss when considering what to get next.12:00
crashoverrideeverything is terrible.12:00
Wonkahope there'll be a FairPhone 4 soon, featuring 5G12:00
Wonkayep12:00
crashoverrideyeah, fairphone 4, now with unibody. But hey, you can easily access the battery. It's glued, but you can look at it!12:00
djphugh12:01
crashoverridedjph: I mean...12:01
systemdleteThe worst part for me is the dread of calling upon the tech support, which is always in some dark cavern somewhere on the other side of the planet where people do not understand basic things like "I did not check the box.  In fact, I never saw it."12:01
crashoverridesystemdlete: what box?12:02
djphwhy... WHY ... does RAID-5 have to mock me?12:02
WonkaI have had really bad luck with four of my last five Android devices. Optimus 2X - Nvidia Tegra chipset, closed source. Nexus 7 (2012) - that flash problem. Asus Zenfone - x86, support got cancelled. bq Aquaris X2 Pro - first Android One to receive support for only two years instead of usually guaranteed 3 years.12:02
systemdleteOr, how about this one:  "I gave you my credit card to PURCHASE the item or service.  THat was NOT permission to bill me for other things."12:02
crashoverridedjph: because RAID60 told it to laugh in your face12:02
djphfair enough12:02
djph... but now I gotta go find something big to backup stuff to12:03
Wonkacrashoverride: fairphone with glued battery will not happen...12:03
djphand burn it to the ground and start over with big drives12:03
crashoverrideWonka: yeah, and before that we said: fairphone with mainboard and screen in one part will not happen12:03
crashoverridedjph: solid plan12:03
djphcrashoverride: yeah, I'm just more disappoint that "oh, you don't have to worry about this with LVM" was not accurate12:04
crashoverrideWonka: sorry, my bad, those are the EXACT two parts of the FP3.12:04
crashoverridenothing more, but still, you DO have a separate "everything" and a screen.12:04
djphfurther reading / digging was "oh, you don't have to worry about it when going to N to N+M drives"12:05
crashoverridedjph: when people say "you don't have to worry", it almost always mean "I got lucky and I didn't realize"12:06
crashoverrideyou can even find gifs of people doing things incredibly stupid and lethal, being miraculously spared; and going at it again believing they can't die.12:07
crashoverridebut due to the nature of that bias, I don't think there's enough scientific evidence of people who have it and stay alive for it to ever be named.12:08
Wonkaon https://shop.fairphone.com/en/spare-parts?phone_type=4&ref=header, I see "Display", "top module", "bottom module", "speaker module", "battery", "back cover", "camera". That's seven parts.12:09
crashoverridebattery isn't really a part.12:09
crashoverridesame for "back cover"12:10
crashoverridenow, lemme have a look at the rest, because I'm pretty positive "top" and "bottom" modules are ridiculous 10g modules with barely anything in them12:10
crashoverrideand "speaker" and "camera" are actually removable on most phones.12:11
crashoverrideso yeah, don't even need to look at the link12:11
crashoverridehttps://upload.wikimedia.org/wikipedia/commons/c/c4/Fairphone_3_modules_on_display.jpg12:11
crashoverridehere it is.12:11
systemdleteIt's a good thing that every model of phone of every different manufacturer have not tried to standardize their battery sizes.  It's more fun this way, and you can't try to cheat them out of THEIR profits by using the battery from an old phone in a different phone.12:11
djphcrashoverride: yeah, I went and dug into the documentation, and it's vague as to whether you can just extend the RAID virtual drive or not12:11
crashoverrideyou can CLEARLY see that 90% easily of the bulk of the phone is copmletely sealed on the mainboard on the right12:12
djphseems it might be related to the specific firmware, MAYBE12:12
systemdleteThis way, you have to hunt for the exact battery model you need.  Good luck finding it.12:12
systemdletebut this really should be in O.T., guys...12:12
crashoverridethe entire content of the chan for most of today should have been in #devuan-offtopic12:13
systemdletewell, maybe.  But I did come here with a rather devuan-specific question myself.  Thereafter, my convo wandered...12:13
djphoh well, rebuilding the array gives me reason to just install beowulf on said machine12:13
systemdleteWe'd better run to #devuan-offtopic before the cop arrives12:14
crashoverridedjph: don't derp like me and install an x86 system on an x86_64 machine...12:14
djphcrashoverride: I'm good there.  I don't even download x86 images :)12:14
crashoverrideWonka: yeah so after verification, the FP3 modules are literally cameras and microphones only.12:15
crashoverrideit's like saying your computer is modular because you can unplug your keyboard, mouse and webcam12:16
* systemdlete is rolling on the floor... "derp" I love it!12:16
crashoverridedjph: I had to deploy devuan on an old computer, and I never re-setup my installation medium.12:16
djphoops :)12:17
* systemdlete is too easy to amuse, really.12:17
crashoverridedjph: and worst part, up to yesterday, I actually had 2.6GiB of total ram.12:17
crashoverridedjph: on a 16GiB system.12:17
djphhah12:17
crashoverrideyeah12:17
crashoverridedidn't know it was x86 y'see?12:17
crashoverrideI was just assuming "modern software" was getting all cosy and dandy with all this ram12:18
crashoverrideand windowsed its way into most of it because "DeVeLoPeR tImE iS mOrE eXpEnSiVe ThAn RAM!!!!1111eleven"12:19
djphyup12:20
systemdleteBeD TiMe  BbL12:21
Wonkacrashoverride: that's still way more than other phones.12:22
crashoverridesystemdlete: no need to take that tone bro :P12:22
crashoverrideWonka: ok, that's fair.12:22
crashoverrideWonka: kinda like saying "this anti-vaxxer got sick, so in terms of antibodies, that's still way more than the other ones"12:23
* systemdlete looks at his comments to examine tone. None found. Rolls over again and back to sleep...12:23
crashoverridebut aside from that, it also is true12:23
x_waiting... on boot (Confuguring network interfaces ... ifup... waiting...)16:26
masonx_: Is that a recently converted system?16:27
x_devuan 3.1.0 + some mx packages16:28
masonx_: If so, that's around where I do this in my migrations: apt --purge install eudev systemd- libnss-systemd- elogind-16:28
x_and autologin doesnt work16:29
x_thank you16:31
crashoverridemason: couldn't someone write a script to convert debian to devuan?16:49
masoncrashoverride: The trouble is that the process includes (for me) a couple reboots. This can be scripted with various bits being handed off, but the process ought to be simple enough to have manual application actually be simpler.16:50
masoncrashoverride: Freshly re-annotated: https://bpa.st/4JVQ16:54
crashoverridemason: technically, an RC script can make it automatic even with reboots.16:54
masoncrashoverride: Right. But the actual steps are minimal enough that hiding them away might not be a service to the user.16:54
crashoverrideRC services can display text.16:55
crashoverridemason: is there a version that I can read the link without a multiple-MiB SGML parser?17:02
masoncrashoverride: Other than the contents of sources.list, that's all stuff I paste into my root shell. I see no SGML here. And it's 31 lines including whitespace and comments.17:05
crashoverridemason: it is literally 122 lines.17:05
masoncrashoverride: bpaste renders here without needing JavaScript and it presents no ads, so I'm not sure what your software is doing. Are you using wget or something?17:06
crashoverridecurl.17:06
masoncrashoverride: try w3m or elinks instead17:07
crashoverrideliterally: curl -SsL https://bpa.st/4JVQ | less17:07
crashoverridemason: w3m and elinks are SGML parsers.17:07
crashoverrideI was asking for a raw paste.17:07
masoncrashoverride: You might want to make use of a World Wide Web client here.17:07
crashoverrideokay I'll grep17:07
masonI'm afraid I can't help with this. It's a bit silly.17:08
crashoverridecurl -SsL https://bpa.st/raw/4JVQ17:08
crashoverrideit was too hard for you to visually identify the "raw" link, it's much easier for me to guess what it's gonna be called and grep it.17:08
crashoverridefine.17:08
crashoverrideanyhow, I can now read.17:09
crashoverridemason: is that enough to migrate?17:34
masoncrashoverride: It's what I've used, directly. I'd tend to pull out anything complex I know I can just reinstall, to simplify the process.17:37
crashoverridelike?17:42
crashoverrideXorg?17:42
crashoverrideany idea how to restrict a package to a specific repo?17:56
ShorTiepin it18:08
ShorTiea list for a hole repro like raspi.list18:09
crashoverridemyeah well18:15
crashoverrideI have no idea what you wrote.18:15
crashoverrideweird18:17
crashoverridedevuan is missing some packages...18:17
crashoverrideI have actually ZERO idea how bad this is18:19
crashoverridebut I just downloaded a couple debs from debian sid and `dpkg -i` them.18:20
crashoverrideproblem solved (for now, this sounds like it's gonna blow in my face later)18:20
tuxd3vcrashoverride, what is your hardware?18:20
crashoverridemy hardware or my arch?18:21
tuxd3vyes, what board are you running?18:21
crashoverrideone sec, will look up18:21
tuxd3vx86?18:21
crashoverridemy board is a 20AN0072GE18:22
tuxd3vho you are running x86, I tough for a moment that you were running arm or arm64, my bad :/18:24
crashoverrideI should be running x86_64 really, but my current arch is i686.18:24
crashoverrideI find your questions confusing.18:25
crashoverrideyou asked for my hardware, and you ended up talking about the arch for which the software I run has been built.18:25
tuxd3vcrashoverride, if you find it confusing its because its confusing :)18:25
crashoverridethere's nothing confusing.18:25
crashoverrideThere's a board: 20AN0072GE18:25
crashoverrideThere's a CPU: i7-4810MQ18:26
crashoverrideThere's a software arch: i68618:26
tuxd3vyeah you should have replied that you have a lenovo 14" laptop, and I would now at that very moment that you were running x86 or x86_6418:26
crashoverridewell, you asked about my "hardware"18:27
tuxd3v would now ->  would know18:27
crashoverrideyou're lucky I didn't tell you what gun I own or what car I drive :D18:27
tuxd3vcrashoverride, indeed18:27
crashoverrideeither way18:27
tuxd3vwhat packages are you complaining about?18:28
crashoverridenng-utils, libnng1, and libnsl2.18:28
crashoverridethe rest was available in the normal repositories.18:29
crashoverrideis it a hell of a lot of noise for 2 files...18:31
crashoverride(and two symlinks)18:31
crashoverrideok, 3 files, actually.18:31
crashoverride3 packages, 3 files.18:31
tuxd3vI am in Chimaera( i686 ), and those packages are available18:32
crashoverrideyeah but that's not stable is it?18:32
tuxd3vnot yet18:32
crashoverrideexactly.18:32
crashoverridewhat versions do you have?18:32
tuxd3vhttps://paste.debian.net/hidden/53d183bd/18:34
crashoverrideSGML again.18:34
crashoverridecan't you people use normal pastebins?18:34
tuxd3vnope, is has tons of javascript trying to mess with my cookies and the rest :)18:34
tuxd3vcookies are a precious thing to me, I guard them close to my chest :)18:35
crashoverrideI see.18:36
crashoverrideso I got *exactly* the same versions installed.18:36
crashoverridewould be cool, maybe, to put them into backports?18:37
masoncrashoverride: Sorry, there's enough going on that I often won't see replies on IRC if they don't hit a highlight. Yes, graphical stuff, things like LibreOffice, web browsers... Big things that have lots of dependencies.18:38
crashoverrideyep18:38
tuxd3vhave you tried the backports way?18:38
crashoverridetuxd3v: no, that's why I'm complaining about them not being on backports. Because I didn't look. Otherwise, I might know they aren't there.18:38
gateway2000crashoverride what release are you on18:44
gateway2000those packages aren't in buster either18:45
crashoverridegateway2000: < crashoverride> but I just downloaded a couple debs from debian sid and `dpkg -i` them.18:51
crashoverrideand to answer your question, I'm using beowulf.18:51
masoncrashoverride: Better to manually backport rather than trying to fit binary packages from another release.18:59
crashoverridemanually backport?19:00
crashoverridewhat's that?19:00
masonhttps://wiki.debian.org/SimpleBackportCreation for example19:00
crashoverrideah yeah, no.19:01
crashoverrideI'm not gonna be a debian maintainer because of 3 files.19:01
crashoverridethank you very much.19:01
masonJust trying to help. It's really not that hard.19:02
crashoverridebut this is literally 3 files.19:03
crashoverridenot exagerating here.19:03
crashoverrideif you'd check how much data is from the actual payload and how much is from the package management and conventions, I wouldn't be surprised more than half is added as overhead.19:04
crashoverrideplus, according to tuxd3v, those files are in Chimera, so when I upgrade, I'll be fine.19:07
golinuxThe files will be in chimaera not chimera.  ;)19:08
gateway2000I don't know who was here earlier but I had a problem with tofu characters on the desktop from the live install - I did end up reinstalling from the installer iso and the problem was resolved - not sure why this happened19:13
gateway2000Oops nevermind I'm doublechecking and they're still there - in the Display xfce4 settings module, my monitor name is nonsense19:14
gateway2000https://i.imgur.com/J7Ydq7x.png19:19
gateway2000This is what I'm talking about - it's supposed to say IBM 17" - I don't know if anybody else's run into this.19:20
gateway2000I've tried regenerating locales and it didn't do anything. It's also strange that in the live usb session this doesn't happen, and it actually didn't happen in the installed OS either at first, I just rechecked it and it's come back19:21
gateway2000Since it doesn't have after install, but happens eventually, maybe that means it has to do with packages I installed post-installation19:36
gateway2000Sorry to flood the channel, if anybody has any ideas I'm all ears.19:36
crashoverridegateway2000: what are tofu characters?19:37
gateway2000crashoverride Nonsense filler characters - for example like in the screenshot I shared19:38
crashoverridegateway2000: are you maybe using a non-UTF8 locale?19:40
gateway2000I thought maybe that had happened on accident, but locale outputs en_US.UTF-8 down the line19:41
crashoverridewell, clearly *something* isn't UTF8 here.19:42
gateway2000I turned off one of my monitors in grub with a kernel line, when my build is finished I'm going to try undoing that and rebooting to see if that makes a difference19:42
gateway2000Yeah something's up for sure19:42
gateway2000in xsettings the monitor name is correct. just not in the gui display module19:43
crashoverridegateway2000: do you know what the first character is supposed to be?19:44
gateway2000should be I19:44
gateway2000the name is supposed to be IBM 17"19:45
crashoverrideIP?19:45
crashoverrideah19:45
rmFrance o.o19:47
gateway2000yeah I know right19:47
gateway2000Never seen anything like it19:47
crashoverrideThe only thing I can sortof make up with it is 倐19:49
rmtry a memtest19:49
rmwith memtest86+, i.e. rebooting into it19:50
rmI'd use a live CD from https://grml.org/, and it has that memtest in "add-ons"19:50
crashoverrideah, no, could also be ၐ19:51
crashoverrideI'm guessing "France" is placeholder for your font for the UTF8 French flag.19:52
gateway2000Okay thanks all. Gonna reboot, try a few things. Will be back19:57
rmnext we know the thing is overclocked :)19:57
crashoverridequick question19:58
crashoverridewhen I dpkg --get-selections, the stuff that is listed as 'deinstall' is explicitely black listed, right?19:58
rwpcrashoverride, No.  Those are packages that are in the "rc" state.  Removed but Conffiles remain.  The package is still half-installed in that state.20:43
rwpTry this: dpkg -l | grep ^rc20:44
crashoverriderwp: thanks :)20:44
rwpThat is the difference between "removed" and "purged".20:44
crashoverriderwp: it's a package from the devuan installer that is in that state.20:44
crashoverrideit's "deinstalled" literally20:44
gateway2000memtest said the memory's fine, my fiddling didn't do anything. guess I'll just forget about it20:47
brocashelmanyone here using a cyberpower ups with their devuan systems? i installed their powerpanel deb (said to work on debian 8) on ceres (unstable) and it works just fine. great cli tool to communicate with my ups as i'm tired of sudden power shutoffs (for a split second) in my area20:47
brocashelmhttps://www.cyberpowersystems.com/product/software/power-panel-personal/powerpanel-for-linux/20:47
brocashelmi can check it out with sudo pwrstat -status20:47
plasma41brocashelm: Yes, I'm using a CyberPower UPS with my Devuan system.21:59
plasma41I've never heard of PowerPanel before. I'm reading about it now22:01
rwpcrashoverride, Right.  "deinstalled" is the string stored in dpkg's /var/lib/dpkg/status file. Maps to "rc" in the "dpkg -l" output. Probably will also have "config-files" in the package status field.22:04
brocashelmyeah, it works really nicely. had no issues installing it on ceres22:04
rwpcrashoverride, See "man dpkg-query" for documentation on package status states.22:05
plasma41brocashelm: It has a non-free license, unfortunately22:05
brocashelmtrue22:06
plasma41brocashelm: The only other dedicated UPS monitoring and control software I've heard of is NUT (Network UPS Tools). It's packaged in De{bi,vu}an and is free software.22:10
brocashelmi'm going to test that one out22:11
plasma41brocashelm: There's also some UPS related functionality in the Xfce Power Manager. It does what little I need, so I've never done a deep dive into NUT.22:13
brocashelmplasma41: i can see that22:16
rwpFor APC UPS devices apcupsd works well and is free software and is packaged for the OS.22:47
Centurion_DanI always stuck with nut - it has a profile for APC ups's22:58
flingWhat is needed for openvpn to set nameserver?23:19
rwpVPNs and nameservers are interesting, if you need to have a specific one in order to get local names resolved from the VPN.23:22
rwpI don't have an alive configuration right now but I remember that I had used resolvconf with a custom script to do the right thing, coupled with server side config with OpenVPN.23:23
rwpSorry that is such a vague, vague, vague description of one possible solution.23:23
rwpHaving always been a bind user it pains me to recommend unbound which makes overriding nameservers trivially easy.23:24
rwpAdditionally depending upon what you are doing and how you need to do it "sshuttle" is trivially easy to use with specific dns and works well, but is not a full VPN so it all depends.23:24

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!