libera/#devuan/ Monday, 2024-07-08

Xenguyrwp, Good tip, seems to work in both tmux and screen, so no need to remember different methods for each05:31
Xenguy'q' FTW05:32
rwp\o/06:08
gnarfacei guess i didn't notice the scrollback thing because most my headless systems are on older kernels than that...07:49
gnarfacewhat a shame07:49
cousin_luigiSo, does anyone use hardware flow control on console ports?09:11
_al1r4di have a question09:13
_al1r4dhow to disable cups-browser?09:14
gnarfacenot clear on what it's for really, but on the one set of ARM devices i regularly use, the instructions from the community have been to explicitly disable hardware and software flow control09:14
_al1r4dalready did update-rc.d but not ok for the results09:14
gnarface_al1r4d: not sure what cups-browser is, do you just mean CUPS the printer daemon?09:14
_al1r4dyeah, cupsd, gnarface09:15
gnarface_al1r4d: oh, it's not good enough to remove the symlinks. change them all from S to K09:15
gnarfacedepending on how you called update-rc.d it may have just removed them09:15
_al1r4dupdate-rc.d cups remove09:16
gnarfacei think you want disable instead of remove, but check the man page (i usually just do them by hand)09:16
gnarfacei also put K01cups symlinks even in the runlevels were there wasn't a corresponding S* one09:16
gnarfacethat took it offline09:16
gnarfaceyou'll probably want to do the same thing for saned if you have it too09:16
gnarfacethere's a way to call update-rc.d with more specificity, it might help09:18
gnarfacereally, creating the symlinks by hand is easier to figure out though09:18
gnarfaceshould look like this when you're done: https://paste.debian.net/1322647/09:19
gnarface(note that they're relative-path symlinks instead of absolute-path is important, though i can't remember exactly why)09:19
_al1r4dokay09:19
_al1r4dthank you09:19
gnarfaceno problem09:19
_al1r4dsudo update-rc.d -f cups defaults k09:20
_al1r4dlike this?09:20
gnarfaceuh, not sure09:20
gnarfacejust check the output of "ls -l /etc/rc*.d/*cups*" when you're done to verify09:21
gnarfacei thought it was something more like: update-rc.d -f cups disable 12345609:21
_al1r4dgnarface09:22
_al1r4dhttps://paste.debian.net/1322648/09:22
_al1r4dnot all services are K01 like yours09:22
gnarfaceyea, cups-browsed is different. you'll have to disable that one too i guess. i don't know what that even is, i print fine without it installed.09:22
gnarfaceymmv (maybe differnet printers need it?)09:23
gnarface*different09:23
gnarfaceit should be disableable the same way as cups, this in fact should work in general for all the daemons09:23
_al1r4dls -l /etc/rc*.d/*cups09:25
_al1r4dlrwxrwxrwx 1 root root 14 Jul  8 14:19 /etc/rc1.d/K01cups -> ../init.d/cups09:25
_al1r4dlrwxrwxrwx 1 root root 14 Jul  8 14:19 /etc/rc2.d/S04cups -> ../init.d/cups09:25
_al1r4dlrwxrwxrwx 1 root root 14 Jul  8 14:19 /etc/rc3.d/S04cups -> ../init.d/cups09:25
_al1r4dlrwxrwxrwx 1 root root 14 Jul  8 14:19 /etc/rc4.d/S04cups -> ../init.d/cups09:25
_al1r4dlrwxrwxrwx 1 root root 14 Jul  8 14:19 /etc/rc5.d/S04cups -> ../init.d/cups09:25
_al1r4dokay09:25
_al1r4dhmm09:25
_al1r4dnot all k0109:25
gnarfacethey're just symlinks, you can create them by hand with ln -s [target] [name]09:38
gnarfaceor rename them with mv09:38
gnarfacethey're neat but they're neither complex or magical09:39
systemdleteI suppose I shouldn't complain, and I don't want to jinx it, but the apt-cacher-ng has been working without problems for the better part of a week now.  I say "better part of" because it has occasionally glitched here and there, although not anywhere close to the headache it was creating earlier on.09:42
gnarfacei still think that eventually it's going to be traced to an anomaly that can only be fixed at the repo side09:42
gnarfaceit's gotta be something that just a few of them do... differently09:43
systemdletegnarface, I hope whatever we find out, we find out sooner than later.  Thank goodness there haven't been any great number of updates recently, although there was that one great big one this past week with the new kernel and dozens of other things.09:43
gnarfacemaybe not wrong exactly, but something apt-cacher-ng can't handle because of its inherent nature as a proxy09:43
systemdleteAnd--get this, gnarface--not a single glitch updating about a dozen VMs and hosts.09:44
gnarfacehuh, interesting09:44
gnarfacethat's probably statistically relevant09:44
systemdleteactually, no.  I lied.  about half a dozen.09:44
systemdleteso:  Only do VERY LARGE UPGRADES.09:45
systemdlete(infrequently)09:45
systemdleteI had all of them chugging simultaneously, too.09:45
gnarfacehmm09:45
systemdleteyeah.  Makes ya think09:45
systemdleteOn a slightly differnt topic:  Any experience with NUT?  (Network UPS Tools)09:47
systemdleteI have 2 different models of APC UPSs that have identical USB vendor and product IDs.  (The standards people draft a standard, and then the mfrs... ah, nvm)09:48
adhocthe vendor probably uses the exact same USB module in their products ...09:49
systemdletethe drive, usbhid-ups, continually dies and restarts, making its usefulness somewhat shaky09:49
adhocthe vendor just sees the USB thingy as a replacement for the Derial port?09:49
adhocserial port09:50
systemdletesure, makes sense enough09:50
adhocand not product in itself09:50
systemdletebut that means the API to the device should be the same... oh, wait.09:50
systemdletemaybe not.09:50
adhocthe module has "a" venfor and "a" product ID09:51
systemdletebut anyway, aside from the problems of differentiating them, which NUT does all right,09:51
adhocand not that the thing it plugs into is different09:51
adhocNUT probes the UPS direclty09:51
adhoc?09:51
LucentWsystemdlete: you might make your own serial cable, usually APC UPSes have both USB and serial09:51
systemdletewell, the driver keeps flaking on me09:51
LucentWso you can use NUT instead of HID shenanigans09:51
adhocusd the serrial port directly09:51
systemdletethese are both USB only09:52
adhoc=/09:52
adhocit may have a serial port internally09:52
adhocAPC has dropped its standards a lot over the years09:53
LucentWit should have a fake Ethernet port which actually is serial09:53
adhocan RJ45 connector?09:53
LucentWyes, it's a RJ45 but actually it's a custom serial, it's not the "standard" RJ45 to RS232, you need to wire your own09:54
adhocjoy09:54
LucentWbut it's fairly easy, at previous job I made my own to reset the battery status when I replaced them so I didn't have to bother having a whole setup of their PowerChute to do the same thing09:55
systemdleteThe XS1300 has a RJ45 to USB connector, and the cable comes with it.09:55
systemdleteThe ES650 iirc, is just a standard USB A-B cable09:56
systemdletesometimes mfrs will change these features and not so much as change the carton, let alone the UPC code and instructions.09:56
systemdleteso it is quite possible that these had serial ports in the olden days09:57
LucentWif you check the 'net, you should be able to find plenty of resources about folks using serial even on USB or network-aware UPSes09:57
LucentWI mean, serial pretty much needs only 3 pins to work if you don't need further signaling09:57
LucentWas in RX, TX and ground09:57
systemdleteso, let me understand:  I can *possibly* connect that RJ45 to a (somewhat) normal CAT6 cable?09:58
LucentWso over a 8c RJ45 they can fit both USB and serial interfaces09:58
LucentWbut on different pins, obviouslt09:58
systemdlete"can" fit...09:58
systemdleteQ:  Does it cause the unit to catch fire, burn up, and explode if wired incorrectly?09:58
systemdlete(I'm serious.  I'm a EE noob)09:59
LucentWnah, it got protections, if you wire them incorrectly the UPS turns off09:59
LucentWbeen there, done that09:59
gnarfacesystemdlete: heh, no, i dont' attach my UPSes to the network. i watched MR Robot.09:59
systemdleteWhat about the other unit, the ES650?10:00
systemdleteI don't recall anything other than a USB on that one10:00
systemdlete(this is for science, only)10:02
systemdleteYou know what, the ES650 might have surge suppression filters for ethernet.10:09
systemdleteI wonder if those can be used for monitoring?10:09
systemdletespecs for the XS1300:  Interface--USB and Serial10:44
fonkeyhey all14:01
gordonDrogonNot specifically a Devuan issue, but maybe some kind soul here has a pointer: Video playback (MP4 off my GoPro) played in firefox work fine - nice and smooth, but in VLC it's somewhat frame rate limited - audio is fine it just appears a bit jerky. It works well in both on my laptop though. Desktop has been upgraded from B to C and now D. Laptop was a fresh install of D.14:39
gordonDrogonI suspect some configuration somewhere might at an issue - especially as I recall having some video issues when I did the upgrade but I really can't recall what it was back then.14:40
djphsimplest test to rule out a systemic issue --> create a temporary / test user on the desktop, and verify VLC behaves now14:41
gordonDrogonah, so environment thingy.14:43
fsmithredDifferent graphics hardware? Does one need firmware?14:48
gordonDrogonwell ,same with a different user.14:49
gordonDrogonfsmithred, same PC - Firefox: OK, VLC: jittery.14:49
amarsh04installing mpv and trying different --hwdec= options can help identify if a specific hardware decoder has problems14:50
amarsh04it's a little quicker to try that than change vlc settings, then quit and restart vlc14:50
gnarfacegordonDrogon: first idea that comes to mind is make sure you have both the vdpau and vaapi driver packages for your video hardware installed14:53
gordonDrogonI'm fairly sure it's all there else what might firefox be using..14:55
gnarfacethat's the point14:55
gnarfacefirefox is using vaapi14:55
gnarfacemost stuff uses vdpau though14:55
gnarfacemost modern video cards support both14:55
gnarfaceinitially, vaapi was the intel offering, and vdpau was the nvidia offering14:56
gnarfaceif the desktop where it's both working is amd or nvidia, but the laptop is older intel though, it might be an actual hardware limitation...14:56
gnarfaceor, sorry, switch those14:57
gordonDrogonit's a dell something with on-board graphics.14:57
gnarfacethose are usually intel14:58
gordonDrogonit is.14:58
amarsh04lspci can identify it quickly14:58
gnarfacecan't hurt to try it14:58
gordonDrogon00:02.0 VGA compatible controller: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller (rev 06)14:58
gnarfacejust make sure you have both mesa-vdpau-drivers and mesa-va-drivers14:59
gnarfacehere, when in doubt just install everything i've got: https://paste.debian.net/1322667/15:01
gordonDrogonii  mesa-va-drivers:amd64                 22.3.6-1+deb12u1                        amd64        Mesa VA-API video acceleration drivers15:01
gordonDrogonii  mesa-vdpau-drivers:amd64              22.3.6-1+deb12u1                        amd64        Mesa VDPAU video acceleration drivers15:01
gordonDrogonI appear to have both,.15:01
gnarfacehmmm15:01
gnarfacemight just be that the hardware can't do vdpau, and vlc insists on it, but maybe there's a way to force it...15:02
gordonDrogonit's not desperatly important. using ff to do quick video previews is fine for me for now.15:02
gnarfaceoh, one thing you can try15:03
gnarfacei forgot15:03
gordonDrogonvlc has an option in the codes tools - it just has auto/vdpau/disable15:03
gordonDrogonseems to make no difference which I choose.15:03
gnarfacehang on, intel-specific packages incoming15:03
gnarfacetry these: i965-va-driver, intel-media-va-driver, libvdpau-va-gl115:04
gnarfaceand also, double check you have these: https://paste.debian.net/1322668/15:04
gordonDrogonoh, I just installed mpv - silky smooth.15:04
gnarfacempv is probably smart enough to use either15:05
gordonDrogonthose are already installed.15:05
gnarfacethough, for what it's worth that libvdpau-va-gl1 is a wrapper from opengl to vaapi, so for old intel hardware you can set it to regular opengl accel and it'll translate it to vaapi, works decent15:06
gnarface(useful for mplayer, i suppose that's one thing mpv adds over mplayer though015:06
gnarface)15:06
gordonDrogonand I have all the amd64 version of those packages in that paste.15:06
gordonDrogonmaybe vlc's days are numbered...15:06
gnarfacei think hardware that doesn't support both protocols days are numbered15:07
gnarface*these days15:07
gnarfacethese days' days are numbered?15:07
gnarfacenevermind15:07
gordonDrogon:)15:07
gnarfacethere might be a way to build vlc with vaapi support15:08
gordonDrogonhttps://wiki.videolan.org/VLC_VAAPI/15:13
gordonDrogonlooks like it all ought to be there15:13
gordonDrogonand theres a load of 'failed' output messages when trying to open stuff15:15
gordonDrogonhttps://unicorn.drogon.net/vlc0.txt15:15
gordonDrogonbut sometimes when you find another solution.... always odd to change though as vlc has served me well for decades... (or it seems that way)15:18
gnarfacedo you have these? i965-va-driver, intel-media-va-driver, libvdpau-va-gl115:19
gnarfacethere's still every possibility you're just missing a library15:19
gnarfaceor mabye there's just something weird about the debian build...15:20
gnarfaceyou should be able to verify by the output of vdpauinfo and vainfo15:20
gordonDrogonyes, got those.15:21
gordonDrogonlet me just check it again on my laptop - I'm sure it was working just fine there.15:21
gordonDrogonit's also a dell/intel.15:21
gnarfaceif you have a good cpu and it's a small video it might not be obvious by framerate all the time, but there should be an obvious difference in cpu load15:22
gordonDrogonyea, silky smooth with vlc on the laptop - I'm suspecting something left-over from the upgrade process I did some months back15:22
gnarfacei don't actually mess with intel video a lot, so unfortunately i can't really test here15:23
gnarfacejust going off old info from helping other people a couple releases back...15:23
gordonDrogonthanks.15:23
gordonDrogonlaptop has cpu topping at 75% but desktop is peaking 110% for vlc. guess it's not using hardware15:24
gnarfaceoh, hmm... i did have a intel laptop though and something else occurs to me...15:24
gnarfaceyou're using xorg not wayland, right?15:24
gordonDrogonright15:25
gnarfacei recall that the default xorg driver at some point changed from intel to modesetting (at intel's request)15:25
gnarfacebut i discovered that some video decoding profiles didn't work in modesetting15:25
gnarfaceso i had to switch it back to intel15:25
gnarfacethat might only affect certain encodings15:26
gordonDrogonI've not really looked 'under the hood' for a long time now. sort of given up as it was a bit too much to keep up with15:26
gnarfacetry like some h264 videos and see if they behave any better, for example15:26
gnarfacelook in the output of "vainfo" and try something that matches one of them15:26
gnarfaceyou said it's MP4 video, but MP4 is the name of both a container format and a encoding format, and they're not necessarily always used in conjunction (rarely in fact, these days)15:27
gnarfaceso there's ambiguity there15:27
gordonDrogonthe files do appear to be h264 - right off the GoPro H815:28
gnarfacehmm15:28
gordonDrogonyea, I understand the container thingy though.15:28
gnarfaceusually h264 is what everything supports best these days, though with some hardware (my raspberry pi) it seems to be picky about pixel format as well15:28
gnarfacecameras often use yuyv422, or something similar, but certain decoding hardware insists on it being yuv420p or it won't accelerate15:29
gordonDrogon(Ugh. heres another whinge  - why do some applications change the CWD on startup!) vaguely related to this - openshot-qt I'm looking at you - I cd to dir fill of files, start you and you default to $HOME! Curse you.15:30
gnarfaceif the video isn't in yuv420p you could try re-encoding it as such with ffmpeg15:31
gnarfacethen see if it behaves better15:31
gnarfacei think with that i'm out of ideas though15:31
gordonDrogonit's fine,. thanks, I have solutions. Lookis like openshot is just as broken as it ever was though.15:35
gordonDrogonbut since I managed to lose the battery door to the GoPro the other day I may not be taking videos for a while.15:36
gordonDrogonstupid design.15:36

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