systemdlete2 | so when I create an xauth for root I can run virtualbox under xtrace. Just won't run withuot xtrace. Grrrrrr | 00:03 |
---|---|---|
systemdlete2 | AHA! I've got it, by George. (sorry George) | 00:05 |
systemdlete2 | So, this WILL work, but it's a PITA. First create the xauth file for root. THen run "su - root -c 'virtualbox -- -d :0.0'" (single quotes are important!) | 00:06 |
systemdlete2 | I was able to generate a xauth for root by running "xauth add ${HOST}:0 . $(xxd -l 16 -p /dev/urandom)" | 00:07 |
systemdlete2 | So, this can be done, but what a fiasco. | 00:07 |
systemdlete2 | fsmithred: Who do I seek out about addng to the release notes? Is that the appropriate place for this info? Will anyone remember them when they get to installing vbox, perhaps months later? | 00:08 |
fsmithred | me | 00:09 |
systemdlete2 | oh. Hi me. | 00:09 |
fsmithred | you can start root apps if you revert to the old behavior of su | 00:09 |
systemdlete2 | So there are two ways to do it. | 00:09 |
fsmithred | don't need to mess with xauth or xhost | 00:09 |
fsmithred | I know five | 00:09 |
fsmithred | (I think) | 00:10 |
fsmithred | I think the easiest is to put 'ALWAYS_SET_PATH yes' in /etc/default/su | 00:10 |
systemdlete2 | how do you get past the fact that the x11 client (in this case, vbox) can't connect to the display? | 00:11 |
fsmithred | I haven't tried it with vbox | 00:11 |
fsmithred | anything else works | 00:11 |
systemdlete2 | We are talking about virtualbox, not anything else, working or not. | 00:12 |
fsmithred | you want to run the regular vbox gui as root? | 00:13 |
systemdlete2 | IOW, how can a beowulf user install vbox extension pack. That's what this has been about the whole time. No "goalposts" moved, or even wiggled. That's all I've been after. | 00:13 |
fsmithred | I don't know. I'm trying to help you find out. | 00:13 |
fsmithred | but I'm not going to install it here. | 00:14 |
rrq | systemdlete2: perhaps starting the X server with "-auth /var/run/common-xorg-access.auth" is an option? | 00:14 |
systemdlete2 | No, as I said before, *I* do not want to run vbox gui as root. The problem, again, is that a regular user can no longer install them themselves. | 00:14 |
fsmithred | and what's the problem with installing them as root? | 00:14 |
systemdlete2 | rrq: Would you really want to do it that way | 00:14 |
systemdlete2 | fsmithred: virtualbox is a gui. | 00:15 |
fsmithred | and? | 00:15 |
systemdlete2 | There IS a command line option I think, and that might circumvent all of this. | 00:15 |
fsmithred | you have to run that as root? | 00:15 |
systemdlete2 | fsmithred: WHy not try installing virtualbox extension yourself? Then you will find out the joy I have. | 00:16 |
systemdlete2 | I am NOT talkig about guest additions. Haven't gotten to that. | 00:16 |
systemdlete2 | These are the extensions that enable things like USB, etc | 00:16 |
fsmithred | no thanks. | 00:16 |
fsmithred | I know. I've used it in the past. | 00:17 |
fsmithred | I'm glad not to have it bothering me about updates all the time | 00:17 |
fsmithred | and having to reinstall guest additions | 00:17 |
systemdlete2 | OK... so how do we proceed from here? | 00:17 |
systemdlete2 | Not everyone in the world will be keen to switching from virtualbox | 00:18 |
fsmithred | well, if you want to install from the gui | 00:18 |
fsmithred | and it will only let root install | 00:18 |
systemdlete2 | In fact, some might even have quite a large investment in it. | 00:18 |
fsmithred | have you tried running the gui as root to install the extensions? | 00:18 |
systemdlete2 | have you? | 00:18 |
fsmithred | and have you tried the commands on the page you linked earlier? | 00:18 |
systemdlete2 | Of course I have. That is how I got to this point! | 00:19 |
systemdlete2 | That's precisely what I haev been doing , yes | 00:19 |
fsmithred | which? | 00:19 |
systemdlete2 | (sorry this keyboard on the testbox is lousy) | 00:19 |
fsmithred | the commands work? | 00:19 |
systemdlete2 | all of the above | 00:19 |
fsmithred | gui as root works? | 00:19 |
systemdlete2 | no | 00:19 |
fsmithred | because it won't run in user session? | 00:20 |
systemdlete2 | what do you mean? | 00:20 |
systemdlete2 | maybe scroll back and see what we've discussed so far. | 00:20 |
fsmithred | if you're in a user session on the desktop, and you get root in a terminal and then start the vbox ui, does it run, or do you get an error? | 00:20 |
systemdlete2 | rather than me retyping it all | 00:20 |
fsmithred | <systemdlete2> how do you get past the fact that the x11 client (in this case, vbox) can't connect to the display? | 00:21 |
fsmithred | sorry, but I don't exac tly know what you mean there | 00:21 |
systemdlete2 | the regular user (person who will ultimately use virtualbox) "owns" the X display | 00:21 |
fsmithred | right | 00:22 |
fsmithred | I already gave you the fix for that | 00:22 |
systemdlete2 | root cannot send to it | 00:22 |
fsmithred | revert the behavior of su | 00:22 |
systemdlete2 | how do you "revert the behavior of su"? | 00:22 |
fsmithred | I think the easiest is to put 'ALWAYS_SET_PATH yes' in /etc/default/su | 00:22 |
systemdlete2 | go back to ascii? | 00:22 |
* Xenguy 's head starts spinning around ... | 00:22 | |
fsmithred | I think the easiest is to put 'ALWAYS_SET_PATH yes' in /etc/default/su | 00:22 |
* systemdlete2 's also | 00:22 | |
fsmithred | take you a couple minutes to try that | 00:23 |
systemdlete2 | fsmithred: even if I gave the entire path explicitly, it doesn't cahnge matters | 00:23 |
systemdlete2 | *change | 00:23 |
fsmithred | I know | 00:23 |
fsmithred | you need to keep user's environment but get root's path | 00:23 |
systemdlete2 | so what would setting the path do for the x permissions rproblem | 00:24 |
fsmithred | and use 'su' instead of 'su -' | 00:24 |
systemdlete2 | OK. Now I see the problem. | 00:25 |
systemdlete2 | You said "revert the behavior of su" | 00:25 |
fsmithred | yeah, like it was in ascii and before | 00:25 |
systemdlete2 | I thought you meant installing an older versoin of su or something | 00:25 |
fsmithred | ewww | 00:25 |
fsmithred | lol | 00:25 |
systemdlete2 | yeah | 00:25 |
systemdlete2 | that's generally what people mean by "revert" | 00:26 |
systemdlete2 | but thanks, that solves it I think. | 00:26 |
fsmithred | I hope so | 00:26 |
systemdlete2 | at least, the gui comes up | 00:26 |
systemdlete2 | let me try doing the actual install though | 00:26 |
fsmithred | yeah | 00:26 |
systemdlete2 | that worked. thanks again. sorry for the mismatch in English. | 00:28 |
fsmithred | yw | 00:28 |
systemdlete2 | (did you know that "deprecate" means to insult someone or something? I've never met a software that felt insulted by anyone or anything, except, perhaps, the person who wrote it. They often got angry, yes. But not the software, incredibly, despite the mis-use of the word.) | 00:29 |
systemdlete2 | (and there are more words like this in softwaredom. It never ceases to amuse me.) | 00:30 |
systemdlete2 | btw, there is no need to set any envars if you omit the - argument. At least, I didn't make any changes | 00:31 |
systemdlete2 | And why do you put it as "like it was in ascii and before" -- afaik, "su -" has been around since time immemorial (I was using it back in college in the early 80's) | 00:34 |
systemdlete2 | I don't remember a time when it was not an option | 00:34 |
systemdlete2 | maybe my memory plays tricks on me again... | 00:34 |
meep_____ | » [16:34:10] <systemdlete2> And why do you put it as "like it was in ascii and before" -- afaik, "su -" has been around since time immemorial (I was using it back in college in the early 80's) | 00:36 |
meep_____ | since epoch | 00:36 |
systemdlete2 | yes, since the epoch. I am 100% certain that the su command has not changed much since then. | 00:37 |
meep_____ | It hasn't | 00:37 |
meep_____ | It's gotten some user-friendlyness enhancement | 00:37 |
meep_____ | S | 00:37 |
meep_____ | Like help messages | 00:37 |
meep_____ | But that's it | 00:37 |
meep_____ | It doesn't do much | 00:37 |
meep_____ | Just changes uid | 00:37 |
systemdlete2 | sure, but the arguments it supports in beowulf haev nothing to do with dropping the - option to su | 00:38 |
meep_____ | There's a "discrete logic" implementation of SU in the US patent archives | 00:38 |
meep_____ | They did this because software patents wasn't a thing, so they had to represent 'su' as a circuit | 00:38 |
meep_____ | systemdlete i'm not familiar with arguments about su in beowulf | 00:38 |
meep_____ | What's happening with su? | 00:39 |
systemdlete2 | yeah, you can now pass arguments to su. So in this case, one runs "su root -c virtualbox" | 00:39 |
systemdlete2 | that wasn't possible before. | 00:39 |
meep_____ | Oh, yeah that too | 00:39 |
meep_____ | So what, poettering and the gang are trying to remove that in beowulf? | 00:40 |
systemdlete2 | but "revert" to how it worked in ascii would not be helpful, because in ascii, you could not enter that command. | 00:40 |
systemdlete2 | no, meep____ not at all | 00:40 |
meep_____ | You can also do sudo -u user cmdline | 00:40 |
meep_____ | Or su -- user to start a shell | 00:41 |
meep_____ | The OpenBSD folks re-implemented sudo and call it doas | 00:41 |
systemdlete2 | what fsmithred really wanted was to drop the root envirnoment | 00:41 |
meep_____ | Not sure why | 00:41 |
meep_____ | Or if doas offers any real improvement over sudo | 00:41 |
systemdlete2 | to create porting challenges | 00:41 |
systemdlete2 | keep things very interesting | 00:42 |
meep_____ | » drop the root envirnoment | 00:42 |
meep_____ | what do you mean | 00:42 |
systemdlete2 | "su cmd..." just runs the user's own environment and privs | 00:42 |
systemdlete2 | whereas "su - cmd..." runs it with root env and privs | 00:42 |
fsmithred | https://wiki.debian.org/NewInBuster | 00:43 |
fsmithred | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=905564 | 00:43 |
systemdlete2 | uh-oh | 00:43 |
fsmithred | https://www.debian.org/releases/buster/amd64/release-notes/ch-information.en.html#su-environment-variables | 00:43 |
meep_____ | » Wayland is the default session type for GNOME 3. (You may also choose an X session from the display manager.) | 00:43 |
fsmithred | when su was moved from shadow to util-linux, the behavior changed | 00:44 |
fsmithred | that happened in buster/beowulf | 00:44 |
fsmithred | it used to be that typing 'su' gave you root's path. No longer. | 00:44 |
meep_____ | I don't have any strong feelings about this change one way or another, however I do have to ask why | 00:45 |
meep_____ | Is this change for the sake of change? | 00:45 |
fsmithred | I don't know why | 00:45 |
meep_____ | Or was there a reason | 00:45 |
fsmithred | I find it very annoying | 00:45 |
fsmithred | usually when I become root, I want to do something in the current directory, and if the command is in any *sbin, I have to type the full path | 00:46 |
fsmithred | or use 'su -' which moves me out of the current directory | 00:46 |
gnarface | i always used "su -" anyway, because just the path isn't usually enough for what i'd want to do with it... | 00:46 |
gnarface | so i guess when this changed i didn't even notice | 00:46 |
fsmithred | I guess we do different stuff | 00:47 |
gnarface | but yea "su -" will also start you in the user's home directory | 00:47 |
fsmithred | I mount a lot of isos | 00:47 |
fsmithred | hm, maybe I don't need it for that | 00:48 |
systemdlete2 | so, essentially, aside from some of the other new features in su, devuan/debian's su now works the way it did in the 1970s. | 00:48 |
fsmithred | I missed that era | 00:48 |
systemdlete2 | I did too, actually. But by the time I got to college, it still worked that way. I have an ANCIENT AT&T Bell Labs System V Unix manual somewhere in my many possessions I have dragged around the US as I moved many times. | 00:49 |
systemdlete2 | I am pretty sure that su is described the same way then. | 00:50 |
systemdlete2 | (without these new bells and whistles aside from the semantic change ) | 00:50 |
systemdlete2 | in fact, I have been using the "-" option in ascii and jessie too. THat's not actually new, even in devuan. | 00:51 |
meep_____ | fsmithred: i have normally used sudo -i and trained a lot of people at the office to use sudo -i because it has increased auditing ability and fine control, like allowing %wheel to be root any time, allowing %webdev to restart the webserver daemon and applications, syn git repos etc | 00:51 |
systemdlete2 | and the "-" was the pivotal part of the thread that caused the confusion for me. | 00:51 |
fsmithred | 'xhost +' didn't work for you? | 00:52 |
rrq | if PATH setting solved it, it would probably have been just lack of DISPLAY setting | 00:52 |
rrq | if env setting... | 00:53 |
systemdlete2 | fsmithred: THere still needs to be an advisory for virtualbox. Otherwise, others may run into the same issues. | 00:54 |
fsmithred | virtualbox is not in the repository | 00:54 |
fsmithred | I'll expand the section that mentions the changes in su | 00:54 |
systemdlete2 | again, thousands of users and businesses use virtualbox | 00:54 |
fsmithred | and other third-party software | 00:54 |
systemdlete2 | or you could just give the one liner to make it easy | 00:55 |
specing | how did vbox take over qemu so hard | 00:55 |
specing | was it because qemu sucks at GUI? | 00:55 |
fsmithred | it's so much easier | 00:55 |
specing | err, wrong channel | 00:55 |
systemdlete2 | have you looked at vagrant? | 00:55 |
specing | systemdlete2: nah | 00:55 |
systemdlete2 | can qumu migrate VMs to other machines? idk actually | 00:56 |
systemdlete2 | My understanding is that businesses like the vagrant features, and I think vagrant is largely built on vbox, but it may support others | 00:57 |
fsmithred | qemu-img allows you to create, convert and modify images offline. | 00:58 |
systemdlete2 | thats just one example. Also, you can lock down vm sessions so users can't tool with the configuration or the running system (kiosks and the like) | 00:58 |
systemdlete2 | vboxmanage can do those things also | 00:58 |
systemdlete2 | and it actually supports qumu cow images (not sure how well though) | 00:59 |
systemdlete2 | last time I used qemu was back around 2005 I think. It was pretty impressive. I just found vbox easier to use. | 01:00 |
systemdlete2 | There are probably many users as lazy as me out there. | 01:00 |
fsmithred | qemu-img -h (to see supported formats) | 01:01 |
fsmithred | if I can find them | 01:01 |
fsmithred | Supported formats: blkdebug blklogwrites blkreplay blkverify bochs cloop copy-on-read dmg file ftp ftps gluster host_cdrom host_device http https iscsi iser luks nbd nfs null-aio null-co nvme parallels qcow qcow2 qed quorum raw rbd replication sheepdog ssh throttle vdi vhdx vmdk vpc vvfat | 01:02 |
meep_____ | » [16:54:03] <systemdlete2> fsmithred: THere still needs to be an advisory for virtualbox. Otherwise, others may run into the same issues. | 01:19 |
meep_____ | » [16:54:17] <fsmithred> virtualbox is not in the repository | 01:19 |
meep_____ | » [16:54:35] <fsmithred> I'll expand the section that mentions the changes in su | 01:19 |
meep_____ | » [16:54:44] <systemdlete2> again, thousands of users and businesses use virtualbox | 01:19 |
meep_____ | you should be able to switch to QEMU no problem. QEMU does everything virtualbox does and more, also without any proprietary binary blobs with native integration into Linux KVM | 01:19 |
meep_____ | » [16:56:16] <systemdlete2> can qumu migrate VMs to other machines? idk actually | 01:20 |
meep_____ | yes, before i moved to LXC containers i did this | 01:20 |
furrywolf | forcing people to transition to a new program, even if the program has more features, should not be done lightly. of course, I entirely missed the context of that paste. | 01:21 |
fsmithred | I don't think the release notes are the right place for guides to third-party software | 01:22 |
meep_____ | Furrywolf, virtualbox requires binary blobs to do basic things like usb2.0/3.0 passthrough, and i don't even know if it can access raw block devices yet | 01:22 |
meep_____ | And it's owned by Oracle | 01:23 |
meep_____ | It would be different if it was open source free software transitioning to something else | 01:23 |
furrywolf | was it included in the repositories previously? | 01:23 |
meep_____ | I don't know | 01:23 |
fsmithred | free version was in repo | 01:23 |
meep_____ | Qemu is really not that hard | 01:24 |
fsmithred | but there's always been the non-free version from virtualbox.org (oracle) | 01:24 |
furrywolf | will the old version continue to function with everything else upgraded, so existing installs don't break? | 01:24 |
meep_____ | Qemu-system-x86_64 -m 4096 -smp 4 -cdrom linux.iso | 01:24 |
meep_____ | 4 gig, 4 core x86 system in a one-liner | 01:24 |
meep_____ | You can add a hard drive in -hda disk.img | 01:25 |
furrywolf | I really, really, really hate "I ran apt-get upgrade and now things don't work". heh. | 01:25 |
fsmithred | I haven't tried an upgrade of ascii with virtualbox installed | 01:25 |
meep_____ | I like to do fancy stuff like virtio or emulated sas to ZFS Zvolumes | 01:25 |
fsmithred | it's actually in ascii-backports | 01:25 |
viciouspiranha | aqemu is a qt frontend to qemu, check it out | 01:26 |
specing | meep_____: how do I add an USB stick (file) while the VM is running? | 01:26 |
meep_____ | Not sure | 01:26 |
specing | See, it is hard | 01:26 |
meep_____ | There's a command line interface into the running qemu virtual machine | 01:26 |
onefang | libvirt allows you to be independent of VM type. It even has a GUI. It did suck enough for me to go back to qemu command lines and self written scripts. | 01:26 |
meep_____ | You may be able to attach it at runtime there | 01:26 |
furrywolf | so far, php and python have been the #1 offenders for upgrades breaking things. heh. | 01:27 |
specing | meep_____: there is, I've tried and it did not work | 01:27 |
meep_____ | onefang: libvirt is just a redhat abstraction layer. It ends up calling qemu-system-<cpuarch> | 01:27 |
onefang | Thus earning their place in my "I hate languages beginning with the letter P" list. | 01:27 |
meep_____ | Pavascript? | 01:27 |
onefang | I know that. It also abstracts other VM systems. | 01:28 |
onefang | So it would make it easier to migrate from vbox to qemu for example. In theory, never tried that. I always just stuck with qemu. | 01:28 |
meep_____ | You can use qemu-img to convert the virtualbox disk formats to standard one | 01:29 |
meep_____ | Word of advice though | 01:29 |
* furrywolf can't think of any other p languages | 01:29 | |
meep_____ | Use qcow2 instead of raw img files | 01:30 |
onefang | Ruby is an honourary member of that list of P languages, JavaScript is also called Ecmascript, if you add the numeric value of J and E together, you got close enough to P. B-) | 01:30 |
furrywolf | pascal! yeah, that was pretty bad too. | 01:30 |
meep_____ | furrywolf: is it? | 01:30 |
onefang | Perl, PL1. | 01:30 |
meep_____ | I've used gddrescue | 01:30 |
meep_____ | It's written in pascal | 01:30 |
furrywolf | oh yeah, perl.... | 01:30 |
meep_____ | onefang: oh no Perl is a fantastic language | 01:30 |
furrywolf | .. pascal still exists? I thought it died like 30 years ago. | 01:30 |
meep_____ | No complaints here about perl | 01:30 |
meep_____ | furrywolf: yes. You can get the Lazurus IDE from the ASCII repos | 01:31 |
furrywolf | have you ever tried reading someone else's perl? or, for that matter, your own? :) | 01:31 |
meep_____ | furrywolf: yes | 01:31 |
meep_____ | Have you ever tried reading someone else's C++? | 01:31 |
furrywolf | yes | 01:31 |
onefang | I started off writing APL, I can read line noise. | 01:31 |
meep_____ | How many times did you smash your desk | 01:31 |
furrywolf | although C is still my favorite language. | 01:32 |
onefang | Lua, C, and assembler are mine. | 01:32 |
meep_____ | I don't think saying Perl is unreadable has much merit | 01:32 |
meep_____ | It comes from SH, Awk, and Sed | 01:32 |
furrywolf | c++ tries to encourage clean programming... perl does not. | 01:32 |
meep_____ | And regular expressions | 01:32 |
fsmithred | yes, I've tried reading perl script written one year prior, and I did not understand it. | 01:32 |
onefang | You can write unreadable code in any language. | 01:32 |
meep_____ | If you understand those your golden | 01:32 |
furrywolf | python attempts to force you to write readable code | 01:33 |
meep_____ | Yes | 01:33 |
meep_____ | And i despise python for that | 01:33 |
meep_____ | It's like the Chasity belt of language syntax | 01:33 |
onefang | lol | 01:33 |
meep_____ | And don't get my wrong i have written a lot of software in python | 01:33 |
fsmithred | this really isn't the right channel for a discussion of different programming languages | 01:33 |
furrywolf | I started writing my own programming language once... maybe I should re-start that project one of these days. | 01:33 |
meep_____ | I've been writing all my newer stuff in perl | 01:34 |
rrq | please go to #debianfork | 01:34 |
furrywolf | it's mostly C syntax, but with all the nice things (proper strings, hashes, regexs, etc) as part of the language, everything passed by reference, a built-in rdbms wrapper,... | 01:34 |
meep_____ | Built in rdbms? | 01:35 |
meep_____ | In perl? | 01:35 |
meep_____ | What are you talking about? | 01:35 |
furrywolf | <furrywolf> I started writing my own programming language once... maybe I should re-start that project one of these days. | 01:35 |
meep_____ | I didn't know there was a built in rdbms | 01:35 |
rrq | please go to #debianfork | 01:35 |
onefang | Yeah, my fault, take the language discussion to #debianfork. | 01:35 |
meep_____ | Hold on | 01:35 |
* furrywolf notes onefang is not in #debianfork to continue being part of said discussion | 01:37 | |
onefang | I'm a compulsive reader, I try to stay in only a few low traffic channels, or I'll never get anything done. I'm here coz I'm a Devuan dev, so I need to help support Devuan like the rest of us. | 01:38 |
meep_____ | I'm concerned about Devuan | 01:41 |
meep_____ | Because I no longer trust Debian | 01:41 |
meep_____ | I know how dependent Devuan is on Debian | 01:41 |
meep_____ | Also, I want to contribute some software packages. Normally told to go upstream it on Debian instead of working downstream | 01:42 |
meep_____ | I don't trust Debian anymore and I really don't want to write anything for systemd integration | 01:42 |
onefang | You want to contribute software packages that are not in Debian? | 01:44 |
meep_____ | Y | 02:11 |
systemdlete | is it possible to have multiple video ports putting out the same output? I have a video card with vga and hdmi (and hdi, but I don't use that). What I want to do is route the video through both vga and hdmi simultaneously to 2 different devices. Is that possible? | 02:34 |
systemdlete | (there is also onboard vga and hdmi, but I think those are disabled by the bios) | 02:35 |
systemdlete | tia for any support on this. | 02:35 |
onefang | Mirroring a screen is a thing, so yes it's possible. | 02:36 |
systemdlete | how do I do that? Is that in the monitor settings? And what if one of the displays is routed through a kvm switch? Will it still work? | 02:46 |
meep_____ | » [18:34:59] <systemdlete> is it possible to have multiple video ports putting out the same output? I have a video card with vga and hdmi (and hdi, but I don't use that). What I want to do is route the video through both vga and hdmi simultaneously to 2 different devices. Is that possible? | 02:48 |
meep_____ | Yes, it's called mirring | 02:48 |
systemdlete | ok, thanks, I'll install it | 02:48 |
meep_____ | systemdlete the xfce has a standlone monitor configuration tool | 02:48 |
meep_____ | Gui | 02:48 |
meep_____ | Although you can also diy by editing xorg.conf directly | 02:48 |
meep_____ | Xfce4-display-settings | 02:49 |
onefang | A KVM shouldn't matter. | 02:49 |
systemdlete | apt install mirring does nothing (spelling? | 02:50 |
meep_____ | » [18:49:17] <meep_____> Xfce4-display-settings | 02:50 |
onefang | If I recall correctly, since I've only ever done this long ago by accident, you can even mirror screens at different resolutions, the smaller one will just have bits cut off. | 02:50 |
MinceR | you can also just use xrandr on the command line | 02:56 |
meep_____ | » [18:50:42] <onefang> If I recall correctly, since I've only ever done this long ago by accident, you can even mirror screens at different resolutions, the smaller one will just have bits cut off. | 04:12 |
meep_____ | The proprietary nvidia driver does that with with the framebuffer on boot | 04:12 |
meep_____ | Except is uses scaling | 04:12 |
meep_____ | So i'll have a 1080p image scaled on a 1600x900 monitor | 04:13 |
meep_____ | It looks cool | 04:13 |
meep_____ | I wonder if it could offer any usability improvement on lower res screens, at the cost of some blurriness | 04:13 |
onefang | I was just messing with some GUI on top of xrandr, on my VR test box that at the time had it's own monitor, and was sharing my main monitor via my KVM. | 04:14 |
gnarface | it should work both ways, you should be able to mirror with scaling or panning | 04:15 |
gnarface | (driver bugs not withstanding) | 04:15 |
onefang | Mostly I was just trying to get the left monitor to be on the left, and the right on the right. | 04:15 |
gnarface | (though with nvidia's binary drivers the bug is just usually "doesn't obey xrandr right so you have to use nvidia-settings instead") | 04:16 |
gnarface | heh, yea that's something the nvidia driver seems to be particularly bad at parsing when it comes from the xorg.conf for the xrandr command-line | 04:16 |
gnarface | but the nvidia-settings thing works usually | 04:17 |
onefang | This is a nVidia Optimus chip, an evil hybrid between Intel and nVidia graphics chips. That computer is small, made from laptop parts, but was the cheapest thing I could buy quickly so I could do VR development on when someone hired me to. | 04:18 |
onefang | Later Oculus decided that Optimus was too evil to support. lol | 04:18 |
systemdlete | xserver-xorg-video-s3 ? | 05:52 |
systemdlete | is this no longer supported? only on debian? | 05:53 |
systemdlete | or never existed or renamed? | 05:55 |
systemdlete | I found an old s3 video card to run the second vga monitor. I can see it in inxi -F (both video cards), but xorg does not seem to detect them. | 05:56 |
yeti | try the vesa server. that card is a pain as there is only acceleration upto 16bit modes and everything™ expects 32bit today | 06:25 |
yeti | there was a s3 xserver but it kind of is useless now maybe itÄs just no longer built by the deb-src's rules | 06:26 |
yeti | I think I've it in an old thinkpad | 06:26 |
yeti | ok... maybe my priblems were/are specific to the S3 variant in that notebook... take all that with a ton of salt | 06:29 |
yeti | P3M era notebook | 06:29 |
gnarface | systemdlete: yea it's gone. the cirrus logic driver is gone too. | 06:38 |
systemdlete | soooo. I can chuck the s3 card? | 06:39 |
systemdlete | (I've got more) | 06:39 |
gnarface | well, yeti's suggestion to just use the vesa driver is probably worthwhile | 06:39 |
gnarface | it's not like the hardware acceleration features are going to be good enough on that card that having support for them is going to buy you any practical improvement in any software currently in the wild | 06:40 |
gnarface | it would still be cripplingly slow either way | 06:40 |
gnarface | so yea i would just shrug it off | 06:41 |
gnarface | or, if you're really bored, get to patching it | 06:41 |
gnarface | that card has what, 2MB of video ram? 4? | 06:42 |
gnarface | heh | 06:42 |
gnarface | it's fine for virtual terminals | 06:42 |
plasma41 | systemdlete: xserver-xorg-video-s3 was removed after Wheezy. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=732476 | 06:42 |
systemdlete | I'm sure there is a recycler somewhere who would like it. | 06:43 |
gnarface | i'd take it off your hands for free but it's not worth the shipping cost | 06:43 |
gnarface | (i have a few here too) | 06:43 |
gnarface | BSD might still have a real s3 driver integrated properly though, if it ever did. whatever reasons it was pulled from linux and/or debian wouldn't apply there. | 06:44 |
systemdlete | Your junk pile and my junk pile probably look a lot alike. | 06:46 |
systemdlete | there is a mainboard vga -- is it possible to enable it again? It seems to shut off when another card is inserted | 06:47 |
* systemdlete switching to testbox... | 06:47 | |
gnarface | in theory, but check the bios options | 06:48 |
gnarface | didn't we go over this already and there were no bios options, though? or was that someone else? | 06:48 |
systemdlete2 | bios -- what I figured, but didn't want to go through all that... | 06:48 |
plasma41 | systemdlete: see also https://lists.x.org/archives/xorg-announce/2019-July/003015.html | 06:48 |
systemdlete2 | someone else, bc I don't recall | 06:48 |
gnarface | check the bios options, and if it's a really old board it might be worth checking for bios updates too | 06:49 |
gnarface | (make sure only to get them from the original manufacturer's site though) | 06:49 |
systemdlete2 | I've updated the bios as far as it can go (v17.17) | 06:49 |
gnarface | oh good | 06:49 |
systemdlete2 | thakns for the reminder | 06:49 |
systemdlete2 | not some fake cia front, huh? | 06:50 |
gnarface | huh? | 06:50 |
gnarface | oh, right, yea like fileshack. fuck those guys | 06:50 |
systemdlete2 | Yeah. I'm very careful about that. (you said only get bios updates from orig mfr) | 06:50 |
systemdlete2 | ok, taking the testbox down for bios config... | 06:51 |
systemdlete2 | I enabled internal graphics. But the system still boots on the same video card (1st pci video card,not the s3) | 06:57 |
systemdlete2 | do I need a cold boot? | 06:57 |
systemdlete2 | nvm. | 07:00 |
* systemdlete2 wants to kill himself. (please do not inhibit this impulse) | 07:00 | |
systemdlete2 | I forgot to move the cable over... | 07:00 |
systemdlete | gnarface: When I listed the hardware on the testbox, there was no specific info on video. That is prob why I'm not getting video on that output -- maybe there is no driver for it | 07:08 |
systemdlete | inxi -F does not list it; only the PCI card | 07:08 |
systemdlete | help me choose which pos card to use: Trident card dated 1994 or creative card dated 1999 (prob pref, if one can call it much better). | 07:13 |
systemdlete | There is a junk shop near me. The guy has bins full of video cards. | 07:13 |
systemdlete | I'm pretty sure they are dated as well. | 07:13 |
systemdlete | OR | 07:14 |
systemdlete | I could be patriotic and buy a new video card online and help keep "fantastic American businesses" like Amazon plenty flush. | 07:14 |
systemdlete | It would probably cost less than the electricity it is taking to sort this out. | 07:15 |
systemdlete | the rest of the cards here are either gaming cards without a vga port, or agp (no agp slot on this board). | 07:16 |
systemdlete | In fact, after much reflection (about 3 seconds since realizing this), I will do that. | 07:17 |
systemdlete | [crowd cheers with a collective sigh of relief] | 07:18 |
onefang | lol | 07:18 |
systemdlete | lol | 07:19 |
onefang | I did wake up this morning to 300 messages about this, 302, 314, ... | 07:19 |
meep_____ | Where do you buy gpu cards for linux that aren't buggy pieces of shit and requiring special snowflake drivers? | 07:19 |
systemdlete | I might give this creative card a shot. It's 21 years since mfr, but if it saves me my next order of Buffalo Wild Wings... | 07:20 |
systemdlete | Actually, I sort of inherited a lot of it, but some I have collected over the years. | 07:20 |
systemdlete | (trying to remember) | 07:20 |
systemdlete | I used to "shop" at ReCyclePC in Seattle when I lived there back in the 1990s/early 2000s | 07:21 |
systemdlete | Then I lived with a widow whose husband had been into all sorts of electronics (not to mention 48 other hobbies) | 07:21 |
systemdlete | He made a lot of stuff himself -- you should see some of the adapters he made... | 07:21 |
onefang | Ah, so some of this stuff was old unsupported stuff people got rid of twenty years ago. | 07:21 |
systemdlete | (I don't use them.) | 07:21 |
systemdlete | not exactly. | 07:21 |
systemdlete | Some of this stuff was relatively recent when I obtained them. | 07:22 |
systemdlete | The creative board was relatively new when I got it c. 2005 or so | 07:22 |
systemdlete | but they followed me over the years. | 07:22 |
systemdlete | I tried to make them go away, honest. | 07:22 |
systemdlete | People sometimes gave me their old computers so I could use them for testing and to play with. | 07:23 |
systemdlete | Anyway... | 07:23 |
onefang | Telling a graphics card "Oy, you, you are old enough to drive, time to go and live on your own" doesn't work that well. | 07:23 |
systemdlete | no | 07:24 |
systemdlete | (funny, onefang!) | 07:24 |
onefang | Though if they where old enough to drive, they wouldn't need drivers. | 07:24 |
systemdlete | I mean, I've tried to collect stuff for recycling, then I forget, and when I'd move, I didn't think to toss them | 07:25 |
systemdlete | (we should take this to fork, btw) | 07:26 |
systemdlete | I don't think this mainboard will let me do it. BIOS allows internal or pci-e only. It is silent on plain pci video cards. And having a pci and a pci-e card would seem simple for linux, but maybe the board gets in the way a bit | 07:46 |
systemdlete | The option says PRIMARY Graphics Adapter (my emphasis) so I figured maybe there could be a secondary, and so forth. | 07:49 |
gnarface | systemdlete: yea the issue is that it's entirely up to the bios whether to support it or not. if it doesn't do it, it doesn't do it... but note that it is worth checking to see if internal only works when there's no add-on card | 10:06 |
systemdlete | yeah, I thought of running that test. It means removing the card though. I am doing some testing on adelie atm; as soon as I can shut it down, I'll remove the card and see. One thing for sure: It seems the internal does not work at all if there is a card also | 10:08 |
gnarface | systemdlete: as for replacement cards, i dunno anything about the creative labs video cards at all, but the trident ones should be well supported - they're just cheap crap though | 10:08 |
gnarface | systemdlete: see if the bios options for the internal video card change when there's no other video card | 10:09 |
systemdlete | well, additional problem is it seems to be limited to just the pci slot(s), and I am not even sure that you can have both pci-e slots with a video card. | 10:09 |
systemdlete | I tried setting the bios to internal with the card still present. It auto switched to the card though; it was the only port getting output. | 10:10 |
systemdlete | I should have said, limited to just the pci-e slots (there's a pci slot, but it doesn't seem to be covered by the whatever logic they use). | 10:11 |
systemdlete | saddest part is that, after all this work, the board might just go blow up tomorrow. It is a rather old board, and I've moved it around a bit from case to case. | 10:12 |
gnarface | some of the pci slots may be paired with pci-e slots in a way that it's not supported to use both of the pair at once | 10:12 |
gnarface | yea i'm also wondering if maybe the onboard video was nvidia and it just burned out... | 10:12 |
gnarface | any way to tell? got the model & vendor name? | 10:12 |
systemdlete | iirc, I bought the board in 2014 or so. | 10:12 |
gnarface | the manual, what does it say? | 10:13 |
systemdlete | yeah. It's the 760GM-P25 (the board is marked 6531 I think, something like that) | 10:14 |
systemdlete | 7641 | 10:14 |
systemdlete | It's actually the 760GM-P34 (which matches my box it came in). I may have the wrong manual (for the p25) | 10:15 |
gnarface | the bios may also have some setting like "reset hardware configuration data" that you should check when you go in there before you save changes, whenever you change the add-on cards | 10:17 |
gnarface | AMD System Drivers for 7xx/8xx/SB7xx/SB8xx Series (except RS690, 740) < | 10:18 |
gnarface | Description | 10:18 |
gnarface | AMD System & VGA Drivers | 10:18 |
gnarface | i'm inferring from this ^ that the onboard video is something by amd | 10:19 |
gnarface | so maybe the radeon or amdgpu drivers would recognize it | 10:19 |
gnarface | but you need to be able to find it in the output of lspci first | 10:19 |
systemdlete | right, right | 10:19 |
gnarface | you saw this, right? https://us.msi.com/Motherboard/support/760GM-P34-FX | 10:20 |
systemdlete | yep | 10:20 |
systemdlete | been there many, many times now | 10:20 |
gnarface | it's not one of those ones that requires a special "apu" cpu to enable, is it? it looks too early for that but i'm not sure... | 10:22 |
systemdlete | no, i don't think so. I used it for years without the video card. | 10:22 |
gnarface | oh, so you've seen the onboard video work before | 10:23 |
gnarface | ok | 10:23 |
systemdlete | If you look at that manual, btw, you will see that there is no video spec in the manual! | 10:23 |
systemdlete | everything else, yes, but not the videospec | 10:23 |
gnarface | that's the other thing that leads me to question whether this is assuming you need a APU... | 10:23 |
systemdlete | thanks for all your help and especially the info you offer. | 10:27 |
systemdlete | I'm going to knock off for a while. It's 2:30am here and I've been pounding at this for at least 8 hours. | 10:27 |
gnarface | alright, lemme know how the test with no add-on cards works. | 10:28 |
systemdlete | I notice no one makes a dual vga output pci card. LOL why would they? | 10:28 |
systemdlete | I'm probably the last guy left on the planet who needs one. | 10:28 |
gnarface | hmm, they certainly must have existed, but i'm sure they were much more rare... | 10:28 |
systemdlete | pci-e I meant | 10:29 |
gnarface | oh, yea, no it's hard enough to even find dvi ones anymore | 10:29 |
gnarface | everything is going hdmi | 10:29 |
systemdlete | This much I know: The card must go in one of the pci-e slots. The only choices in the bios are internal and pci-e | 10:29 |
gnarface | interesting | 10:29 |
systemdlete | would linux recognize the pci card though? I saw it listed in ixni | 10:30 |
systemdlete | inxi | 10:30 |
gnarface | can you paste your lspci output to paste.debian.net? | 10:30 |
systemdlete | or, I thought I did | 10:30 |
gnarface | ideally i'd like to see it with and without the add-on card | 10:30 |
gnarface | might not matter though depending on what we see | 10:31 |
gnarface | actaully you can just /msg it to me even | 10:31 |
systemdlete | sure. | 10:32 |
xrogaan | I don't know if it's related to consolekit, so I'll be switching back to elogind to figure it out, the .xsession-errors file isn't being created or rotated for some reason, and it annoys me so much. | 11:08 |
xrogaan | (beowulf) based on /etc/X11/Xsession, a xsession-errors file should be created following: http://dpaste.com/0CDVCVV | 11:09 |
xrogaan | it is not, and no message gets logged anywhere. | 11:09 |
xrogaan | Oh, weird, I just found out that there is a .xsession-errors.log, no idea how it gets created | 11:11 |
gnarface | rsyslogd perhaps? | 11:12 |
gnarface | or the X server itself | 11:12 |
gnarface | dunno though for sure | 11:13 |
xrogaan | slim starts the session this way: login_cmd exec /bin/bash -login /etc/X11/Xsession %session | 11:15 |
xrogaan | I don't know why there is a .log at the end of the xsession-errors | 11:15 |
xrogaan | I can't find any reference to that file anywhere. | 11:16 |
gnarface | yea that's what made it look like a rsyslogd thing to me, but then i remembered the Xorg.0.log ends in .log too... | 11:16 |
gnarface | but it's just a guess | 11:16 |
xrogaan | the rsyslog is minimal here | 11:17 |
xrogaan | rsyslog conf* | 11:17 |
xrogaan | I remember why I don't want to go back to elogind: xfce is crappy under it and we won't get an update to xfce until next debian stable. | 11:19 |
xrogaan | so if I change, I change of WM too | 11:19 |
gnarface | maybe enlightenment will behave better for you? | 11:19 |
xrogaan | no, it's a crappy desktop | 11:35 |
xrogaan | the defaults are so out of my norms that it's just too much work to figure out | 11:35 |
xrogaan | Like the mouse cursor automatically being placed on the button of a message window. | 11:36 |
gnarface | i admit that's annoying and i turn it off too | 11:45 |
gnarface | but there are so many settings... | 11:46 |
gnarface | it's very flexible | 11:46 |
kgkknll | ping | 16:30 |
MSIVITALER^3I_ | hey there how would one suspenf in openrc would loginctl suspend work isnt that sytemd command? | 16:31 |
gnarface | there's probably a bunch of ways to do it | 16:32 |
gnarface | something would be tied to acpid probably... | 16:32 |
MSIVITALER^3I_ | what would be the easiest with a one line coomand | 16:33 |
MSIVITALER^3I_ | Im pretty green to openrc based systems, but Im glad I switched after a bit of research I dont know how something like systemd come to existence | 16:34 |
gnarface | i'm sorry i don't know openrc either and i'm not sure it is even the part of the system that handles suspend | 16:45 |
gnarface | i use sysvinit which really does not | 16:46 |
gnarface | suspend would be handled by acpid and your window manager... | 16:46 |
gnarface | like, there should be something in the gui to click on | 16:46 |
gnarface | there isn't? | 16:46 |
gnarface | if you have a suspend button on the keyboard it might work too | 16:47 |
gnarface | see if you have a package called pm-utils, and try the pm-suspend script in there | 16:47 |
gnarface | if all else fails | 16:47 |
MSIVITALER^3I_ | gnarface: cheers found answer on acpid arch wiki | 16:53 |
gnarface | MSIVITALER^3I_: cool, glad you found it | 16:59 |
MSIVITALER^3I_ | gnarface: thk for the help | 17:29 |
gnarface | np | 17:32 |
meep_____ | gnarface: doesn't pm-suspend rely on policykit? | 20:01 |
meep_____ | How did we suspend before the redhatification? | 20:01 |
meep_____ | I remember back when shutdown -P NOW worked | 20:02 |
meep_____ | And then i had to change to poweroff | 20:02 |
meep_____ | Or shutdown -H NOW before ACPI | 20:02 |
mason | policykit? no | 20:03 |
gordonDrogon | once upon a time it was: sync;sync <pull the plug> | 20:10 |
gnarface | meep_____: i have it here without policykit, i can't claim to know if it works right or not though | 20:29 |
meep_____ | This https://upload.nuegia.net/ac42c189-115c-432c-8a5a-0943d335496a/screenshot.png | 20:30 |
meep_____ | This is bullshit | 20:30 |
meep_____ | It looks like an iphone | 20:30 |
meep_____ | Gtk+3 is so bad | 20:30 |
meep_____ | Now I have to fork notifyd too | 20:31 |
avbox | I want to setup a new mail server (smtp/587, imap/143, webmail, spam). What can you recommend? Is there a howto for Devuan? | 20:43 |
meep_____ | tom: fixed https://upload.nuegia.net/6b2d9050-9068-44be-a907-bab754f0f4fd/screenshot.png | 20:54 |
tuxd3v_ | meep_____, what is the toolkit used? | 21:00 |
meep_____ | Gtk2 | 21:00 |
meep_____ | avbox: OpenSMTPd, Dovecot, Spamassisin, and SquirrelMail | 21:01 |
meep_____ | tuxd3v_: I reverted the GTK+3 version back to gtk2 | 21:02 |
avbox | meep_____: SquirrelMail, is there any further development' | 21:02 |
meep_____ | Idk | 21:02 |
meep_____ | I don't use webamil | 21:02 |
meep_____ | I have programs for email, like mutt and claws-mail | 21:03 |
mason | gordonDrogon: Sometime trace the origin of "sync; sync" - I did recently and I was surprised. | 21:32 |
furrywolf | sync sync... that's the new york prison where they stick dirty write caches, right? :P | 21:37 |
gordonDrogon | heh... my first unix was v6 on a pdp11 - it was just the "done thing" ... | 21:38 |
mason | gordonDrogon: I still do it reflexively, but evidently it hasn't actually been a thing for decades. | 21:45 |
gordonDrogon | I know... old habits though. | 21:45 |
gordonDrogon | arguably the start of bloat - why write a shutdown/reboot when you can 'wall' everyone, sync;sync, then flip the run switch :) | 21:46 |
expert975 | That's how I poweroff my raspberrypi: sync and unplug :) | 21:53 |
aitor | meep_____: good job | 22:00 |
meep_____ | » [13:46:45] <gordonDrogon> arguably the start of bloat - why write a shutdown/reboot when you can 'wall' everyone, sync;sync, then flip the run switch :) | 22:01 |
meep_____ | to allow the daemons to stop | 22:01 |
meep_____ | Your forgetting a few steps | 22:02 |
ukine | i was taught after unmounting/disk ops to do it 3 count 'em 3 times | 22:02 |
meep_____ | Yes you wall | 22:02 |
ukine | syncing | 22:02 |
meep_____ | Wall, stop daemons gracefully, remount disks read-only, sync, then halt | 22:02 |
meep_____ | Then ask acpi to turn off the power | 22:02 |
gordonDrogon | the 2nd sync supposedly blocked until the first one was done, but I never did a 3rd. as for daemons... the small, lonely pdp11 only ran lpd that I recall, but memory isn't what it was .. | 22:11 |
meep_____ | I don't think the second sync is needed | 22:12 |
meep_____ | Is it? | 22:12 |
meep_____ | As far as i know that's like people who press the off button on their light switches because they have ocd or something | 22:14 |
meep_____ | And want to really make sure it's off | 22:15 |
meep_____ | *light switches 5 times | 22:15 |
expert975 | REISUO | 22:37 |
meep_____ | Ebola? | 22:38 |
initfr33d0m | hello | 23:45 |
Guest72323 | \o | 23:48 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!