tom_work | hello | 00:44 |
---|---|---|
tom_work | I'm having some trouble with Devuan. My sftp downloads are incredibly slow despite being on the same ISP | 00:44 |
tom_work | https://upload.nuegia.net/f9ebaa7e-0eba-4bd6-9c06-a3399cb2d5f9/screenshot.png | 00:44 |
tom_work | Are there any known causes for this or troubleshooting steps for fuse sftp? | 00:45 |
tom_work | 40 and 5kB/s seems a bit wrong | 00:46 |
tom_work | is there something wrong with the way fuse handles sparse files? | 00:49 |
gnarface | tom_work: can you reproduce the problem with scp? | 00:55 |
gnarface | there is every possibility there is a problem with the fuse driver in your particular kernel version that is specific to sftp, but it is far more likely the ISP is just indiscriminately throttling encrypted traffic | 00:58 |
tom_work | gnarface, the coffee shop i'm in is the same ISP I use at home, and both are using a static ip. I know this ISP not to be doing that and not have the capability to do that even if they wanted to. | 00:59 |
tom_work | I used to work there | 00:59 |
tom_work | so I'm pretty confident ruling that out | 00:59 |
tom_work | It's a small local ISP not a comcast | 00:59 |
tom_work | hmm. | 01:01 |
gnarface | tom_work: ok, that's fair. but every Comcast customer who ever ran into this problem told me the same thing | 01:01 |
gnarface | (just for the record) | 01:01 |
tom_work | scp does not seem to fluctuate as much | 01:01 |
gnarface | best thing you can do is try some other ports and tranfer protocols, see if you can narrow down the exact problem | 01:02 |
tom_work | I mean there is still a pretty significant pause between when each file starts up the transfer but not nearly as bad as fuse | 01:02 |
gnarface | also, if you can see the problem beteween two computers within the same LAN, and then reproduce that with two machines in a *different* LAN that would be a strong indicator that it is a kernel issue | 01:02 |
tom_work | with fuse the whole file manager locks up for a good 6 seconds if I list out a directory with more than 30 files in it | 01:02 |
tom_work | ok | 01:03 |
tom_work | is the fuse sftp implementation particular sensitive to network latency? | 01:03 |
gnarface | i don't know | 01:04 |
tom_work | is there a better unix-xattr native protocol I could use instead of sftp? | 01:04 |
gnarface | sshfs? | 01:05 |
gnarface | it also uses fuse | 01:05 |
gnarface | but i don't really know either | 01:05 |
tom_work | only thing I'm aware of is NFS but that sucks pretty bad outside of a lan or even over wifi | 01:05 |
gnarface | i always use scp | 01:05 |
gnarface | well, almost always | 01:05 |
tom_work | do you know if sftp://home over fuse uses the legacy SCP protocol or the current SFTP protocol? | 01:05 |
gnarface | it would depend on your client program | 01:06 |
tom_work | scp as in a simple in-band cat %s where sftp is an actual file transfer protocol of it's own over a ssh channel | 01:06 |
tom_work | thnar | 01:06 |
tom_work | thunar | 01:06 |
gnarface | i was not aware there was a significant distinction between scp and current sftp in how data is transferred | 01:08 |
gnarface | it was my understanding that the "legacy" protocol was the old sftp (some sort of tls+ftp hybrid operating on a non-ssh port) | 01:08 |
gnarface | encrypted file transfers are cpu-bound operations though... you're sure this isn't just being caused by high load or insufficient hardware resources at the server end? | 01:09 |
gnarface | and low RAM availability could impact all file transfers, not just the encrypted ones | 01:10 |
gnarface | anyway | 01:10 |
gnarface | so far i've not seen this issue except on windows with Putty's pscp.exe | 01:11 |
tom_work | gnarface, no what's going on is SCP is actually the deprecated protocol where it's a simple catting out of some file over the encrypted SSH channel sorta like oldstyle x-kermit transfer over serial links or dialin modems. where SFTP is it's own transfer protocol implemented by the /usr/sbin/openssh-sftpd binary. sftp is the new protocol your supposed to use and most operating systems just symlink scp to sftp because the syntax is | 01:11 |
tom_work | mostly the same | 01:11 |
tom_work | gnarface, oh yeah. the cpu-bound of my link is more ~970 megabits | 01:12 |
tom_work | I made a small bash script that tests SSH ciphersuite performance over loopback | 01:12 |
tom_work | funny thing is that chacha20 is faster than blowfish, but still slightly slower than hardware accelerate AES | 01:13 |
tom_work | chacha20 being both faster and stronger than blowfish | 01:13 |
tom_work | in software implementations | 01:13 |
gnarface | is this ascii you're testing? | 01:16 |
gnarface | if it's ascii with a stock kernel and there were known performance issues like this in FUSE, the backports kernel may have already integrated a fix | 01:17 |
tom_work | gnarface, it is indeed ascii | 01:38 |
tom_work | hey gnarface | 01:39 |
tom_work | since I'm using this as a desktop should I be using the rt kernel? | 01:39 |
tom_work | or laptop rather | 01:39 |
tom_work | I normally use rt kernels on workstations | 01:39 |
tom_work | linux-image-rt-amd64 | 01:41 |
tom_work | should desktops be using a rt kernel instead of the regular kernel build for interactive workloads not server workloads | 01:45 |
debdog | last time I've checked, like 10 years ago, RT did not improve user respons time, quite the opposite. could be irrelevant with modern, faster CPUs | 01:50 |
debdog | back then RT was useful for real time audio processing | 01:51 |
gnarface | tom_: no, don't use the RT kernel unless you have a specific reason to | 02:33 |
gnarface | some people used to advocate using them for gaming but the actual performance improvements were mostly speculative with a couple isolated corner cases that, as debdog says, should be pretty much unnoticeable now in the +GHz CPU speed range. all you're likely to do is introduce weird corner-case style bugs | 02:35 |
gnarface | tom_: recompiling the stock kernel for desktop-style latency however instead of the default that is tuned for servers actually might make a noticeable improvement in some UI experiences though | 02:36 |
gnarface | but note that a low-latency kernel is not the same thing as a "real-time" latency kernel. | 02:36 |
gnarface | they're fundamentally different approaches to processing jobs and data, though the details are above my head | 02:37 |
debdog | right, kernel option Timer frequency, Debian's kernel is set to 250 Hz. but 300 or even 1000 Hz are better for desktop systems | 02:45 |
nexgen | hello | 07:52 |
nexgen | please let me know how to show kernel output in a terminal from which qemu is executed | 07:53 |
nexgen | https://forums.gentoo.org/viewtopic-t-991672-start-0-postdays-0-postorder-asc-highlight-.html | 07:54 |
nexgen | they suggest to uncomment #s0:12345:respawn:/sbin/agetty -L 115200 ttyS0 vt100 | 07:54 |
nexgen | but I do not see it in the inittab | 07:54 |
nexgen | I have openrc | 07:54 |
nexgen | actually it is a modified qemu | 07:55 |
nexgen | actually it is a modified *heads linux | 07:55 |
nexgen | not qemu | 07:55 |
nexgen | why I cannot lynx to https://dev1galaxy.org/register.php?action=register | 09:09 |
nexgen | lynx https://dev1galaxy.org/register.php?action=register | 09:10 |
nexgen | Looking up dev1galaxy.org | 09:10 |
nexgen | Making HTTPS connection to dev1galaxy.org | 09:10 |
nexgen | Alert!: Unable to make secure connection to remote host. | 09:10 |
nexgen | lynx: Can't access startfile https://dev1galaxy.org/register.php?action=register | 09:10 |
debdog | lynx https://dev1galaxy.org/register.php?action=register --- works here, though might be because I've been there with a 'normal' browser and have the certificate | 09:19 |
nexgen | are lynx and other browser certs being shared? | 09:30 |
gnarface | yes | 09:33 |
gnarface | the package is ca-certificates, i think | 09:33 |
gnarface | though i think some of them come bundled with some of their own and add to the same directories | 09:33 |
gnarface | i could be wrong | 09:34 |
gnarface | you can add your own, too | 09:34 |
gnarface | (to /usr/local/share/ca-certificates/) | 09:37 |
nexgen | well, tried to install alpine from another host | 10:18 |
nexgen | works fine | 10:18 |
nexgen | the old host from which I tried is most likely hijacked | 10:19 |
nexgen | MITMed | 10:19 |
nexgen | it is right this one | 10:19 |
nexgen | btw :) | 10:19 |
nexgen | not sure why | 10:19 |
nexgen | is not it silly | 10:19 |
nexgen | I will switch later to I2P,etc. | 10:19 |
nexgen | fully anonymouse | 10:20 |
nexgen | paid from someone else | 10:20 |
nexgen | for hosting | 10:20 |
nexgen | just for some help work done | 10:20 |
nexgen | no mining, no bank, no paypal | 10:20 |
debdog | where the heck is https://packages.debian.org/buster/linux-image-4.19.0-5-amd64 in http://ftp.de.debian.org/debian/pool/main/l/linux/ ? | 10:31 |
debdog | can find the unsigned ones only | 10:31 |
nexgen | still cannot register on the forum | 10:38 |
nexgen | using lynx | 10:38 |
nexgen | The following errors need to be corrected before you can register: | 10:38 |
nexgen | * You answered incorrectly to the "Human or Robot" question, or you are a Bot! | 10:38 |
nexgen | Y | 10:38 |
nexgen | it asked about stable release name | 10:38 |
nexgen | I entered ascii | 10:38 |
nexgen | is it even possible at all to register from text mode lynx? | 10:39 |
nexgen | now asks about default login manager | 10:40 |
nexgen | I even do not know what is it | 10:40 |
nexgen | I use trinity | 10:40 |
nexgen | tdm | 10:40 |
debdog | prolly slim | 10:40 |
gnarface | lynx doesn't support javascript or css, so expect a bunch of stuff not to work | 10:41 |
onefang | "Human or robot" thing likely needs a graphic browser with javascript. They tend to follow your mouse as you move it to the "I'm human" button. | 10:41 |
nexgen | ungoogled cannot either | 10:41 |
onefang | So if your mouse movement looks like a human moved it, that's the real test. | 10:42 |
nexgen | what is the sense of this protection if it yet can be overcomed by zennoposter for example | 10:42 |
nexgen | it can move mouse | 10:42 |
nexgen | I am not goint to zennopost of course | 10:43 |
onefang | Don't ask me, ask Google or who ever wrote it. | 10:43 |
nexgen | just to ask some questions about init | 10:43 |
gnarface | well they might do other stuff too like burn a bunch of cpu cycles on wasted work to make sure it's not a tiny cloud bot swarm | 10:43 |
gnarface | (maybe mine some bitcoin on the side) | 10:44 |
nexgen | do you refer recapcha? | 10:44 |
gnarface | yea | 10:44 |
nexgen | I do not see any recapcha there in ungoogled | 10:44 |
nexgen | may be blocked :( | 10:44 |
nexgen | will try to change IP | 10:44 |
gnarface | are you blocking cookies? it likely needs those | 10:44 |
gnarface | at least one of them anyway | 10:45 |
onefang | It's an arms race, build better bot detectors, the opposition builds better bots, rinse and repeat. | 10:45 |
nexgen | most likely no | 10:45 |
nexgen | it could login to gmail and other recapcha protected sites | 10:45 |
nexgen | I tried an android device - Opera browser | 10:52 |
nexgen | no recapcha either | 10:52 |
nexgen | and I see a recapcha here on another site: | 10:55 |
nexgen | https://whois.domaintools.com/test.com | 10:55 |
nexgen | it is like recapcha blocked specifically for me specifically on your site | 10:56 |
nexgen | is it possible? | 10:56 |
nexgen | some black list | 10:56 |
nexgen | all my IPs | 10:56 |
nexgen | and browser cancases | 10:56 |
nexgen | *canvases | 10:56 |
gnarface | yes it's obviously possible in theory. i don't know if that's actually a feature or not though | 10:59 |
nexgen | I tried an android never used yet | 11:00 |
nexgen | and a computer of another owner | 11:00 |
nexgen | none of them dosplays recapcha | 11:00 |
nexgen | just a questions | 11:00 |
nexgen | I tried to answer it so many times already | 11:00 |
nexgen | may be it is easy to ask the question here | 11:01 |
nexgen | please help to configure qemu for a textual output from devuan | 11:01 |
nexgen | I do not see | 11:02 |
nexgen | #s0:12345:respawn:/sbin/agetty -L 115200 ttyS0 vt100 | 11:02 |
nexgen | in the inittab | 11:02 |
nexgen | using openrc | 11:02 |
nexgen | can I just add this line | 11:02 |
nexgen | not familiar with init yet | 11:02 |
nexgen | man inittab does not help much | 11:03 |
rrq | s/#s0/T3/ | 11:03 |
rrq | also needs args to qemu | 11:03 |
nexgen | please suggest qemu parameters | 11:04 |
rrq | add "-serial mon:stdio" | 11:04 |
rrq | then as it has started: use "C-a c" to shift to serial | 11:04 |
rrq | in the terminal | 11:04 |
nexgen | why when using Debian ISO for ARM with EFI | 11:08 |
nexgen | there was no need to C-a c | 11:09 |
nexgen | ? | 11:09 |
nexgen | https://pastebin.com/raw/Nn1q5QHZ | 11:09 |
nexgen | is it correct file | 11:09 |
nexgen | added the line at the end | 11:09 |
nexgen | T3:12345:respawn:/sbin/agetty -L 115200 ttyS0 vt100 | 11:09 |
rrq | possibly the qemu arg "-serial stdio" would make the terminal be serial only, i.e. no monitor | 11:13 |
nexgen | well, I did not see how kernel booted | 11:18 |
nexgen | but see a heads login screen | 11:18 |
nexgen | password not accepted though | 11:18 |
nexgen | and there was not any login request when booted with serial option | 11:19 |
nexgen | shifting to qemu terminal displays a general qemu commands screen | 11:21 |
nexgen | help displays a list of them | 11:21 |
rrq | you can set/change the passwd for the image eg via chroot | 11:24 |
nexgen | sure | 11:24 |
nexgen | but I thought I knew it | 11:24 |
nexgen | will retry | 11:24 |
nexgen | works finally | 11:29 |
nexgen | thank you very much! | 11:29 |
nexgen | your devuan is so cute | 11:29 |
nexgen | though tried alpine today first time | 11:30 |
nexgen | just to overcome lynx MITM | 11:30 |
nexgen | and it is SOOO FAST | 11:30 |
nexgen | apk add joe is instant | 11:30 |
nexgen | though may be it is just fast Linode host compared to my PC, heh | 11:31 |
nexgen | please suggest, can qemu console be enlarged to occupy the whole screen | 11:32 |
nexgen | many time on different KVM hostings I saw the same small console windows | 11:33 |
r3boot | nexgen: checkout https://firmit.wordpress.com/2008/11/30/qemu-running-fullscreen/ (first hit on google for qemu fullscreen) | 11:36 |
nexgen | thanks :), sometimes it is important to know if it is possible at all before googing | 11:39 |
nexgen | btw, do you all see recapcha on Devuan forum registration screen? | 11:41 |
r3boot | I think it's just being lazy, but thats just me ;) | 11:41 |
rrq | nexgen: to get boot output to serial, you'll need to have kernel boot parameter "console=ttyS0,115200" | 11:57 |
fsmithred | For the record: There is no javascript on the forum. | 11:58 |
rrq | nexgen: you ,ight also like qemu argument "-echr 0x1c" which changes its meta key from C-a to be C-\ .. much easier to use emacs :) | 12:00 |
rrq | s/,/m/ | 12:00 |
rrq | (fsmithred: I you look carefully at the source of a thread page you'll actually find some amount of javascript) | 12:04 |
fsmithred | I stand corrected. rrq, we don't use re-capcha, do we? | 12:47 |
rrq | nope. but a small collection of silly questions :) | 12:47 |
fsmithred | yeah, that I know | 12:48 |
systemdlete2 | I'm going to just guess, but the reason shutdown/reboot/hibernate/suspend are disabled in lightdm is due to systemd somehow? | 12:48 |
fsmithred | systemdlete2, is policykit-1-gnome installed? | 12:48 |
djph | systemdlete2: no idea, but sounds as good as any. | 12:48 |
systemdlete2 | it is installed, yes. | 12:49 |
fsmithred | ew | 12:49 |
fsmithred | that usually fixes it | 12:49 |
systemdlete2 | ew? | 12:49 |
systemdlete2 | I added a couple of rules by suggestion of a web page, but it did not help | 12:49 |
fsmithred | ew, like we're in deep shit | 12:49 |
systemdlete2 | why? | 12:50 |
systemdlete2 | maybe I am | 12:50 |
fsmithred | because the thing that usually fixes it didn't | 12:50 |
fsmithred | what p'kit backends are you using? | 12:50 |
systemdlete2 | not sure | 12:50 |
fsmithred | oh, you have conflicting desktops intsalled - that might be the problem | 12:50 |
systemdlete2 | all the ones installed by default if one does not specifically specify during install | 12:51 |
fsmithred | dpkg -l | egrep "consolekit|elogind|policykit|polkit|libpam" | 12:51 |
fsmithred | brb | 12:52 |
systemdlete2 | https://pastebin.com/kFVtZfsN | 12:54 |
fsmithred | which desktops? cinnamon and what? | 12:59 |
systemdlete2 | omg | 12:59 |
fsmithred | ? | 12:59 |
systemdlete2 | kde, gnome, cinnamon, mate, xfce | 12:59 |
fsmithred | lol | 12:59 |
systemdlete2 | hey, don't laugh at ME. Laugh at the guys who configured the system installation tool! | 13:00 |
fsmithred | I have no idea what will work in that case. | 13:00 |
systemdlete2 | Well, actually, I don't mind getting rid of them except for cinnamon, which is working really well on this tablet! :) | 13:00 |
fsmithred | the whole p'kit landscape is a clusterfsck | 13:00 |
fsmithred | that might simplify things | 13:01 |
systemdlete2 | add in systemd, then try to tear out systemd from that clusterfck, and you have an even messier clusterfck | 13:01 |
fsmithred | yeah, we know | 13:01 |
r3boot | well, in principle, lightdm will run some Xsession, and if you trace those scripts, you will get a clue as to how things are running now | 13:01 |
r3boot | another hint is to run 'export' in a terminal, to see which env vars are exposed | 13:02 |
systemdlete2 | http://paste.debian.net/1099653/ looks like cinnamon | 13:04 |
systemdlete2 | that's export in a regular users shell, not root | 13:04 |
lioh | is it already possible to do a dist-upgrade to beowulf? | 13:05 |
systemdlete2 | lioh: If running devuan in a VM, you could make a snapshot and try it. If it fails, you still can go back. | 13:06 |
r3boot | systemdlete2: ok, check, so you have the proper DBUS_* vars; Did you check with dbus-monitor what response you get back from dbus when you do a suspend? | 13:07 |
systemdlete2 | I've never tried suspend. | 13:07 |
r3boot | the reason those are disabled, is because your user (likely) does not have permisssions to talk to the ACPI/power related functions on DBUS. | 13:08 |
systemdlete2 | you mean the user running lightdm? Remember, there is no one logged in yet in lightdm! | 13:09 |
r3boot | yeah, that also includes the user under which lightdm runs itself. | 13:09 |
systemdlete2 | "also"??? | 13:10 |
systemdlete2 | who else could it be at that point? | 13:10 |
systemdlete2 | once I am logged in as my regular user, the shutdown/reboot/etc all work fine. | 13:10 |
systemdlete2 | really fine, I might add. | 13:11 |
systemdlete2 | But since there is no real user as such logged in in lightdm, the only such "user" could be the user id of the running instance of lightdm's greeter or the like | 13:11 |
systemdlete2 | looks like user "lightdm" is running dbus-launch, dbus-daemon, and at-spi-registryd | 13:14 |
systemdlete2 | but it is root that is running lightdm itself it appears to be | 13:14 |
fsmithred | maybe remove xfce and mate first, since those are the only ones that want consolekit | 13:14 |
systemdlete2 | And I have not reconfigured anything with regards to permissions foreither root or lightdm, to the best I can recall | 13:15 |
systemdlete2 | ok | 13:15 |
r3boot | Hmm, digging through my memory; So freedesktop based systems, talk via the dbus system bus to the org.freedesktop.upower service for power management. So the first thing to check would be, is upower installed? | 13:15 |
fsmithred | good call ^^^ | 13:15 |
r3boot | Once that's installed, there usually is some system group ('power' on archlinux) that you need to be a member of to be able to call power functions over dbus | 13:16 |
systemdlete2 | yes upower is installed | 13:16 |
systemdlete2 | ah! | 13:17 |
systemdlete2 | ok | 13:17 |
systemdlete2 | looks like root is running the upower daemon | 13:18 |
systemdlete2 | and I don't see any group related to "power" -- could it be something like "sys"? | 13:18 |
systemdlete2 | fsmithred: To remove xfce, do I remove both tasks -- desktop and live? | 13:19 |
fsmithred | huh? | 13:20 |
r3boot | Ok, and do you see the org.freedesktop.upower service on the system dbus? dbus-send --system --dest=org.freedesktop.DBus --type=method_call --print-reply /org/freedesktop/DBus org.freedesktop.DBus.ListNames | 13:21 |
fsmithred | remove the specific desktop metapackage: xfce4 | 13:21 |
fsmithred | I'm not sure what it is for mate | 13:21 |
fsmithred | mate-desktop | 13:22 |
systemdlete2 | yeah, if you remove "mate" it automatically does mate-desktop (it's smart?) | 13:24 |
systemdlete2 | ok, then I ran apt autoremove just to get rid of all traces | 13:24 |
systemdlete2 | hmmm. maybe i should do purge... | 13:25 |
systemdlete2 | ok all gone. | 13:25 |
systemdlete2 | (mate and xfce) | 13:25 |
systemdlete2 | r3boot: Looks like Upower is there | 13:29 |
fsmithred | maybe 'aptitude purge ~c' to get rid of leftover config files | 13:30 |
systemdlete2 | ok,, here is my theory: lightdm, a very nice dm, was originally written for sane linux systems. Then it gets rewired for systemd. Strip out systemd for much more sane devuan system. Leaves functionality for systemctl-dependent code hanging in the wind. | 13:31 |
systemdlete2 | (basically, but there is probably much, much more nuance...) | 13:31 |
fsmithred | I switched from lightdm to lxdm (with xfce) | 13:31 |
fsmithred | not for the problem you're having | 13:31 |
r3boot | systemdlete2: you might be right, but there /are/ replacements for that functionality afaik (eklogind, or what was it called), which should help out here | 13:32 |
fsmithred | elogind | 13:32 |
r3boot | if that doesnt work, you could resort to a different login manager which is not linked to systemd; Maybe slim or xdm or so | 13:32 |
fsmithred | slim needs consolekit | 13:33 |
systemdlete2 | fsmithred: did you mean "-c" or really meant "~c" | 13:33 |
fsmithred | tilde | 13:33 |
fsmithred | use 'search' instead of 'purge' and it will show you a list | 13:34 |
systemdlete2 | slim's config will be removed... | 13:34 |
systemdlete2 | done | 13:34 |
systemdlete2 | test? | 13:34 |
fsmithred | sure | 13:34 |
systemdlete2 | bbs... | 13:35 |
fsmithred | reboot | 13:35 |
systemdlete2 | nope. sorry, guys... :( | 13:38 |
systemdlete2 | I will say this much: The tablet is very usable at this point. Ive even fought victorious over the rotation problems. (I've got a button on the toolbar to switch it!) It even starts in the correct mode, and that includes the dm greeter. | 13:41 |
systemdlete2 | I really deeply appreciate all the help. I knoww I ask a lot of help | 13:41 |
systemdlete2 | I really had to dig for some of this info... googled some items for hours | 13:42 |
fsmithred | long shot: run pam-auth-update and make sure elogind is checked. | 13:44 |
systemdlete2 | yep. all are checked. | 13:46 |
fsmithred | all? | 13:50 |
fsmithred | I've got Unix and elogind checked. | 13:51 |
fsmithred | consolekit is not in my list. Is it in yours? | 13:51 |
systemdlete2 | there are 4 listed,the 2 you listed and gnome keyring and inheritable capabilities management | 13:54 |
systemdlete2 | all 4 are asterisked | 13:54 |
fsmithred | ok, I don't know what the management thing is | 13:54 |
fsmithred | mine also has Create home directory on login | 13:55 |
fsmithred | un-checked | 13:55 |
systemdlete2 | mine does not | 13:55 |
systemdlete2 | this is ascii, right? | 13:55 |
fsmithred | oh, no. | 13:55 |
fsmithred | I just looked at that on beowulf | 13:55 |
fsmithred | hang on | 13:55 |
systemdlete2 | that begs a conversation actually (beowulf) | 13:56 |
fsmithred | my ascii just has Unix and elogind | 13:56 |
systemdlete2 | I left some space on the emmc for other distros. It might be interesting to try out beowulf and see if performance and other support improves on this laptop. But that's for another day... | 13:56 |
fsmithred | this is on refracta - xfce installed in pieces, not metapackages | 13:56 |
fsmithred | you installed from regular installer isos or from live iso? | 13:57 |
systemdlete2 | netinstall | 13:58 |
fsmithred | ok | 13:58 |
systemdlete2 | pretty much chose all the defaults | 13:58 |
systemdlete2 | found I needed to do expert install due to system instability (original kernel). after install, i could update the kernel, which I did and it seriously improved stability all around. (4.19) | 13:59 |
fsmithred | new hardware needs new kernel | 13:59 |
systemdlete2 | those options to reboot etc in lightdm are greyed out. That tells me that it is the lightdm tool itself that is noting the inavailability of these functions. | 14:00 |
systemdlete2 | IOW, it is not merely that they don't WORK. It's that lightdm knows they don't or doesn't want the user to have access to them. | 14:01 |
fsmithred | yeah, but the source of the problem is likely not in lightdm | 14:01 |
systemdlete2 | I looked at the lightd config also. There were no noticeable options for governing access to the power functions. | 14:02 |
systemdlete2 | so, it must be "testing" for access and graying them out then? | 14:02 |
fsmithred | yeah, that sounds reasonable | 14:02 |
fsmithred | I get lost trying to understand this stuff | 14:03 |
systemdlete2 | me too. Just trying to remember all the minutia and administrivia... | 14:03 |
systemdlete2 | my brain overloads and wants to exit | 14:04 |
fsmithred | need food. bbl. | 14:07 |
systemdlete2 | same here | 14:11 |
dalme | Hi | 16:32 |
dalme | Is it possible to download an ISO file of the latest non-stable version of Devuan? | 16:33 |
dalme | I mean, development version | 16:33 |
enyc | pass ;p | 16:33 |
debdog | not entirely certain but http://pkgmaster.devuan.org/devuan/dists/unstable/main/installer-amd64/current/images/netboot/mini.iso could be one. might be outdated. preferably install ASCII and dist-upgrade to Beowulf | 16:41 |
fsmithred | if you select expert install on the netiso, I think you get a choice of suites | 16:44 |
fsmithred | or maybe I'm thinking of the mini.iso | 16:45 |
xet7 | Is there Devuan for m68k Amiga ? | 22:11 |
fsmithred | I don't think so | 22:11 |
xet7 | Ok | 22:11 |
fsmithred | you can check at files.devuan.org | 22:11 |
fsmithred | might possibly be with the embedded images | 22:11 |
fsmithred | but I don't recall anyone ever saying Amiga around here | 22:12 |
fsmithred | is it new or old? | 22:12 |
r3boot | ancient | 22:12 |
r3boot | amiga is pre-386 stuff | 22:12 |
r3boot | your best bet is NetBSD imho | 22:13 |
xet7 | Ok | 22:13 |
xet7 | Yes I think so too. OpenBSD dropped support. | 22:13 |
r3boot | yeah :/ | 22:14 |
r3boot | only Alpha and Sparc are still well supported under openbsd | 22:14 |
danielinux | hello, I'd like to use my sifive-hifive-unleashed to build devuan from sources. Can you point me to the right manual to kickstart a new arch? | 22:14 |
fsmithred | maybe a debian manual talks about that | 22:15 |
fsmithred | and we only have sources for about 5% of the packages. The rest is from debian unchanged. | 22:16 |
fsmithred | what arch are you planning to do? | 22:16 |
danielinux | riscv64 | 22:17 |
danielinux | board is https://www.sifive.com/boards/hifive-unleashed | 22:17 |
danielinux | debian is 'partially ported', and there are riscv64 packages for it | 22:17 |
danielinux | https://wiki.debian.org/InstallingDebianOn/SiFive/HiFiveUnleashed | 22:18 |
fsmithred | you probably want to talk to parazyd | 22:18 |
gnarface | oh but i really want there to be a m68k devuan. when was the last time there was a mainline kernel buildable for m68k? | 22:18 |
fsmithred | and you might find him in, uh, I think it's #devuan-arm | 22:18 |
danielinux | we might need a #devuan-riscv soon :) | 22:19 |
fsmithred | I think there are a few people who would like that | 22:19 |
fsmithred | maybe Centurion_Dan would be interested, too, but I know he's pretty busy with other stuff right now. | 22:19 |
gnarface | i feel like it must have been sometime before kernel 2.5 | 22:21 |
gnarface | i mean 2.4 | 22:21 |
gnarface | i only remember hearing about it | 22:21 |
danielinux | I think I might just need the right pointers, the idea is to compile packages natively, it'll take a while but it should go much smoother than cross-compiling. For now I'm bootstrapping a debian to get build-essential on board | 22:22 |
r3boot | gnarface: hmm, from the looks of it, atleast debian still has some m68k build servers running.. The wiki seems pretty dead tho | 22:22 |
r3boot | this seems to be their last update: https://people.debian.org/~tg/f/m68k/2015-Jan/ | 22:25 |
danielinux | fsmithred, thanks for the ptrs | 22:26 |
fsmithred | yw | 22:26 |
gnarface | r3boot: 2015, kernel 3.6, that's not nearly as abandoned as i had feared... | 22:40 |
Centurion_Dan | fsmithred, gnarface: about m68k... if we have access to a buildhost, we will build it. | 22:40 |
gnarface | kernel 3.16* | 22:40 |
gnarface | Centurion_Dan: now i'm kicking myself for turning down that Amiga a few years back | 22:41 |
Centurion_Dan | I think that the question is, how many m68k's are out there that are or would be put to real world use, and are there new machines being built on that architecture? | 22:42 |
gnarface | yes there is actually | 22:43 |
gnarface | i was just trying to dig up the link... | 22:43 |
r3boot | The biggest question is, who knows enough about m68k and is willing to backport kernel fixes imho | 22:43 |
gnarface | someone made a single-chip 500 core version of them for build farms and hosting | 22:43 |
gnarface | it's a after-market knock-off (not by motorola) but supposedly it rocks and had 100% compatiblity | 22:43 |
gnarface | compatibility | 22:43 |
gnarface | damn i can't spell today | 22:43 |
Centurion_Dan | Ok. Then it may be worth looking at. But it's likely going to require some work, and should be done in close coordination with Debian. | 22:44 |
gnarface | i'm trying to dig up the link | 22:44 |
Centurion_Dan | Is there real interest in having Debian for it. | 22:44 |
Centurion_Dan | s/Debian/Devuan/ | 22:45 |
r3boot | Personally, I'd check to see how many ppl would potentially be interested before embarking on such a project tbh | 22:46 |
r3boot | there arent that many amiga's available, and they are way too slow for modern-day usage, so the usebase might be pretty small to non-existent | 22:46 |
Centurion_Dan | Exactly r3boot, I mean real interest, in that someone will fund some time and hardware... | 22:46 |
gnarface | i would only honestly suggest it was worth it if it could be easily built without heavy kernel patching | 22:46 |
gnarface | but i do think this is one of those "if you build it they will come" scenarios | 22:46 |
Centurion_Dan | riscv64 is a no brainer, and we should get some hardware for that. | 22:47 |
r3boot | mja, hence my remark about netbsd/m68k. That is actively maintained and has xs to pkgsrc | 22:47 |
Centurion_Dan | We also need another arm64 builder.... | 22:47 |
gnarface | https://arstechnica.com/gadgets/2017/05/the-a-eon-amiga-x5000-reviewed-the-beloved-amiga-meets-2017/ | 22:49 |
gnarface | i think this is what i was thinking of, but maybe it is not as relevant as i thought | 22:49 |
gnarface | it seems to be powerpc based | 22:50 |
gnarface | but they do claim full support for all the legacy software due to the fact it's a custom designed hardware emulator chip | 22:50 |
gnarface | only 16 cores not 500 like in my dreams though ;) | 22:50 |
gnarface | and this was as recent as 2017 and they're claiming it booted Linux then | 22:50 |
danielinux | Centurion_Dan, I'm willing to lend out my unleashed to kick-start the arch | 22:51 |
r3boot | gnarface: isnt demo coding on amiga's way more fun? :P https://ada.untergrund.net/ | 22:51 |
r3boot | .15 | 22:51 |
gnarface | well it sure looks more fun anyway | 22:52 |
gnarface | https://games.slashdot.org/story/18/06/17/1942243/new-commercial-amiga-500-game-released | 22:52 |
danielinux | although the reason why I am interested is to make an embedded server for myself, so in the end it might be indeed useful if you have one in the farm | 22:52 |
gnarface | ^ 2018, an commercial game release for Amiga | 22:53 |
gnarface | the platform is dead like that Monty Python sketch | 22:53 |
danielinux | but for sure there'll be more and more interest to run on RV64 | 22:53 |
gnarface | it keeps shouting "i think i'll get better!" | 22:53 |
gnarface | "i feel happeeeee!" | 22:53 |
James1138 | ...that game would not move if you put 4000 volts through it ...it is passed on ...it has joined the choir invisible | 22:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!