| gnarface | kreyren__: you can tell qemu to just use a particular number of host cpu cores directly without actually emulating them | 00:15 |
|---|---|---|
| kreyren__ | gnarface, oh i see | 00:16 |
| gnarface | it's not as efficient as something like the linux-vservers kernel patch, but it does mitigate a lot of the emulation overhead, especially if you have working hardware kvm/iommu acceleration | 00:16 |
| arno11 | kreyren__: your sources.list should be: | 00:17 |
| arno11 | https://paste.debian.net/1324282/ | 00:17 |
| arno11 | ignore the last line (experimental repo) | 00:17 |
| gnarface | arno11: just fyi, Devuan ops prefer people would use deb.devuan.org (a dns round-robin of mirrors of pkgmaster, which also contains pkgmaster) instead of just using pkgmaster directly, to reduce their bandwidth load | 00:19 |
| gnarface | not all the mirrors are super great on bandwidth themselves though, so i don't blame you | 00:19 |
| gnarface | in a fair world, the only difference should be a minor periodic delay of mirror replication during updates | 00:20 |
| gnarface | people who have problems with the dns-rr usually just pick the mirror off the list closest to them (http://deb.devuan.org/mirror_list.txt) | 00:21 |
| gnarface | i am not officially affiliated with Devuan, i just hang out in their channel | 00:22 |
| arno11 | fair enough :P | 00:28 |
| kreyren__ | arno11, ye i just added -devel thinking that's enough, trying to reboot now with these sources thanks for clarification | 00:31 |
| kreyren__ | gnarface, vservers? Hmm i am in the middle of deploying Xen infra-wide in nixium i wonder if it would help for this | 00:32 |
| gnarface | kreyren__: did you "apt-get update && apt-get dist-upgrade" after changing the sources.list? | 00:32 |
| kreyren__ | gnarface, yep | 00:32 |
| kreyren__ | wait no, i did full-upgrade | 00:32 |
| kreyren__ | afaik it does upgrade and dist-upgrade | 00:32 |
| gnarface | i am not sure | 00:33 |
| kreyren__ | https://salsa.debian.org/apt-team/apt/-/blob/1.6.12/cmdline/apt.cc#L74 apparently same command | 00:33 |
| kreyren__ | {"full-upgrade", &DoDistUpgrade, _("upgrade the system by removing/installing/upgrading packages")}, | 00:33 |
| kreyren__ | seems that full-upgrade is the modern and the dist-upgrade is obsolete and kept there for legacy and muscle memory? | 00:34 |
| gnarface | http://linux-vserver.org/ if you're interested, and their channel is over on OFTC, but if you're deploying it for work Xen might be a safer choice just because it's got current support. the linux-vserver project has kinda fallen on hard times and doesn't have as much current kernel support, and also isn't really for emulation, so running non-linux guests isn't an option | 00:35 |
| kreyren__ | like the problem on xen is that it's arm support is experimental as well | 00:36 |
| gnarface | but, it's lightweight enough you could conceivably run it on your actual phone with no overhead | 00:36 |
| kreyren__ | interesting | 00:36 |
| kreyren__ | to what benefit considering the limited processing resources of N900? | 00:36 |
| gnarface | maybe none :) i mostly use it for my quake server farm | 00:37 |
| kreyren__ | heh | 00:37 |
| gnarface | it's good for if you're running internet-connected services you're afraid might get hacked and you want some way to keep any potential breaches from being used to breach other services | 00:38 |
| gnarface | it's got a lot of functional overlap with something like bsd "jails" | 00:38 |
| gnarface | in modern linux kernels, the various "namespaces" features have started to largely obsolete the old vserver features, but still don't quite reach the same level of feature-complete paranoia, especially with regards to the network layer | 00:39 |
| kreyren__ | NiXium is trying to manage that threat by using lokinet for intraweb comms and as a layer that deploys services for clearweb with PQS SSH that requires pubkeys and services in their invidual sandboxes on impermanent system | 00:39 |
| kreyren__ | rebooted the phone and it seems to work so far ^-^ | 00:40 |
| gnarface | cool! | 00:40 |
| kreyren__ | > Backend ring/tel/account0 | 00:40 |
| kreyren__ | instead of ofono | 00:40 |
| kreyren__ | The phone also seems much more responsive ^^ | 00:42 |
| gnarface | nice | 00:42 |
| arno11 | kreyren__: it is bit late for me but tomorrow i can provide you few tweaks to improve perfs a LOT :P | 00:45 |
| kreyren__ | arno11, thanks that would help a lot too, i would like to update the wiki | 00:45 |
| arno11 | zzz time | 00:49 |
| kreyren__ | arno11, gn~ <3 | 00:54 |
| sicelo | Wizzup: https://github.com/maemo-leste/libicd-network-wpasupplicant/pull/2 ... at least helps with https://github.com/maemo-leste/bugtracker/issues/730 and https://github.com/maemo-leste/bugtracker/issues/729 | 10:55 |
| Wizzup | sicelo: cool, ty, will look when I'm home in ~5 hours | 11:18 |
| Wizzup | btw, looks like phoenix/leste is down again for a few hours still, maintenance/renovation is almost done at least... | 11:19 |
| Wizzup | will get a UPS tomorrow.. | 11:35 |
| Wizzup | sicelo: on the PR, do you need to free err2 if it is not null? | 11:36 |
| sicelo | ah yes i forgot to add it. how is err freed? (from tge preceeding GetAll dbus call) | 11:42 |
| Wizzup | I think g_error_free, but I'm sure there's other instances that file | 11:47 |
| Wizzup | so something like if (err2 != NULL) { WPALOG_ERR("Foo: %s" err2->message); g_error_free(error); ... } | 11:48 |
| Wizzup | or you might want to decide not to log the error if it's not important/fatal | 11:48 |
| Wizzup | or log it differently | 11:49 |
| Wizzup | bbl | 11:49 |
| sicelo | I'll add it next to the g_variant_unref's | 11:49 |
| freemangordon | you also can do g_clear_error(&error) | 11:52 |
| freemangordon | it also sets error to NULL | 11:52 |
| freemangordon | just a note, up to you what to use | 11:52 |
| sicelo | thanks | 11:53 |
| arno11 | using n900-pm script (off mode disabled), i get 44mA idle on 3.5g network now with 6.6, giving at least 30 hours idle. not bad | 13:36 |
| arno11 | but once i plug any usb charger and unplug it, i get 10-12mA more | 13:38 |
| arno11 | i didn't remember this 'old' bug | 13:38 |
| arno11 | maybe i should open an issue | 13:39 |
| arno11 | wow 46mA with 3.5g connected | 13:41 |
| arno11 | maybe one day someone will find the trick to hit RET. we are not so far imo | 13:43 |
| sicelo | s/RET/OFF/ | 14:05 |
| sicelo | that'll be a great day! | 14:05 |
| sicelo | but yes, really amazing to get this far without OFF mode | 14:05 |
| Wizzup | sicelo: we don't hit RET either I think | 14:27 |
| Wizzup | on n900 | 14:27 |
| Wizzup | enabling it currently causes funky stuff | 14:27 |
| arno11 | yes still the same funky stuff :) | 14:53 |
| arno11 | and indeed we still even don't hit RET | 14:54 |
| Wizzup | ok, phoenix/leste should come back online ina bit | 14:56 |
| Wizzup | interesting, on this 3g network the mtu of 1500 is too high and I can't ssh to my home machine, but dropping it to 1200 makes things work | 15:45 |
| sicelo | the mtu is 1500? | 15:47 |
| Wizzup | on wwan3 yeah | 15:47 |
| bencoh | some router probably doesn't allow fragmentation and/or drops frag packets | 15:48 |
| bencoh | I'd guess actual mtu is higher than 1200 though :) | 15:48 |
| Wizzup | yeah, was testing mostly, but still | 15:49 |
| freemangordon | Wizzup: hmm something's not ok with the autobuilder | 17:26 |
| Wizzup | yes, will fix | 17:27 |
| Wizzup | like 60mins | 17:27 |
| freemangordon | ok | 17:27 |
| Wizzup | just changed ips and I need to make em static | 17:28 |
| Wizzup | btw what did you need xkbmap wise from me | 17:28 |
| freemangordon | please, spin a build of mafw-tracker-source chimaera-devel | 17:28 |
| Wizzup | ok | 17:28 |
| freemangordon | no idea | 17:28 |
| Wizzup | shift+space says it switchen for me but ofc hw layout does not | 17:28 |
| freemangordon | what about xkbmap? for cyrillic? | 17:28 |
| Wizzup | yeah | 17:28 |
| freemangordon | whatever, not now, I am packing stuff :) | 17:29 |
| Wizzup | kay | 17:29 |
| freemangordon | just wanted to fix the rest of tracker-source issues | 17:29 |
| Wizzup | cool | 17:29 |
| freemangordon | hopefully I did, but well, lets see | 17:29 |
| freemangordon | at least sorting works and I moved everything to prepared statements and not clear strings | 17:34 |
| Wizzup | sweet | 17:34 |
| freemangordon | so now everything should be properly ordered in omp | 17:34 |
| freemangordon | bbl | 17:35 |
| dsc_ | so uhh | 20:45 |
| dsc_ | anyone know how to disable apparmor | 20:45 |
| dsc_ | at boot, just turn it off in general | 20:46 |
| dsc_ | ok ok ill write some apparmor rules | 20:54 |
| dsc_ | for Matrix I have join/create and 1:1 E2EE working in conversations | 20:59 |
| arno11 | dsc_: that's nice. able to send/receive or? | 21:27 |
| Wizzup | dsc_: sudo aa-teardown | 21:42 |
| sicelo | i wasn't even aware we're using apparmor | 21:51 |
| sicelo | Wizzup: just tried the wireguard plugin for the first time. looks like there's a small usability issue - as soon as you enable it, status shows "Connected" even if there's no connection yet | 21:55 |
| sicelo | could lead to false sense of security, at least for someone using it for that kind of thing | 21:55 |
| sicelo | but my internet is super bad this evening, so i don't know if this is PEBKAC somewhat | 22:07 |
| dsc_ | arno11: for initial support just text messaging | 22:12 |
| dsc_ | the file transfer stuff for e.g xmpp, matrix, telegram etc. is probably doable but requires significant changes | 22:12 |
| dsc_ | ill do vcard's first | 22:13 |
| dsc_ | (avatars) | 22:13 |
| Wizzup | sicelo: that just means the wg interface is up | 22:16 |
| Wizzup | it doesn't know whether the wg actually works or not | 22:16 |
| Wizzup | but good point I suppose.. | 22:16 |
| sicelo | in my case, i don't see an interface anywhere either | 22:21 |
| arno11 | dsc_: E2EE messaging is already really cool | 22:24 |
| dsc_ | :) | 22:25 |
| Wizzup | sicelo: strange, that is not great indeed | 22:25 |
| Wizzup | sicelo: we don't use apparmor on arm it seems ,but we do on amd64 | 22:25 |
| Wizzup | I think maybe just because we don't have it enabled in kernel on arm | 22:25 |
| dsc_ | Wizzup: could you create "olm" repo @ https://github.com/maemo-leste-upstream-forks (no hurries) | 22:26 |
| dsc_ | we need a more recent version of https://gitlab.matrix.org/matrix-org/olm | 22:26 |
| Wizzup | freemangordon: done @ jenkins build | 22:28 |
| Wizzup | dsc_: done | 22:29 |
| dsc_ | ty | 22:30 |
| sicelo | Wizzup: i've added the missing g_free now. made it non-fatal, since if we somehow fail with SignalPoll(), we fallback to the 'old' method of using the BSS properties | 22:33 |
| Wizzup | freemangordon: so if you do this: | 22:34 |
| Wizzup | setxkbmap -layout "us,bg" | 22:34 |
| Wizzup | setxkbmap -option grp:alt_shift_toggle,grp_led:scroll -variant ,phonetic | 22:34 |
| Wizzup | then you can use the 'caps lock' and 'sym' key on the d4 to switch hw keyboard input | 22:34 |
| Wizzup | but you might lack ч ш щ :D | 22:35 |
| Wizzup | sicelo: agree, ok, cool, shall I merge and do a -devel build? | 22:35 |
| sicelo | up to you. it's working great for me :-) | 22:36 |
| sicelo | maybe yes @-devel, so it can get some testing, and we get feedback regarding any regressions (i don't see where they'd come from though) | 22:38 |
| Wizzup | we always try to do devel first | 22:41 |
| sicelo | it works so well (for me) that i now frequently open the status menu 'by accident' just to see the bars adjusting | 22:51 |
| Wizzup | :) | 22:52 |
| Wizzup | sicelo: I think it looks good to me otherwise, and the int16 vs int32 is also in line with the icd2 headers / dbus signature? | 22:56 |
| sicelo | SignalPoll's rssi signature is gint32, while BSS.Signal is gint16. both are reporting same thing though. i think it's some sort of bug wpa side | 23:00 |
| arno11 | Wizzup: do you have few min to check @n900 headset PR ? | 23:03 |
| Wizzup | arno11: link? :) | 23:12 |
| Wizzup | sicelo: I think we need to match the dbus signature of icd2 | 23:12 |
| Wizzup | I'll see if I can figure out what it is supposed to be | 23:12 |
| sicelo | makes sense, yes | 23:14 |
| arno11 | Wizzup: leste-config/pulls/48 | 23:17 |
| arno11 | *pull | 23:18 |
| Wizzup | arno11: looks good, does regular headphone still work | 23:22 |
| sicelo | yes, also looks good to me (the headset stuff) :-) | 23:22 |
| Wizzup | building | 23:23 |
| arno11 | regular headphone still works (in fact jackcontrol makes no diff between headset and headphone) | 23:23 |
| arno11 | cool :) | 23:24 |
| sicelo | ideally the volume would move to BootSequence, but i guess that's something for future work/investigation. alsaucm folk don't recommend having volumes changing in Enable/Disable sequences, to respect user settings | 23:24 |
| arno11 | yes, true | 23:24 |
| arno11 | there is still a bit of work on ucm but works already pretty fine. hope we'll be able to add bluetooth :P | 23:26 |
| arno11 | *one day | 23:27 |
| freemangordon | arno11: hmm, no diff HS vs HP? not good | 23:28 |
| Wizzup | sicelo: does it have to be 32 bit int, given that in dbus you do G_VARIANT_TYPE_INT16 | 23:29 |
| Wizzup | oh wait that's the old path | 23:29 |
| arno11 | freemangordon: indeed but you can check through alsactl monitor: no diff when you plug HS, HP or TV out | 23:30 |
| freemangordon | Wizzup: please, do me a favor while I am MIA - find why OMP/mafw loses audio and needs PA restart | 23:31 |
| Wizzup | I will try | 23:31 |
| Wizzup | sicelo: so it's icd2's link_stats | 23:31 |
| freemangordon | thanks | 23:31 |
| Wizzup | sicelo: ok its gint there | 23:32 |
| Wizzup | so 32 bit seems fine | 23:32 |
| Wizzup | arno11: you probably know this but n900 trs pins is different from regular/modern headsets | 23:33 |
| Wizzup | like a modern headset will make it think you plugged in a tv out | 23:33 |
| Wizzup | are you using the nokia headset, or? | 23:33 |
| arno11 | yep, using the nokia headset | 23:34 |
| arno11 | (and yes i know @specific n900 jack stuff) | 23:35 |
| Wizzup | sicelo: building for devel | 23:38 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!