tmlind | i just do this in my /etc/acpi/kp_slide_close script: kill -SIGSTOP $(pidof mosh-client) | 07:50 |
---|---|---|
Wizzup | I thought of that but just assumed it would actually work well ;) | 08:00 |
Wizzup | (it=mosh) | 08:00 |
buZz | sooo, did we change repo/branch for 5.18.9 for mapphones? | 16:00 |
buZz | or can i just git pull the wip-n900 one ? :P | 16:00 |
buZz | i guess the former | 16:01 |
buZz | btw, recompiling mosh kept failing on the tests? maybe the tests freak out on my high timeouts | 16:01 |
buZz | gotta figure out how to dpkg-buildpackage without doing the tests :P | 16:01 |
buZz | ooo, maybe DEB_BUILD_OPTIONS=nocheck | 16:03 |
buZz | now i just need to know how to make this version a different versionnumber from the repo one? | 16:12 |
Wizzup | buZz: you can always 'git checkout' the tag in the changelog | 16:14 |
buZz | i am not sure what that means :P | 16:14 |
buZz | oh, for the kernel | 16:14 |
buZz | i'm on my custom mosh build now with huge timeouts and intervals | 16:20 |
buZz | gonna try a day of 'mosh in my pocket' to see if it improved | 16:21 |
bencoh | for gprs I just use ssh, but ... | 16:22 |
uvos__ | buZz: kernel branch is back to the usual sane maemo-x.y | 16:28 |
uvos__ | checkout my fork of the maemo repo for the feature branches, if you want to change something in those & remerge | 16:28 |
uvos__ | so mosh sux? | 16:30 |
uvos__ | i was wanting to try it | 16:30 |
Wizzup | well, it's worse than ssh for pm at least | 16:31 |
Wizzup | unless you SIGSTOP it like Tony does | 16:31 |
buZz | or patch it | 16:32 |
buZz | i'm trying the patchroute now | 16:33 |
buZz | well, doesnt -seem- to have dropped a lot | 16:50 |
buZz | meh | 16:50 |
Wizzup | yeah I think you're better off with sshg | 16:52 |
Wizzup | ssh* | 16:52 |
buZz | oh, had a browser open still, lol | 16:54 |
buZz | ok mosh running seems to 'cost' about 50 mA | 16:54 |
buZz | browser with js thingies open about 100-200 | 16:55 |
Wizzup | so a 4-5x battery life reduection | 16:55 |
buZz | i'll measure more | 16:55 |
buZz | if its reall 4-5x, i could fork it for leste? | 16:55 |
buZz | really* | 16:55 |
buZz | or maybe jus fork it fully to a mosh-lowpower or something | 16:56 |
buZz | i wonder if those typos are mine or a effect of my patches, hehe | 16:57 |
Wizzup | buZz: no, I'm saying running mosh will make your battery drain 5 times faster | 16:58 |
Wizzup | I suspect that in general udp might not be good for power saving | 16:58 |
buZz | oh like that, maybe yeah | 16:58 |
buZz | maybe i should remove 'seconds' from my irc timestamp too, so it doesnt get new data every second, lol | 16:59 |
buZz | or fully | 16:59 |
buZz | hmm | 17:00 |
Wizzup | yeah that won't help | 17:00 |
buZz | right | 17:00 |
buZz | maybe mosh as a whole isnt that optimal for irc | 17:01 |
buZz | if only ssh was able to reconnect over different ips :P or survive ifconfig eth0 down, etc | 17:03 |
bencoh | it could survive down interfaces, depending on how you conf it | 17:04 |
buZz | hmm | 17:04 |
Wizzup | buZz: I mean if you ifconfig down with mosh it'll work ok with pm | 17:05 |
Wizzup | but you can also just SIGSTOP it then | 17:05 |
buZz | hmhm, or just detach/logout | 17:05 |
buZz | is that sigstop daemon a bit eh, fluent now? | 17:06 |
buZz | might be just what i want, just sleep the tool until i open phone again | 17:06 |
buZz | tnx, wouldnt have thought about the seconds in display without rubberducking a bit ;) | 17:11 |
buZz | https://space.nurdspace.nl/~buzz/irssi-on-d4.png | 17:19 |
buZz | :P | 17:19 |
Wizzup | uvos__: so the trace from yesterday I think might relate to the ofono timeouts we're seeing | 18:22 |
uvos__ | Wizzup: yeah i did see it | 18:22 |
uvos__ | buZz: if you mean sigstopd, that can only gui tasks | 18:23 |
uvos__ | (ie tasks that have at least one xwindow) | 18:23 |
uvos__ | *can only stop | 18:24 |
buZz | oh, so not a process inside a xterm? | 18:24 |
uvos__ | nope | 18:24 |
buZz | aww | 18:24 |
uvos__ | but a extension that stops other tasks on mce display off would be cool | 18:24 |
uvos__ | (rn it stops based on what windows are visible) | 18:25 |
uvos__ | no window viisble = stopped | 18:25 |
uvos__ | buZz: actually inside of a xterm would work | 18:26 |
Wizzup | that could be catastropic for quite some windows/tasks too | 18:26 |
uvos__ | since it stoppes all children too | 18:26 |
uvos__ | so it wouls stop when the xterm becomes hidden | 18:26 |
uvos__ | Wizzup: sure but you have to select which tasks to monitor by hand | 18:26 |
uvos__ | Wizzup: so dont select those :P | 18:26 |
buZz | uvos__: and hidden == screen off? | 18:27 |
uvos__ | buZz: yes | 18:27 |
buZz | nice | 18:27 |
uvos__ | any form of hidden | 18:27 |
uvos__ | actually its not screen off per say | 18:27 |
uvos__ | but hidden by tklock | 18:27 |
bencoh | sigstopd is very cool for firefox btw :) | 18:27 |
uvos__ | it only stops when locked | 18:27 |
bencoh | like, *very* | 18:27 |
uvos__ | bencoh: firefox wasting power was why i made the thing :) | 18:29 |
bencoh | :) | 18:29 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!