libera/#maemo-leste/ Tuesday, 2025-01-07

arno11inky: no i don't use the 20$ stuff00:22
arno11you spend money depending of the number of tokens used and the model00:23
arno11gpt-4o-mini is really cheap00:25
arno11and powerful enough for text00:25
arno11and yes you can use api key for few $00:26
arno11and not pay them 20$ a month00:26
arno11to find the trick, i asked...chatgpt00:27
arno11Wizzup: the leste assistant works fine on n900 with libcurl :)00:28
arno11it mainly provides answers based on irc logs but able to provide links to the leste wiki as well00:31
arno11(or other links from logs ofc)00:33
arno11bluffy...00:35
inky> to find the trick, i asked...chatgpt01:27
inkywow01:27
freemangordonhmm, kmscube doe not run :(07:48
freemangordon*does07:48
freemangordonXorg segfaults07:49
freemangordonok, upstream kmscube works :)08:24
freemangordonglmark2 2021.02 works too08:57
freemangordonso, I have to find why xorg driver segfaults08:58
freemangordonglmark2 Score: 4909:03
freemangordonthat's vsynced09:03
freemangordonWizzup: d4 booted to h-d, have to run now, will push the needed stuff later on09:26
freemangordonglmark2 Score: 6509:34
Wizzupfreemangordon: getting mixed signals but sounds like it works?11:58
dsc_https://x.com/Scobleizer/status/187632899398685930514:36
Wizzupfreemangordon: for sunxi images, do we use our current sunxi kernel?15:42
sicelodsc_: that looks like a nice little machine. i like the simple UI as well16:00
dsc_sicelo: I think this powers that UI https://github.com/mecha-org/mctk16:00
dsc_but not sure16:00
siceloyes16:06
g4rrgl3n0m4d /616:19
dsc_g4rrgl3n0m4d blessing us with the unexpected by 6 division16:21
g4rrgl3n0m4dsorry just typed in the wrong window16:23
dsc_its alright16:25
freemangordonWizzup: had to fix xorg omap driver, but everything works17:34
freemangordonand to pull latest sgx binaries ofc17:35
freemangordonre sunxi - I use 6.12 with patches17:35
freemangordonanyway, just came home, will push the changes later on, hopefully17:36
freemangordonWizzup: a side note: neither kmscube nor glmark-es2-drm in debian repos work17:37
freemangordonhad to compile upstream kmscube (which works) and some older glmark217:37
Wizzupok17:38
Wizzupok, so, when you push those fixes I can make a new image17:38
Wizzupas far as sunxi goes, I guess we might add those patches to our droid kernel, or do we want to maintain separate sunxi kernel?17:38
Wizzupboth are easy for me17:38
Wizzupjust checking what to do for sunxi images17:38
freemangordonwe will have to enable lots of options17:46
freemangordonso not sure it is possible to have the same kernel17:53
freemangordonbut, we can try17:53
Wizzupseems like it should be possible17:54
freemangordonyay, both n_gsm patches seem to make it to upstream :)17:57
freemangordontmlind: ^^^17:57
freemangordonunfortunately silence for all others17:57
freemangordon*other17:58
freemangordonWizzup: what the?!? https://github.com/maemo-leste-upstream-forks/mesa/actions/runs/1263552926621:01
sicelofreemangordon: i am running the updated ofono for the first time on my D4 today ... no SIM in device. org.ofono.Modem.GetProperties() shows Powered = true and Online = true. i can't check further for now, but thought to mention so long21:34
freemangordonsicelo: umm... you report a bug or?21:34
freemangordonWizzup: who and why did this? https://github.com/maemo-leste-upstream-forks/xorg-server/commit/af4c84ce8855e84c0ad89b929bc972e884f0b8e321:35
Wizzuplooking (in mtg)21:38
Wizzupfreemangordon: github probably does silly stuff21:38
Wizzupfreemangordon: looks like they themselves did this21:39
freemangordonbut this is not in upstream21:39
freemangordonnor in deadalus xorg21:39
freemangordonwhere did we gat the git tree from?21:40
freemangordon*get21:40
freemangordonfrom debian?21:40
freemangordonanyway, I will fix that21:41
freemangordonumm, will fix omap driver to use anotehr propertyu21:41
sicelofreemangordon: i mean ofono that's now in Leste (2.13)21:44
freemangordonok, but don't understand if there is some issue21:44
siceloi don't expect Online = true when no Sim is present21:45
freemangordonhow is that?21:45
freemangordonyou can;t make emergency calls if modem is offline21:45
freemangordonsee doc21:45
sicelook.21:46
freemangordonhttps://github.com/maemo-leste-upstream-forks/ofono/blob/maemo-ofono/doc/emergency-call-handling.txt21:46
freemangordon"Online property of org.ofono.Modem interface should be set to TRUE."21:46
freemangordon"Case 4: Call in No SIM state"21:47
freemangordonanyway, this is not related to ofono, it is cellulard that sets the property21:47
siceloyes i know that doc, just don't remember this behavior on the droid 4. guess false alarm21:50
freemangordonwell, maybe qmi interface behaves differently21:50
Wizzupfreemangordon: yes from debian21:51
sicelothe doc is simply saying to go online if an emergency call is being attempted from UI. i guess it's still fine if we always go online21:51
Wizzupare you sure it's not in daedalus xorg?21:51
freemangordonyes, I am21:51
Wizzupwtf21:51
Wizzupok, I can look later how this happened (work mtgs atm)21:52
Wizzupbut if you can fix it, that's fine too21:52
freemangordonok, no hurry21:52
freemangordonthis is daedalus https://github.com/maemo-leste-upstream-forks/xorg-server/commits/maemo/daedalus/config/udev.c21:52
freemangordonvs https://github.com/maemo-leste-upstream-forks/xorg-server/commits/maemo/chimaera/config/udev.c21:53
freemangordonsicelo: well, I guess it does not make sense to have a phone without SIM :)21:53
freemangordonhowever, if there is any rationale to not go online without SIM, please elaborate21:54
freemangordonI mean - this is what my understanding was back then, but if you think there is an issue we can change it21:54
siceloi think it's fine21:56
Wizzupfreemangordon: can you check debian xorg?22:01
Wizzuphttps://salsa.debian.org/xorg-team/lib/mesa/-/tree/debian-bookworm/debian/patches?ref_type=heads22:01
Wizzuphttps://salsa.debian.org/xorg-team/lib/mesa/-/tree/debian-bookworm?ref_type=heads22:01
freemangordonumm... this is mesa22:01
Wizzupsorry that is mesa22:02
Wizzuphttps://salsa.debian.org/xorg-team/xserver/xorg-server/-/tree/debian-bookworm?ref_type=heads22:02
freemangordonHTTP 502: Waiting for GitLab to boot22:02
freemangordon:)22:02
freemangordonhttps://salsa.debian.org/xorg-team/xserver/xorg-server/-/commits/debian-bookworm/config/udev.c?ref_type=heads22:03
freemangordonreverts are there22:04
sicelofreemangordon: only nit would be that https://github.com/maemo-leste/cellulard/blob/master/src/cellulard.c#L231 should be get_mode instead of set_mode :p22:04
sicelobut it's mostly cosmetic, so doesn't matter too much22:04
Wizzuphttps://salsa.debian.org/xorg-team/xserver/xorg-server/-/commits/debian-bookworm/config/udev.c22:04
Wizzupyes reverts are there22:04
Wizzupso that's what debian did22:05
Wizzupor someone22:05
WizzupI think oliver fourdan is from xorg folks22:05
freemangordonok, but we don;t have those in our tree22:06
freemangordondo we use upstream xorg?22:06
sicelodidn't we use newer xorg to accomodate something ... maybe pp?22:07
sicelonewer than debian/devuan, that is22:07
freemangordonI mean - I am fine, I already fixed xorg driver to use  ODEV_ATTRIB_DRIVER instead of ODEV_ATTRIB_BUSID22:07
freemangordonsicelo: hard to say :)22:07
freemangordonthat's what I am trying to understand - what do we use in daedalus22:08
Wizzupfreemangordon: we don't have those?22:09
freemangordonI don't see those reverts in our daedalus tree22:09
Wizzupdid you also check the tag?22:09
Wizzupjust to confirm22:09
freemangordonhttps://github.com/maemo-leste-upstream-forks/xorg-server/commits/maemo/daedalus/config/udev.c22:09
freemangordonsorry, what tag?22:11
Wizzupsorry, I shouldn't try to do this during mtgs22:15
WizzupI mean the git tag that we build src from22:15
Wizzup21.1.7.122:15
freemangordonno, I didn;t22:15
freemangordonbut, isn;t checking the tree enough?22:16
freemangordonand given that omap xorg driver segfaults because busid is not set?22:17
freemangordonanyway, not on hurry, please finish you mtg22:17
freemangordonif you have time later on, please try to understand what's going on with xserver we use on daedalus22:17
freemangordonI will not push anything to xorg driver until that's cleared, as it may turn out it is not needed22:21
freemangordonsicelo: re ofono - what concerns me most is if there is any change in power usage22:24
freemangordonas afaik ofono does not have any logic to control the frequency of notifications over qmi22:24
freemangordonand I have seen RSSI events coming avery few seconds22:24
freemangordon*every22:25
freemangordonbut, this is general issue we have to address, not sure where22:25
siceloin your use so far, noticed any higher usage?22:25
freemangordonRN there is a hack in d4, but it is device specific, we shall find a solution that's applicable for all devices22:26
freemangordonit is hard to say as I am torturing the device with lots of changes, including 6.1222:26
arno11btw ofono 2 seems to work fine on N90022:38
siceloyup, n900 side nothing changed. support for it still rock solid (except when they accidentally broke it in version 2.6, which we skipped on Leste)22:46
arno11ah ok22:46
siceloso from 2.6 to 2.12 N900 was broken. fun result - after the fix, now the modem is always /n900_0 instead of /n900_2 (which was sometimes annoying) :p22:51
arno11oh i didn t notice/check22:51
siceloyeah. not a big deal really, but did annoy me a few times ... e.g. when doing dbus calls directly22:53
arno11oh i still get n900-2 lol22:56
sicelommm, what ofono version?22:56
arno112.1322:57
arno11let me check again22:57
arno11yep 2.1322:58
siceloweird22:58
arno11but i just looked @list-contexts...22:59
sicelommm, what about /list-modems?22:59
arno11and it was n900_2 when registered22:59
sicelook. i don't know what's up there. doesn't really matter, as said before, but after https://git.kernel.org/pub/scm/network/ofono/ofono.git/commit/?id=116b6dfef7a7a4cfffaf3d02e713baa232195c73 i've never seen anything besides n900_0 on mine23:00
arno11ok, it s still n900_2 in modems23:01
arno11yes indeed, doesn t matter, it works fine23:01
Wizzupfreemangordon: ok, I am here23:17
Wizzupdo you need me to check something?23:17
Wizzupfreemangordon: on my maemo/daedalus branch I don't think I see the revert :/23:18
Wizzupwth23:18
Wizzupfreemangordon: so where did you see the revert in our maemo/daedalus23:18
Wizzuphm gitlab says it is part of that branch23:19
Wizzupgithub23:20
Wizzupbut I think it's lying23:20
Wizzuphttps://github.com/maemo-leste-upstream-forks/xorg-server/commits/maemo/daedalus/config/udev.c23:20

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