Wednesday, 2021-06-23

rinigusmorning! is there a way to disable systemd service that is installed by one of SFOS packages? ideally, by adding something into /etc/systemd via config/sparse09:03
rinigusI have limit-cpu.service messing up actdead for tama/aarch64/aosp1009:03
rinigus... found it! just have mask the service via symlink to /dev/null09:13
piggzrinigus: ah, i thought i had a similar issue on volla, but it was an android service
riniguspiggz: not in this case :) .10:08
rinigusI have framerate for compositor going down to 37fps while there is no app active in foreground. is there some knob to turn to keep it at 60 (62) fps while device screen is on?13:02
malrinigus: which device?13:03
rinigusmal: xz2c (tama)/aarch64/aosp1013:04
malI think x10 II has a bit similar issue13:04
rinigusmal: so it gets choppy while interacting with lipstick?13:04
malthere are some framerate issues, not fully sure in which situations13:14
rinigusmal: but any ideas on why such slowdown is happening?14:47
malnot at the moment14:47
malrinigus: have you yet tried to fix hall sensor?14:50
rinigusmal: no, I haven't. as it wasn't flagged with aosp10 label. should look into it - just making new image which I expect will fix actdead mode for us14:51
malrinigus: which platform is that device?14:52
malI mean qcom platform14:52
rinigussd845, if it is what you mean14:52
rinigusmal: ^14:52
malrinigus: so that is still using 4.9 kernel even on aosp10?14:53
rinigusmal: no, it is on 4.1414:53
rinigusbut that issue was filed against aosp9 port14:53
rinigus(hall sensor)14:54
malrinigus: hmm, this is then confusing
rinigusyes, it was mixed up by the reporter. or he used very old AOSP 10 release14:54
malbut anyway, try to update 4.14 kernel to latest from sony14:54
rinigusmal: is this update available via mer-hybris kernel repo as well? I tend to sunc from there14:55
mal and the fix I made was also for sdm84514:55
malat least I updated the mer-hybris sony kernel14:55
rinigusnice, will use that14:56
malso if you rebase your changes on top of that it should be ok14:56
rinigussounds like it.14:57
piggzlbt: how about a psa for obs fully up and working ;)16:19
lbthehe ... sorry I am swamped at the minute16:19
lbtwhat's the main issue atm ?16:20
lbtI know webhooks is not working16:20
piggzlbt: id appreciate 4.0 and 4.1 on all archs working16:21
lbtbut afaik builds for all arches doesn't it ?16:21
piggzdoes it?16:21
piggzlbt: no targets visible in the repo adder ui16:21
lbtit does build - I didn't know about the repo adder16:21
lbthas that ever worked?16:22
lbtpiggz: edit the meta as per:
piggzlbt: yeah, always used to work, but ill try manually16:23
piggzwhat does latest point to?16:23
piggzi usually do manually, but use the UI to see whats available16:24
piggzlbt: what is the 4.1 version?16:26
piggzis that working?16:26
piggzcompared to
lbtit's not yet16:28
piggzthat would be a major plus, getting 4.1, and then i could update ports and apps16:29
piggzlbt: also, there must be something changed with the names .. do you think we should update, or you will make the names match ... see
malbtw, there is no latest_aarch64 yet in
piggzlbt: thx, was that the names?17:15
piggzaarch64 is saying failed on hw:common due to missing packages17:15
lbtno, latest_aarch64 is there now17:15
piggzlbt: oh, but those projects say suceeded now!17:16
lbtI've been kicking things17:16
malpiggz: my bad17:16
malpiggz: had a mistake in meta for aarch64 in devel common17:16
piggz :)17:17
lbtah good17:17
lbt4.1.0.24 is broken for some reason17:17
lbtI wonder if it's on jolla side17:18
mallbt: is it possible to do multi target repo like this that
malit seems to only build aarch64 at the moment17:19
piggzlbt: define broken :)17:20
lbtmal: it may not do what you want it to do17:20
lbtpiggz: missing some cross- stuff17:20
piggzhmmm, shame, hope u canfind a fix17:21
mallbt: hmm, wondering how to handle that then because we didn't think long time ago about the naming so the repo name doesn't contain arch17:22
lbtyeah - proof:
lbtit kinda should work but I suspect things are not truly multi-arch in sfos17:26
lbtand I bet our noarch isn't good enough17:27
malhmm, yeah, noarch might be a bit problematic17:27
piggzone failure here
malpiggz: I bumped it, it was old version17:33
malpiggz: now all built17:34
piggzso, that just leaves 4.1, and all will be good with the world17:35
piggzoh, and the telegram bridge17:35
malwe still have the repo naming issue in testing17:35
piggzmal: you can fix that righ? will ports need to point to a different url?17:36
malnot sure if it fixable without changing repo urls on devices17:39
piggzyeah, seems so17:39
mallbt: does obs support repo aliases?17:40
maljust thinking if we would rename the repos but keep the old arm naming until devices have moved to new naming17:40
lbtI don't know - I doubt it. There's not much control over publishing like that17:40
malof course we can have the old naming as separate repo, just a bit of useless rebuild17:41
piggzis it becuase the normal test: url doesnt include _%(arch)17:43
malpiggz: yes17:45
piggzmal: i always wondered why that was in devel: but not testing: :D17:45
malyeah, it makes very little sense17:45
piggzi would make the repos build, and porters or users can fix the url17:46
piggzatleast that way it works .. the other way doesnt build!17:46
malit would be simple to change the url if people would just update from the repo with a new community-adaptation package and repos would update automatically17:47
malbut for transition period we probably want to have the old one there for a release or two17:47
piggzmal: yeah, so, like you said, just have an extra duplicate repo for arm17:48
malyes, this way we could also have i486 in testing17:48
malunlike now17:48
mallbt: how does that sound ^17:49
piggzmal: that was my next question ... wheres x86?17:49
piggzi guess i never noticed as i never moved latte from devel!17:50
malwe only would need to change these and and one sed command in the .inc file17:51
malI think first one is not even needed unless we want new naming for adaptation repo also17:51
lbtit seems sane to keep the old name around as a duplicate build of vanilla armv7hl17:52
lbtuntil you feel you've transitioned17:52
piggzmal: best to be consistent and change both17:52
mallbt: yeah17:52
malpiggz: the common repo change would be automatic but adaptation repo would need porter to change the obs project17:54
piggzmal: sure, ive no issue doing that for consistency17:55
piggzand its me and rinigus shouting mostly about obs :D17:55
malI'll ask sledges17:56
lbtso are things a touch better now piggz?18:08
lbtI tried to get up but there's something weird happening there so it'll need some digging18:08
piggzlbt: a touch ... just _reeeeeeeeeealy- need 4.1 ;)18:08
piggzim now building kf518:09
piggzmal: when changing repos, consider also fixing sailfish/sauifishos descrepancy18:15

Generated by 2.17.1 by Marius Gedminas - find it at!