rinigus | lbt sweet! Now would have to review all what needs to be updated... | 06:14 |
---|---|---|
rinigus | How do we get packages to chum repo? Would it fit for something like TTS engines? | 06:15 |
T42 | <adampigg> Rinigus, what is chum repo? | 06:26 |
rinigus | @adampigg: https://forum.sailfishos.org/t/community-obs-refurbished-and-re-floated/6556/10 | 06:34 |
rinigus | not sure it is used yet, but would be great to have a repo where we can add all kinds of libs/tools missing in SFOS | 06:34 |
rinigus | tts for example | 06:34 |
T42 | <adampigg> Cool | 06:37 |
T42 | <adampigg> Im busy this morning, but looking forward to testing 4.1 out later on pro1 | 06:38 |
T42 | <spodermandowebswings> how to install this libm : https://irc.thaodan.de/.imgstore/9cfce15f/file_2216.jpg | 07:48 |
T42 | <spodermandowebswings> recently im trying sailfish 3 on my old device | 07:49 |
T42 | <spodermandowebswings> but mainly due to libm.so.6 im not able to install packages | 07:49 |
T42 | <elros34> you have outdated system, either update it, try older package for application you are trying to use or recompile it | 08:58 |
rinigus | Is there a way to know whether device is currently showing Lipstick (tiles, notifications, app selector) or is showing an app? Nothing pops up on session DBus, as far as I can see. | 09:51 |
rinigus | Asking as Tama/AOSP10 looks to have AOSP/BLOBs bug leading to slowdown of the graphics while showing lipstick | 09:52 |
rinigus | but all is fine as soon an app is active | 09:53 |
T42 | <elros34> IIRC not very reliable and not excatly what you are looking for but there is com.jolla.lipstick coverstatus . There is also org.nemomobile.compositor privateTopmostWindowProcessIdChanged 0 for lipstick which might be better for detecting | 10:08 |
rinigus | @elros34: thanks! coverstatus as far as I understand is just how many covers are shown? | 10:41 |
rinigus | will have to look into nemomobile.compositor. it is on system bus which actually would fit nicely | 10:43 |
T42 | <elros34> not sure what coverstatus is really used for but it's about transition. According to my notes: 3 => 0 - events or other window, 1 => 2 - home screen. It change when you start peek, cancel, activate window | 10:45 |
rinigus | @elros34: that org.nemomobile.compositor privateTopmostWindowProcessIdChanged looks to be perfect! it is on system bus and accessible to root easily. So I can plugin it to my zgovernor and bump GPU frequency when lipstick is shown | 10:47 |
rinigus | this should resolve glitches and slow rendering with min battery hit - gonna try that one. | 10:48 |
rinigus | if I would use coverstatus, I would need to run a separate service to track the transitions and communicate it to the system service before GPU could be adjusted | 10:49 |
rinigus | @elros34: thanks again for pointing in this direction! | 10:50 |
lbt | PSA: Webhooks should be working again | 10:50 |
T42 | <elros34> rinigus: yeah I was struggling with coverstatus only and end up using both methods to detect when I am on homescreen | 10:52 |
T42 | <elros34> lbt: will it be possible to build for older release like 3.4.0.x or you dropped older releases? | 10:53 |
lbt | I should be able to enable them at some point | 11:04 |
T42 | <elros34> great | 11:05 |
spiiroin | rinigus: regular glob() is used [in mdy_governor_apply_setting()], so "/sys/devices/system/cpu/cpu[4-7]" should be a usable pattern | 18:21 |
spiiroin | while testing feel free to relax the probably too draconian path validator, or disable it altogether | 18:22 |
spiiroin | the wip branch is rather old, so yes: that script is python2 | 18:24 |
spiiroin | ... IIRC I did some logic sanity checking for it on xa2 or something, but then something else came up. so it is in rather raw state | 18:27 |
rinigus | spiiroin: had to use my own zgovernor still - it is tracking also whether lipstick shows some window or is displayed itself. thanks to some bug in aosp10 and/or blobs for sony tama, interactions with lipstick were choppy | 18:40 |
rinigus | so, it is rather port-specific combo and it makes sense to offload it all to a single daemon and not split between mce and it | 18:41 |
rinigus | which is pity, but unfortunate state of tama aosp10 base | 18:41 |
piggz | mal: @notkit figured out the last 2 remaining parts to get sailjail working on volla... | 18:42 |
piggz | [defaultuser@VollaPhone ~]$ cat /etc/firejail/whitelist-common.local | 18:42 |
piggz | whitelist /dev/mali0 | 18:42 |
piggz | whitelist /dev/pmsg0 | 18:42 |
piggz | whitelist /dev/ion | 18:42 |
piggz | whitelist /dev/hwbinder | 18:43 |
piggz | whitelist /dev/__properties__ | 18:43 |
piggz | private-etc ld.so.preload | 18:43 |
mal | hmm | 18:46 |
mal | piggz: so my way of debugging was the correct then | 18:47 |
mal | piggz: meaning going through the strace and seeing what needs to be whitelisted | 18:47 |
piggz | pretty much ... only thing i missed was __properties_..... | 18:47 |
piggz | and, the private-etc is needed for our tls padding preload | 18:47 |
mal | ok | 18:51 |
piggz | mal: the original volla port, based on android 9, had the device name volla:yggdrasil ... the new one, based on los17/arm64, has the device name gigaset:gs290 .... | 21:33 |
piggz | should I use that for a new obs project? or volla:gs290 ? | 21:33 |
piggz | what do u think | 21:33 |
T42 | <nephros> rinigus: anything one can help testing for zgovernor? I have a J1, Gemini, and X10 if that is useful. | 21:46 |
piggz | mal: i think gigaset:gs290 is the way to go, keeps the pot seperate from the old arm32 version | 21:48 |
b100dian | Mister_Magister: do you remember how you solved this? "OMXNodeInstance: !!! Observer died. Quickly, do something, ... anything..." | 22:52 |
Mister_Magister | eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee | 22:52 |
Mister_Magister | maybe i didn't solve it | 22:52 |
Mister_Magister | do you have link to logs? | 22:52 |
Mister_Magister | context matters | 22:52 |
b100dian | I have that on video recording | 22:54 |
b100dian | you probably had it too long long time ago https://piggz.co.uk/sailfishos-porters-archive/index.php?log=2017-08-23.txt#line512 | 22:55 |
Mister_Magister | i wonder what device was that | 22:56 |
b100dian | well.. I don't know either. | 22:58 |
b100dian | did you had any port with missing video recording:) | 22:58 |
Mister_Magister | it was possibly an x force and x force has still broken video recording | 22:58 |
Mister_Magister | so like i said i mostlikely didn't fix it | 22:58 |
Mister_Magister | b100dian: multiple xd | 22:58 |
Mister_Magister | my x2 which i daily drive still has video recording brooken | 22:59 |
b100dian | understandable:) | 22:59 |
Mister_Magister | but yeah it was mostlikely x force and most likely never fixed | 22:59 |
Mister_Magister | video recording on motorolas is a biatch | 22:59 |
b100dian | Well I have that on a Xiaomi, so :hmm: | 22:59 |
Mister_Magister | b100dian: you gonna have haaaaaaaard time | 23:00 |
Mister_Magister | as its never-fixed-before problem | 23:00 |
Mister_Magister | we figured out most of them past years but not all of 'em | 23:00 |
b100dian | ok, challenge accepted :-S | 23:00 |
b100dian | was going to ping you before actually starting debugging myself | 23:01 |
Mister_Magister | 2017 mal didn't have idea but maybe 2021's mal is better xd | 23:01 |
Mister_Magister | also abranson which doesn't seem to be present has good knowledge around camera | 23:01 |
Mister_Magister | who* | 23:02 |
Mister_Magister | dang i've been doing this shit for around 6 years already huh | 23:02 |
Mister_Magister | didn't learn a thing | 23:02 |
b100dian | Do you also get yourself speaking in the logs when searching a current problem:p | 23:03 |
b100dian | I get that, and it's only 5 months:) | 23:03 |
Mister_Magister | sometimes | 23:03 |
Mister_Magister | sometimes i intentionally search for myself cause i remember i had this issue in past | 23:03 |
b100dian | :) | 23:08 |
b100dian | the problem with my camera is that logs are soooo verbose they don't fit a pastebin service :) | 23:09 |
b100dian | here is logcat, everything starts around 02:04:25 http://vlad.synology.me/logcat-camera-1.txt | 23:13 |
Mister_Magister | gl | 23:17 |
T42 | <b100dian> I take that's a prayer | 23:20 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!