*** zbenjamin is now known as Guest56735 | 02:53 | |
*** zbenjamin_ is now known as zbenjamin | 02:53 | |
lachs0r | okay, according to dbus-monitor connman is scanning wifi networks every 3 seconds and then dumps all the info through dbus | 03:36 |
---|---|---|
lachs0r | very. slowly. | 03:36 |
lachs0r | using a lot of cpu time and also stalling lipstick as well as all network io on the wifi interface | 03:36 |
lachs0r | also it is infuriating that there is no perf package | 03:37 |
lachs0r | (and no kernel-source for that matter!) | 03:37 |
lachs0r | so I wonder what causes connman to poll wifi networks every 3 seconds. this is definitely not a thing it should be doing | 03:38 |
lachs0r | also jesus fucking christ jolla people need to learn how to write commit messages instead of references to private bug trackers | 03:43 |
lachs0r | this is infuriating. how is anyone supposed to contribute to a project when development is so opaque | 03:44 |
r0kk3rz | kernel sources are available | 04:27 |
r0kk3rz | which device? | 04:27 |
lachs0r | xperia x, but nvm that | 04:32 |
lachs0r | traced the problem to connman/plugins/sailfish_wifi.c | 04:32 |
lachs0r | either the autoscan intervals are too short or the timers actually get added multiple times | 04:32 |
r0kk3rz | https://github.com/mer-hybris/android_kernel_sony_msm | 04:33 |
r0kk3rz | for future reference | 04:33 |
lachs0r | would be nice if those were actually packaged and installable | 04:33 |
lachs0r | instead of just the headers | 04:33 |
* lachs0r mumbles something about gpl violations | 04:34 | |
r0kk3rz | as in source packages? | 04:35 |
lachs0r | yes | 04:36 |
r0kk3rz | im not sure we've got source packages for anything | 04:36 |
lachs0r | we actually do | 04:37 |
lachs0r | I’m getting kinda fed up with the way sfos is built and how little opportunity there seems to be for me to contribute to the base system | 04:38 |
r0kk3rz | yeah well jolla runs the show | 04:39 |
lachs0r | mer packaging is just BAD and a lot of copy/paste from fedora packages (which are NOT GREAT in terms of QA) | 04:39 |
lachs0r | now I kinda want to get a project going that is managed more like opensuse tumbleweed | 04:40 |
lachs0r | but I cannot even do that as I’m unable to even get the OBS interconnect to work and build.sailfishos.org is broken | 04:41 |
lachs0r | sorry if I’m ranting a little too much but I just don’t think this is how you build a community | 04:42 |
lachs0r | the fact that the best we have in terms of independent software repositories is openrepos is just sad. I think we can do a LOT better than that | 04:43 |
r0kk3rz | agreed | 04:43 |
r0kk3rz | theres a number of glaring issues with openrepos, but nobody has built an alternative | 04:44 |
r0kk3rz | the guy who runs it doesnt care about sailfish anymore | 04:44 |
lachs0r | it seems to me like everybody is just expecting jolla or someone else to fix all the glaring issues with sfos for them, and of course that never happens so people jump ship | 04:45 |
r0kk3rz | some people are like that sure, but the community devs either left or were hired by jolla | 04:47 |
lachs0r | definitely need better community resources. a real issue tracker, obs repos with a project hierarchy and actual package review, more documentation on basic app architecture and design (to help beginners develop better apps) | 04:48 |
r0kk3rz | the mer bugzilla was never really used | 04:50 |
lachs0r | not surprising | 04:51 |
r0kk3rz | the public obs is just for the core packages | 04:51 |
lachs0r | yeah, and it’s very messy | 04:53 |
r0kk3rz | yeah its easy to look at the wrong project | 04:54 |
lachs0r | honestly I wish I could just replace connman with networkmanager because at least nm is mature and works (plus it has vpn support already) | 07:16 |
lachs0r | connman cannot even keep track of interface state reliably | 07:18 |
lachs0r | okay I think I found the offending line in connman but I have no idea why it’s there. sailfish_wifi.c has some of the worst state management code I’ve ever seen | 09:22 |
lachs0r | lbt / Sage_: can I get accounts at git.sailfishos.org and build.sailfishos.org? | 09:26 |
lachs0r | (also sorry for overreacting just now, just frustrated because the impact on battery life is so severe) | 09:28 |
PaulePanter | Hi. I installed the Signal App from some Android store three(?) years ago. | 09:44 |
PaulePanter | It looks like it was never updated. | 09:44 |
PaulePanter | What store do you recommend for Signal? | 09:44 |
* PaulePanter is currently using LibreSignal 4.1.0 which was abandoned in 2017. :( | 09:52 | |
r0kk3rz | lachs0r: you have a tjc account? | 10:03 |
lachs0r | yes | 10:03 |
lachs0r | https://together.jolla.com/question/215847/connmanlipstick-stuttering-and-battery-drain-with-wlan-enabled/ btw | 10:04 |
r0kk3rz | ok, message lbt with the username | 10:04 |
r0kk3rz | lachs0r: do you have a hidden ssid network configured? | 10:06 |
lachs0r | r0kk3rz: no | 10:07 |
r0kk3rz | interesting | 10:08 |
r0kk3rz | do you have many wifi networks around you? | 10:08 |
lachs0r | oh just disconnected the wifi and it’s lagging again (more severely) | 10:08 |
lachs0r | so there are two issues | 10:08 |
lachs0r | yes | 10:08 |
r0kk3rz | ah ok | 10:08 |
r0kk3rz | ive seen that before | 10:09 |
lachs0r | well hang on, trying to debug this further | 10:10 |
r0kk3rz | lachs0r: if you want help with middleware layers and debugging tricks or whatever, you can join the #sailfishos-porters channel | 10:12 |
lachs0r | no I see what’s going on | 10:12 |
lachs0r | the autoscan interval is constant 5 seconds | 10:12 |
lachs0r | way overkill imo | 10:12 |
r0kk3rz | just an fyi | 10:12 |
lachs0r | okay looks like it’s resetting the autoscan interval every time | 10:15 |
lachs0r | yeah, wifi_device_bsss_changed | 10:18 |
lachs0r | it forces 5 seconds while the screen is on | 10:18 |
lachs0r | that is dumb | 10:18 |
lachs0r | it should not change behavior just because the screen is on. if the user wants the phone to connect faster, they’ll just do it manually imo | 10:19 |
lachs0r | this will just drain the battery like crazy for no good reason | 10:20 |
lachs0r | https://git.sailfishos.org/mer-core/connman/blob/master/connman/plugins/sailfish_wifi.c#L2859-2868 | 10:21 |
r0kk3rz | lachs0r: share your findings with spiiroin | 10:26 |
ender| | <lachs0r> xperia x barely lasts one day with wifi enabled, several days with it disabled ← i never had this problem - my Xperia X takes about 2 days to discharge to around 50%, at which point i charge it again | 13:42 |
adantes | ender|: turn off the wifi was a great option. Since yesterday, I guess that, in normal use, Ive just lost 10% in gemini, and in XA2 | 13:44 |
ender| | interesting. i've got about 20 wifi networks available here, and it's never been a problem for me | 13:45 |
mal | strange, for me xperia x has very good battery life with wifi enabled, maybe some app is using internet connection a lot? | 14:26 |
*** Renault_ is now known as Renault | 14:29 | |
ender| | https://marc.info/?l=openssh-unix-dev&m=157259802529972&w=2 | 14:55 |
abranson | yeah i think there's something else going on if he's getting such high drain compared to others | 15:27 |
mal | abranson: from my experience xperia x has probably the best battery lifes of the devices | 15:43 |
ol | I was experiencing high battery drain in Riga city centre with WiFi enabled on Jolla C. | 15:49 |
mal | I think it depends on how much wifi hotspots are around | 16:39 |
r0kk3rz | yeah ive seen it start to choke in cities | 21:50 |
r0kk3rz | on jolla c | 21:50 |
lachs0r | well I’m in seattle now, so lots of wifi networks around | 23:47 |
lachs0r | it wasn’t a problem in my home town | 23:47 |
lachs0r | I did notice some battery drain with wifi enabled but it wasn’t nearly as bad | 23:47 |
lachs0r | I’m about to check whether this change fixes the wifi power consumption on the tablet (it’s always been suspiciously high) | 23:49 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!