**** BEGIN LOGGING AT Sat Aug 24 02:59:57 2019 Aug 24 04:42:50 Saket Upadhyay was added by: Saket Upadhyay Aug 24 05:10:15 How's hackathan going on bro's 😄 Aug 24 05:13:21 it will be later Aug 24 05:13:49 14:00 is begining i think Aug 24 05:13:53 GMT +2 Aug 24 05:18:42 @AyraHikari [Hello, anyone know how to solve Unpacking got killed? … https://paste.ubuntu.com/ …], can you try https://gitlab.com/JBBgameich/halium-install/ Aug 24 06:35:22 Kachel was added by: Kachel Aug 24 07:16:03 Ok, it not booting … Refusing to boot. See /diagnosis.log (in initrd only) … $ ssh phablet@192.168.2.76 … ssh: connect to host 192.168.2.76 port 22: Connection refused Aug 24 07:16:09 What should i do? Aug 24 08:13:51 what do uou have in dmesg? Aug 24 08:13:57 on host? Aug 24 08:14:03 and device? Aug 24 08:18:03 Here is dmesg … https://del.dog/ubesigimom.coffeescript Aug 24 08:19:47 try telnet 192.168.2.15 Aug 24 08:19:56 (Photo, 169x202) https://irc.ubports.com/005qIBm1.png Also it added some drives in my pc Aug 24 08:20:24 let them be Aug 24 08:20:25 @Luka Panio [try telnet 192.168.2.15], Thanks, it works now Aug 24 08:20:41 what works? Aug 24 08:20:48 telnet or booting? Aug 24 08:21:37 telnet Aug 24 08:21:46 i need to fix boot.img in /diagnosis.log Aug 24 08:22:37 yes Aug 24 08:22:39 maybe Aug 24 09:24:03 i've add missing config in kernel and now my device is not booting … https://del.dog/cidabihivo.coffeescript Aug 24 10:03:24 @AyraHikari [i've add missing config in kernel and now my device is not booting … https://del. …], what do you mean with that? Aug 24 10:07:08 anything in dmesg? Aug 24 10:08:28 @Luka Panio [what do you mean with that?], Stuck at EDL mode ig Aug 24 10:08:31 Not booting Aug 24 10:08:42 what is yor device? Aug 24 10:08:46 your* Aug 24 10:08:49 No detection from adb, fastboot, port, storage, etc Aug 24 10:08:52 @Luka Panio [what is yor device?], land Aug 24 10:08:55 Redmi 3S Aug 24 10:14:53 ./halium/halium-boot/check-kernel-config path/to/my/defconfig -w Aug 24 10:28:44 oh sorry it is for ut Aug 24 11:51:07 malditobastardo was added by: malditobastardo Aug 24 12:01:46 It is 14h UTC+2: welcone to everybody in the list guesting this first Ofono Hackathon Aug 24 12:02:56 I'm Guf and together with many other Ubuntu Touch fans, we are working on UBports porting: we would like to organise this event to fix some issues regarding ofono in halium-7.1 porting Aug 24 12:03:48 We also invited other developers and contributors from projects using ofono pkg and being affected by the same issue. I meant LuneOS and Mobile Plasma Aug 24 12:04:26 We all have in common: starting from halium project and integrating an OS using **ofono*. Aug 24 12:05:07 The bug I was referring is the fact there is (or was) any sound in calls, either outbound from ported device and inbound Aug 24 12:06:12 During the last week, in UT porting we found a solution and now it is time to make it automatized in system.img and rootfs.img and test on every device Aug 24 12:07:02 Last thing to mention is that we have relied on a common pad to share ideas and everything can be useful: https://pad.ubports.com/p/ofono-hackathon Aug 24 12:07:42 We are also happy to invite to share some thoughts that can be part of the next porting podcast ;) Aug 24 12:07:53 This is an ongoing version, so feel free to read, add contribution, experience and everything you consider useful Aug 24 12:08:38 Do we have people around who would like to introduce themselves and the projects they represent? Aug 24 12:09:02 Feel free to do so as you please. Aug 24 12:09:53 Of course if there is someone from the other projects, they all are welcome Aug 24 12:10:58 Hey there Aug 24 12:11:13 If not, Alfred, can you resume the hard work of this week? especially what we could obtain and what workaround we used to fix audio in calls in UT Aug 24 12:11:41 @bhushanshah [Hey there], Hi Bhushan, welcome Aug 24 12:13:14 So I'm Bhushan, one of founder of Halium and PlaMo dev. Aug 24 12:13:36 @stefanomelchior [If not, Alfred, can you resume the hard work of this week? especially what we co …], finally, what you did to get audio in call on suzu and to integrate changes in halium-boot and telepathy-ofono Aug 24 12:14:18 hi Aug 24 12:14:24 Alright. So on Ubuntu Touch we realized that ofono itself was doing its thing as advertised. The issues we were having relate to audio rerouting being buggy, resulting in audio-less calls. What I've done to circumvent the issues for now is modify telepathy-ofono, a component which handles switching in- and outputs depending on whether Aug 24 12:14:24 it detects said in- and outputs, to read the h2w kernel switch (from the land of Android kernels) before prefering the headset jack for output. Aug 24 12:14:50 Before you continue hacking I've topic I like to address Aug 24 12:15:05 Give me 2-3 mins to type it out Aug 24 12:15:59 https://wiki.ubuntu.com/Touch/Audio?action=AttachFile&do=get&target=Ubuntu+Touch+-+Audio+Architecture+Overview.png Aug 24 12:16:06 So ofono which UBports is using is 1.21.x version as I understand it i Aug 24 12:16:10 [Edit] So ofono which UBports is using is 1.21.x version as I understand it. Aug 24 12:16:52 I think it is interesting to look at this pic to understand how halium and UT/PlasmaMobile and LuneOS interact mutually Aug 24 12:16:57 @bhushanshah [So ofono which UBports is using is 1.21.x version as I understand it.], It is a fork of 1.17 Aug 24 12:17:29 Right Aug 24 12:18:22 So, at one point canonical tried to upstream patches to ofono hosted on git.kernel.org Aug 24 12:18:35 And they upstreamed squashed version of patchset upstream Aug 24 12:19:42 But after that more patches were added on both ends Aug 24 12:19:49 And thing diverged pretty badly Aug 24 12:20:29 Also there's third fork of ofono from sfos. Which uses totally different mechanism then ofono upstream of UBports/canonical fork Aug 24 12:21:00 And as far as I'm aware they don't plan to upstrem that Aug 24 12:21:11 Or no immediate plans anyway Aug 24 12:21:25 So question is, is it worth hacking on canonical fork at moment? Aug 24 12:21:40 Because that can't be used as is by plasma mobile or LuneOS Aug 24 12:22:29 Or we invest time in investigating if we can use sfos fork? Aug 24 12:22:39 Or something else. I'm not sure Aug 24 12:22:42 @bhushanshah [Because that can't be used as is by plasma mobile or LuneOS], stop have we problem in plasma? Aug 24 12:22:49 i cant test Aug 24 12:22:57 modem is dead on plasma Aug 24 12:23:11 and i cant open app on plasma Aug 24 12:23:12 @bhushanshah [Because that can't be used as is by plasma mobile or LuneOS], good question! maybe we can start at making it work to speedup porting and then try to fing a common solution Aug 24 12:23:30 @bhushanshah [Because that can't be used as is by plasma mobile or LuneOS], Why not? Aug 24 12:23:44 @bhushanshah [Or we invest time in investigating if we can use sfos fork?], I would prefer us to move away from using forks Aug 24 12:24:10 @dohbee [Why not?], It is outdated fork, and in long term, that won't support Android 8/9 as investigated by the NotKit Aug 24 12:24:31 @Luka Panio [and i cant open app on plasma], Likely different issue. Aug 24 12:24:33 I'd say we should prefer an upstream with financial backing behind them, which either can be "official" ofono of Jolla's version Aug 24 12:24:58 (assuming "official" is still paid) Aug 24 12:25:39 @bhushanshah [It is outdated fork, and in long term, that won't support Android 8/9 as investi …], I think those are two separate issues, and separate from what we need in immediate near-term Aug 24 12:26:03 upstreaming current changes would be hard, no doubt, but doable Aug 24 12:26:24 @dohbee [I think those are two separate issues, and separate from what we need in immedia …], I mean well, I'm fine with UBports investing time in fixing their issues. But if you want Halium level solution, then it is useless for me Aug 24 12:27:23 something that speaks for the UBports variant: it has shipping/shipped devices with full support for at least up to Android 7.1 devices Aug 24 12:27:34 @bhushanshah [I mean well, I'm fine with UBports investing time in fixing their issues. But if …], well, ofono is not in the halium side, but rootfs; so plasma or luneos could ship different ofono version if needed, no? Aug 24 12:27:35 @bhushanshah [I mean well, I'm fine with UBports investing time in fixing their issues. But if …], Since halium is the common part of all the project, what would you suggest either for halium point of view and plasma mobile's one? Aug 24 12:28:50 @stefanomelchior [Since halium is the common part of all the project, what would you suggest eithe …], I'm totally unsure what would be solution tbh. For PlaMo I would prefer to use fully upstreamed solution. Aug 24 12:28:59 @fredldotme [(assuming "official" is still paid)], well, kernel.org seems actively developed. 1.30 release was 12 days ago Aug 24 12:29:10 good to know Aug 24 12:29:15 thanks for checking Aug 24 12:29:28 @dohbee [well, ofono is not in the halium side, but rootfs; so plasma or luneos could shi …], They can, sure. Ofono is not at all part of Halium scope, but it would be good to have something common working 😊 Aug 24 12:30:07 I guess the question is how many halium devices are out there with ofono working? Aug 24 12:30:09 Pinging some people btw, CC: @peat_psuwit @NotKit Herrie Aug 24 12:30:34 @bhushanshah [They can, sure. Ofono is not at all part of Halium scope, but it would be good t …], yes, and i think we should move to tracking true upstream ofono (git.kernel.org version), and getting our fixes upstreamed, or at worst case shipped as "distro patches" or external plug-ins, rather than forks Aug 24 12:30:41 @fredldotme [I guess the question is how many halium devices are out there with ofono working …], mine cedric Aug 24 12:31:12 specifically, are there Plasma Mobile devices with ofono working? Aug 24 12:31:36 maybe we could also work with jolla to get their ofono changes upstreamed, later Aug 24 12:31:48 @fredldotme [specifically, are there Plasma Mobile devices with ofono working?], not mine Aug 24 12:32:05 @dohbee [maybe we could also work with jolla to get their ofono changes upstreamed, later], We'd need to ask them through their IRC and together.jolla websites Aug 24 12:32:55 at least they are cooperative on other components, there's a chance they'd want to do the same for ofono. Aug 24 12:32:57 @fredldotme [We'd need to ask them through their IRC and together.jolla websites], right, and they have different needs from us, so i think we shouldn't worry too much about it yet, and concentrate on our problems related to switching to upstrearm ofono Aug 24 12:33:07 it's hard to take a decision right now, maybe it would be better to share with marius, dalton, florian during Q&A session tonight Aug 24 12:33:36 the questions are what do we need short-term and what would we like to see long-term Aug 24 12:34:19 Maybe it would be good to listen to @fredldotme to explain how he made audio available on calls, this is the UT solution Aug 24 12:34:26 @fredldotme [the questions are what do we need short-term and what would we like to see long- …], right. i think short term we are stuck with current forked ofono (not least because of mtk support). and medium/long term, we want to move to upstream ofono Aug 24 12:34:44 @stefanomelchior [Maybe it would be good to listen to @fredldotme to explain how he made audio ava …], yep, and outside ofono itself :) Aug 24 12:35:04 long term on our (Ubuntu Touch) side is unclear as to how long we are going to support our current MTK devices altogether Aug 24 12:35:44 @stefanomelchior [Maybe it would be good to listen to @fredldotme to explain how he made audio ava …], https://github.com/ubports/telepathy-ofono/commit/d0524cc66f032c7d6453ff84abd0501ed428f1d5 Aug 24 12:35:50 I've upstrem ofono working to some extent on Nexus 5X Aug 24 12:36:01 @bhushanshah [I've upstrem ofono working to some extent on Nexus 5X], what's missing? Aug 24 12:36:20 Fetches network, is able to initiate calls, but that makes whole thing crasg Aug 24 12:36:22 @stefanomelchior [Maybe it would be good to listen to @fredldotme to explain how he made audio ava …], I could forward my late night ubports_porting advertisement here Aug 24 12:36:24 [Edit] Fetches network, is able to initiate calls, but that makes whole thing crash Aug 24 12:36:29 And audio doesn't work etc Aug 24 12:36:43 @fredldotme [long term on our (Ubuntu Touch) side is unclear as to how long we are going to s …], there has been porting work for other mtk devices though Aug 24 12:37:12 @dohbee [there has been porting work for other mtk devices though], do they also need the Canonical-provided MTK plugins which are currently in our ofono? Aug 24 12:37:46 I also at one point tried to find divergence point for rild in UBports fork and upstrem version but failed horribly as patched in UBports fork were squashed when merging upstream Aug 24 12:37:57 @fredldotme [do they also need the Canonical-provided MTK plugins which are currently in our …], i presume so, to use the modem. and also, the goal is to support devices for as long as we can, even if some devices can't have all new features enabled Aug 24 12:38:41 indeed, but it's open ended currently, we can't set dates right now Aug 24 12:39:41 @bhushanshah [Fetches network, is able to initiate calls, but that makes whole thing crash], are there moments where it does not crash? Aug 24 12:39:59 @fredldotme [indeed, but it's open ended currently, we can't set dates right now], right, but from my point of view, it's irrelevant. we should move forward either way, with the assumption we won't drop support for those devices, and try to keep the useful changes around Aug 24 12:40:39 @dohbee [right, but from my point of view, it's irrelevant. we should move forward either …], alright Aug 24 12:41:13 so, should I forward the information about h2w, telepathy-ofono etc here? Aug 24 12:41:30 I think so Aug 24 12:41:40 Fwd from fredldotme: https://github.com/ubports/telepathy-ofono/commit/d0524cc66f032c7d6453ff84abd0501ed428f1d5 … This makes calling through the earpiece work. … Requires h2w in the kernel. … Also introduces quirks to skip the deep_buffer sink/source. Aug 24 12:41:41 Fwd from fredldotme: Remaining issues: … - Switching to the speakerphone doesn't work … - No dynamic headset switching Aug 24 12:41:41 Fwd from fredldotme: If we can guarantee that every supported device has a h2w switch available I'll go ahead and add dynamic headphone switching. Aug 24 12:41:41 Fwd from fredldotme: Porters without a native h2w switch need to add this kernel driver: https://github.com/ubports/android_kernel_oneplus_msm8974/commit/66b614eb9f53b7d5d4e283908fda771faa4ebee8 Aug 24 12:41:41 Fwd from fredldotme: To install this modified telepathy-ofono: `sudo ubports-qa install xenial_-_haliumfixup` Aug 24 12:41:41 Fwd from fredldotme: To enable the new telepathy-ofono quirks: … ```# telepathy-ofono quirks … PRODUCT_PROPERTY_OVERRIDES += \ … t-o.quirk.forcesinkprimary=1 \ … t-o.quirk.forcesourceprimary=1``` Aug 24 12:41:58 this way it would be clearer to bhushan and others how you fixed it Aug 24 12:42:01 @fredldotme [are there moments where it does not crash?], First outgoing or incoming calls are fine Aug 24 12:42:09 the telepathy-ofono code is just an initial test, nothing I'd merge just yet Aug 24 12:43:27 Have you tried looking into ohm instead of the h2w? Aug 24 12:44:15 @bhushanshah [Have you tried looking into ohm instead of the h2w?], my preference was to keep new components out as long as we have others that do the job, not to have two services fighting over the same job Aug 24 12:44:43 At same time, less kernel patches the better imo Aug 24 12:44:46 for me h2w is ok Aug 24 12:44:57 the idea for telepathy-ofono is to have two backends: Aug 24 12:45:04 without patch Aug 24 12:45:21 evdev input based event reading as well as h2w for backwards compatibility Aug 24 12:46:07 evdev input works on suzu/Xperia X for example, it's just not hooked up in any way Aug 24 12:49:10 the "new" h2w detection implementation will be based on libudev, which currently still makes issues (udev_monitor_new_from_netlink always returning NULL) Aug 24 12:50:08 when i insert jack ihave: Aug 24 12:50:10 [ 581.598814] wcd_mbhc_mech_plug_detect_irq: enter Aug 24 12:50:28 and [ 589.058050] wcd_mbhc_report_plug: Reporting insertion 4(0),zl 55455 ohm,zr 55455 ohm … later Aug 24 12:50:49 on Ubuntu Touch you can test evdev based detection with `evtest` Aug 24 12:52:48 I: Bus=0000 Vendor=0000 Product=0000 Version=0000 … N: Name="msm8952-snd-card Headset Jack" … P: Phys=ALSA … S: Sysfs=/devices/soc/c051000.sound/sound/card0/input8 … U: Uniq= … H: Handlers=alsa_to_h2w event8 … B: PROP=0 … B: EV=21 … B: SW=3c0d4 Aug 24 12:53:06 and Aug 24 12:53:08 I: Bus=0000 Vendor=0000 Product=0000 Version=0000 … N: Name="msm8952-snd-card Button Jack" … P: Phys=ALSA … S: Sysfs=/devices/soc/c051000.sound/sound/card0/input9 … U: Uniq= … H: Handlers=kbd event9 keychord … B: PROP=40 … B: EV=3 … B: KEY=40 0 0 0 0 0 0 0 0 0 f0 4 0 0 0 c0000 0 0 0 Aug 24 12:53:18 sorry for much spam Aug 24 12:53:37 please let's keep the channel for discussions only for now Aug 24 12:57:36 ok, can we use pastebin? Aug 24 12:57:50 @Luka Panio [ok, can we use pastebin?], preferably, as usual ;) Aug 24 12:57:57 @fredldotme [If we can guarantee that every supported device has a h2w switch available I'll …], @bhushanshah actually this will be the only part affecting kernel, and for this reason also Halium project. the other changes are more focused on telepathy-ofono Aug 24 12:59:24 which component is responsible for audio switching & rerouting on Plasma Mobile and LuneOS? Aug 24 12:59:28 @fredldotme [To enable the new telepathy-ofono quirks: … # telepathy-ofono quirks … PRODUCT_PROPE …], and these quirks as well (device.mk) Aug 24 13:00:05 @fredldotme [which component is responsible for audio switching & rerouting on Plasma Mobile …], PlaMo uses telepathy-ofono Aug 24 13:00:32 While LuneOS uses voicecall iirc Aug 24 13:02:28 @bhushanshah [PlaMo uses telepathy-ofono], at least we can share tips&tricks with Plasma Mobile Aug 24 13:02:44 How is headset detection done on PlaMo, does pulseaudio handle it by itself? Aug 24 13:02:56 outside of calls that is Aug 24 13:03:12 [Edit] outside of calls as well as during calls that is Aug 24 13:06:06 I am half here, doing b'day prep for my son's 5th b'day tomorrow Aug 24 13:06:37 happy birthday from the internet to your son then :) Aug 24 13:07:11 @fredldotme [How is headset detection done on PlaMo, does pulseaudio handle it by itself?], In fairness, I haven't tested this Aug 24 13:07:33 On LuneOS we only have 5.1 Android devices working using Mer's oFono but with legacy RILmodem driver, not the new Mer one. 7.1 has issues on our Xiaomi devices, probably due to QMI Ril, but not 100% sure yet. Trying to get Hammerhead 7.1 based port working for comparison. Aug 24 13:07:46 We suspect it's something Xiaomi specific somehow. Aug 24 13:08:07 you mean rild not doing its job? Aug 24 13:08:29 With some signals not being sent that oFono expects. But this is already on incoming and outgoing calls, before pa kicks in it seems Aug 24 13:08:42 Yeah the QMI Ril Seems problematic a bit Aug 24 13:08:53 Or at least the binary we have for our target devices Aug 24 13:10:07 so on ut swithing audio is problem Aug 24 13:10:13 which Xiaomi device is that? do we have someone in Ubuntu Touch land with the same device? Aug 24 13:10:30 @Luka Panio [so on ut swithing audio is problem], yes, ofono does its job fairly well Aug 24 13:11:46 @fredldotme [yes, ofono does its job fairly well], yes with my pactl commands or your mod working perfect and usable Aug 24 13:12:02 Tissot, Rosy, Mido Aug 24 13:13:41 All SD 625 based Aug 24 13:20:14 ever tried using UBports' ofono? Aug 24 13:20:47 @Luka Panio [yes with my pactl commands or your mod working perfect and usable], My current experience with tenshi (msm8937, BQ Aquaris U Plus) is that inbound/outbound call audio works well with 'pactl' configuration (https://paste.ubuntu.com/p/fVzDNdDwNc/) before each call Aug 24 13:21:47 at the moment `h2w_switch` doesn-t work on my UT porting, maybe due to initializing twice h2w Aug 24 13:22:13 better to adapt Alfred-s code to my msm8937 Aug 24 13:23:36 for the custom h2w driver to work correctly one needs to investigate which sound driver is the proper one and only call the h2w initialization there instead of for every jack device that the kernel finds. Aug 24 13:24:55 @fredldotme [for the custom h2w driver to work correctly one needs to investigate which sound …], yep, we need more work, but if it makes sense for the majority of devices, it can be a good combined solution (at least for those who use ofono) Aug 24 13:26:22 tbf the h2w switch stuff is just a short-term solution to a larger problem. what we need is proper detection of headsets as well as microphones. h2w can only decide between jack being plugged in or not, where as evdev allows us to check if the microphone is also in use. Aug 24 13:26:46 at least on suzu where there's no distinction between headset with microphone or headphones without microphone Aug 24 13:29:02 Event: time 1566651404.734685, type 5 (EV_SW), code 2 (SW_HEADPHONE_INSERT), value 0 … Event: time 1566651404.734685, type 5 (EV_SW), code 4 (SW_MICROPHONE_INSERT), value 0 … Event: time 1566651404.734685, type 5 (EV_SW), code 7 (SW_JACK_PHYSICAL_INSERT), value 0 … Event: time 1566651404.734685, -------------- SYN_REPORT ------------ Aug 24 13:29:08 for me Aug 24 13:30:10 I was wondering if this piece of code (https://paste.ubuntu.com/p/JxvWM7pCNY/) of `ofonocallchannel.cpp` can be modified to first set default mode the `speaker`, then checking for bluez or `AudioModeWiredOrEarpiece` Aug 24 13:30:56 might it be a temporary solution to adapt telepathy/ofono, rather to use `h2w?switch` Aug 24 13:31:22 Event: time 1566653360.333548, -------------- SYN_REPORT ------------ … Event: time 1566653462.445380, type 5 (EV_SW), code 2 (SW_HEADPHONE_INSERT), value 1 … Event: time 1566653462.445380, type 5 (EV_SW), code 7 (SW_JACK_PHYSICAL_INSERT), value 1 … Event: time 1566653462.445380, -------------- SYN_REPORT ------------ Aug 24 13:31:34 yes it knows have it mic or no Aug 24 13:33:13 is it same for tensi? Aug 24 13:33:45 @stefanomelchior [might it be a temporary solution to adapt telepathy/ofono, rather to use h2w?swi …], something like this> https://paste.ubuntu.com/p/TYTkbWPFT2/ Aug 24 13:35:43 you could try it, telepathy-ofono is one of the few projects that build fine with `crossbuilder` Aug 24 13:35:59 @Luka Panio [Event: time 1566653360.333548, -------------- SYN_REPORT ------------ … Event: tim …], when I remove and insert a mic equipped headset: https://paste.ubuntu.com/p/fgfs5MrtsZ/ Aug 24 13:36:46 the problem with upstream ofono is that ofono itself is developed, but nobody works on RIL plugin Aug 24 13:37:21 and Jolla's ofono has *completely* different RIL plugin implementation Aug 24 13:37:43 but it is maintained, including newer devices support Aug 24 13:38:12 @Luka Panio [Event: time 1566653360.333548, -------------- SYN_REPORT ------------ … Event: tim …], it is without mic Aug 24 13:38:16 current UBPorts RIL plugin won't work for 8/9 and newer as they changed old socket transport for binder instead Aug 24 13:38:23 so are you suggesting building on top of upstream ofono with Jolla's ril plugin? Aug 24 13:38:55 or just Jolla's ofono for the moment? They are rebasing on top of upstream themselves Aug 24 13:39:20 then we would need to carry the MTK patches ourselves for the moment Aug 24 13:39:43 that means having two plugins... Aug 24 13:39:59 Canonical's and Jolla's Aug 24 13:40:36 @Luka Panio [it is without mic], yep, it is a headset with mic Aug 24 13:40:38 as @dohbee suggested it would be beneficial to keep our current lineup of devices compatible for the forseeable future Aug 24 13:41:14 I'm wondering, assuming we want to provide an override, how would we do so for devices that have a closed BSP? Aug 24 13:41:27 anyway, I have experimented with buulding Jolla ofono, it is on CI Aug 24 13:41:42 @fredldotme [I'm wondering, assuming we want to provide an override, how would we do so for d …], it was done by repacking image before Aug 24 13:41:58 Well Jolla's is too specific with MCE etc, that would need to be stripped Aug 24 13:42:17 good to know, sounds reasonable, we could then provide an Upstart override for where the Canonical ril plugin is selected on MTK devices Aug 24 13:43:20 @NotKit [anyway, I have experimented with buulding Jolla ofono, it is on CI], which is? xenial_-_merofono? Aug 24 13:43:36 yes Aug 24 13:44:29 Herries, yes, that's the problem with it as well Aug 24 13:44:52 I noticed the Jolla Ofono adds quite a few additional dependencies, therefore I didn't import it into debian-pm yet. Herrie do you know if any of those can be dropped? Aug 24 13:45:02 Would it make sense to talk to Jolla about making mce integration optional? Aug 24 13:45:56 Well we added the dependencies at our end for now. Most are small libs Aug 24 13:45:59 maybe add more ifdefs in PR? but I didn't check how much reliant is it yet Aug 24 13:46:41 probably glib libs will have to stay, but MCE is useless at our side Aug 24 13:48:22 @Herrie [Well we added the dependencies at our end for now. Most are small libs], I also have everything packaged locally but didn't set up the ci yet. In case the decision here ends up being in favor of using Jolla ofono, we should get that done for Plasma Mobile pretty quickly then. Aug 24 13:49:33 Yeah Mer has been pretty OK with adding build flags etc Aug 24 13:49:56 We PR'd some VoiceCall stuff so we can build it without needing to maintain patches Aug 24 13:51:23 We can send a few more things upstream for the libs at our end so they build with Yocto without patches for toolchain Aug 24 13:51:35 MCE useless for us too Aug 24 13:52:11 But either merging legacy RILmodem or having Jolla's upstream would be good Aug 24 13:52:31 Seeing Mer is quite some oFono versions behind again Aug 24 13:52:46 @NotKit [yes], Got an error> https://paste.ubuntu.com/p/q4HXZdpFS9/ Aug 24 13:56:54 It seems it could be reasonable to converge to Jolla's for all the project and adapt code to support legacy devices, on top of that agreeing MCE is useless for the majority Aug 24 13:57:55 [Edit] It seems it could be reasonable to converge to Jolla's for all the projects and adapt code to support legacy devices, on top of that agreeing MCE is useless for the majority Aug 24 14:09:04 @stefanomelchior [It seems it could be reasonable to converge to Jolla's for all the projects and …], will we share these thoughts with UT leadership team tonight? Aug 24 14:09:29 would be an interesting after-party topic Aug 24 14:10:11 Jolla got 8.x/9.x based working as well, so RIL wise it should be pretty OK Aug 24 14:11:52 it would be very intersting for future porting to assure a certain kind of continuity Aug 24 14:22:07 @stefanomelchior [it would be very intersting for future porting to assure a certain kind of conti …], Hire people full time for continuous improvement Aug 24 14:23:15 Hi guys, Simon from plasma mobile here. Is it possibility to have external plugins for ofono, ie take upstream ofono and have a separate package for the ril plugin? Aug 24 14:24:36 theoretically possible, though may need changes to make it doable; but it wouldn't solve the core issue with the rilmodem differences and needing to ship both Aug 24 14:27:22 @Flohack [Hire people full time for continuous improvement], I meant relying on hired people (Jolla's) and on their 8.x/9.x continous support can be useful even for a community project, as UT, where atm there are few efforts to explore haium-8.x/9 porting Aug 24 14:30:57 @stefanomelchior [I meant relying on hired people (Jolla's) and on their 8.x/9.x continous support …], i would say there's more blockers than lack of effort Aug 24 14:38:43 @dohbee [theoretically possible, though may need changes to make it doable; but it wouldn …], a so there are two different versions of the rilmodem? what part is that actually? Is there an overview somewhere how actual hardware and ofono connect? Aug 24 14:40:48 @Simon [a so there are two different versions of the rilmodem? what part is that actuall …], as mentioned before, mer (sailfish) has a different implementation of ril in their fork. i don't know if there's a shiny overview, but the docs in ofono probably have some info Aug 24 14:41:53 it would also be nice to get some feedback from ubports people on https://github.com/TelepathyIM/telepathy-ofono/issues/3 or https://github.com/ubports/telepathy-ofono/issues/4 Aug 24 14:45:02 @Simon [a so there are two different versions of the rilmodem? what part is that actuall …], Did you see this at the beginning? … https://wiki.ubuntu.com/Touch/Audio?action=AttachFile&do=get&target=Ubuntu+Touch+-+Audio+Architecture+Overview.png Aug 24 14:46:06 https://github.com/mer-hybris/ofono-ril-binder-plugin how does this one fit in? has anybody tried it? Aug 24 14:50:02 Mer's documentation could be better, I've found myself wondering on a number of occasions what components are being used for Aug 24 14:51:05 @John (Marlin) [Did you see this at the beginning? … https://wiki.ubuntu.com/Touch/Audio?action=A …], that just says "ofono" but not which plugin talks via what to the hardware Aug 24 14:52:29 @Herrie [Mer's documentation could be better, I've found myself wondering on a number of …], I guess a great outcome of this would be better docs and some hints how to debug everything Aug 24 15:14:17 Problem with most devs is that they're not very keen on writing documentation. I'm not great at it either but try at least sometimes ;) Aug 24 15:18:29 The best time to write documentation is right after you did something for the first time Aug 24 15:18:32 As you discover Aug 24 15:18:58 That creates the most rich documentation that answers the questions you had Aug 24 15:19:10 Going back and doing it after is... less useful Aug 24 15:27:46 I think there is no more contribution for the ofono hackathon atm: can we stop it here? or would anyone add anything else? Aug 24 15:37:30 @Simon [https://github.com/mer-hybris/ofono-ril-binder-plugin how does this one fit in? …], I think it's a plugin for their ofono RIL driver Aug 24 16:10:36 @NotKit: but I don't understand, in their "ril" ofono plugin, it already talks to the RIL socket there Aug 24 16:11:24 (I tried it on LuneOS, same result as with upstream oFono regarding our "tissot" target) Aug 24 18:14:18 anu updates on ofono? Aug 24 18:20:59 becouse only problem is switching Aug 24 18:25:37 Tofe, that plugin is for binder, not socket Aug 24 20:09:16 binder seems to be the android equivalent to dbus? Aug 24 20:09:46 yes **** BEGIN LOGGING AT Sat Aug 24 20:28:34 2019 Aug 24 21:29:41 Does Pm use F2fs? Aug 24 21:32:25 Does Pm use ext4? **** ENDING LOGGING AT Sun Aug 25 02:59:57 2019