07:00:12 #startmeeting Sailfish OS, open source, collaboration -- 27th October 2022 07:00:12 Meeting started Thu Oct 27 07:00:12 2022 UTC. The chair is flypig. Information about MeetBot at http://wiki.debian.org/MeetBot. 07:00:12 Useful Commands: #action #agreed #help #info #idea #link #topic. 07:00:19 #info Meeting information and agenda can be found here: 07:00:23 #link https://forum.sailfishos.org/t/community-meeting-on-irc-27th-october-2022/13252 07:00:29 I am the meeting's chairperson today, and will be doing my best to keep time and order. Please respect the timings and no twisting my melon, man. 07:00:39 #topic Brief introduction (5 min). Please prefix your name/handle with #info 07:00:49 #info Otto Mäkelä, community 07:01:06 #info sebix, community 07:01:07 #info poetaster Mark, community 07:01:21 #info David Llewellyn-Jones, sailor @ jolla & meeting chair 07:02:01 #info Jacek Jagosz, community 07:02:47 #info Björn Bidar, Sailor @ Jolla 07:03:05 It's great that we have some of the question-askers here 07:03:29 #info Simonas Leleiva -- privateer for Jolla 07:04:15 Looks like it's going to be a win for the community today! 07:05:36 ViGe: Where are you? ;D 07:05:44 A 4:3 win for the community: good game! Please feel free to add your name later if you arrive late. Let's go to the first question. 07:05:58 #info Ville Nummela, sailor @ Jolla 07:06:09 After the final whistle! 07:06:23 First up is your question JacekJagosz[m]. 07:06:24 :( 07:06:26 #topic What is the progress on camera fixes on newer devices? (10 mins -- asked by JacekJagosz) 07:06:33 #info Currently there are 2 issues with cameras on SFOS. 07:06:38 #info 1st is "Camera App crashes when trying to record a video" which affected not just X10 III, but also Volla 22/GS5. 07:06:42 #link https://forum.sailfishos.org/t/4-4-0-64-xperia-10-iii-camera-app-crashes-when-trying-to-record-a-video/12118/ 07:06:47 #info It got fixed by extending firejail whitelist and limiting maximum resolution. 07:06:52 #link https://github.com/HelloVolla/droid-config-halium-mimameid/commit/79cc6086200fe8bc5da54797c0d7e5c61aaa060e 07:06:55 #link https://github.com/HelloVolla/droid-config-halium-mimameid/commit/99d333a9be49e3c897125fb95ab95018dd97e0b6 07:07:02 #info Could the solution be the same on the 10 III, and not a problem with Sony's adaptation? 07:07:08 #link https://github.com/sonyxperiadev/bug_tracker/issues/761 07:07:12 #info Another issue is that 2 cameras on 10 III are inaccessible, because of no Camera2API support on SFOS. I know you wanted someone from community to work on it, but so far no one has stepped up. 07:07:21 #info Has there been any progress on any of those? Because from a user's perspective there is none, and not working camera is not great in a new mobile phone. 07:07:48 Thanks for your question JacekJagosz[m]. Here's our prepared answer. 07:07:49 #info First, for video recording, we appreciate the input, and for highlighting that it might be connected to the similar issue on the Volla. 07:07:57 #info We currently believe that on the Xperia 10 III it's an issue with the upstream adaptation. 07:08:02 #info However, we have an internal issue about it and are liaising with Sony, and if there's any progress we'll aim to share it on the forum. 07:08:11 #info Second, concerning the Camera2 API. We realise full support for all three cameras is important for users, but as you point out, there's not yet been any publicly visible progress. 07:08:22 #info In the longer term this might make a good Sailmates project. 07:08:26 #info In the shorter term, we were hoping for someone from the community to take the initiative, even just to create a minimal non-working PR to crystalise effort. We'd welcome thoughts on this. 07:08:56 That's the answer. Thaodan, I believe you created the bug on Sony's public tracker; would you like to add anything? 07:09:25 Yeah I was already about to answer.. 07:10:32 Further question: If Sony makes corrections/upgrades to their binary blobs (on which the camera system on SFOS/Xperia 10III depends, if I've understood correctly), is there a mechanism to upgrade these? 07:10:51 First: jacekjagosz mixes two issues up that have similar symptoms. The Volla issue is different some of the libraries used in the hal where not whitelisted. 07:11:13 ExTechOp: Please ask later the question is another issue. 07:11:29 OK 07:12:12 On to the Xperia 10 III: The Sony camera blob crashes which gives the impression for the user that doesn't see this that something with the camera app is wrong. 07:12:51 closing the camera app or turning the screen on and off is usually enough to renew the connection to the camera android daemon after it was restarted. 07:14:11 The remaining thing that we can do as Jolla or the community is to look if some configuration is missing that should have been imported from Sony's stock rom while porting. 07:14:24 Any ideas why no AOSP user reported this bug? 07:14:30 Only seen on SFOS? 07:14:53 However since the Sony stock and the Sony AOSP camera stack is so different the issue might be really in the qualcomm blob. 07:15:17 The fact that SFOS uses CameraAPI, and AOSP Camera2API? 07:15:36 Sony AOSP uses Qualcomms default camera api while Sony stock uses their own Camera api on top of Qualcomm stock camera api. 07:15:41 No! 07:16:04 The issue also happens on AOSP, we simply noticed it during testing. 07:16:13 Oh, interesting 07:16:18 Didn't know 07:16:37 Question, where to start looking at the current Camera api to compare with 2? Is this not directly related to the version of QT ? 07:17:07 We finished some parts of the camera part that were missing from the Sony initial port e.g. the hardware enconding for the video recording and encoding. 07:17:31 2 more mins on this topic. 07:18:05 So tldr: the issue is a android issue not related to Sailfish OS, we are waiting for Sony for help, community could try to find things we missed in the open source part. 07:18:11 so done :) 07:18:16 mal, do you have any thoughts on poetaster's question about the Camera2 API? 07:18:38 Also the issue isn't related to Camera1 or Camera2. 07:18:53 so that is entirely unrelated. 07:18:56 Thanks for your detailed explanation Thaodan. 07:19:25 Thank you so much, that makes much more sense! 07:19:26 If we can reproduce an issue on AOSP in the same or slightly different way than Sailfish OS is out of the discussion. 07:19:34 Perhaps we can talk about Camera2 API in the General section, as we already hit time. Is this okay JacekJagosz[m]? 07:19:59 That would be great if we could do that later 07:20:04 Ok by me. 07:20:17 Great, thanks JacekJagosz[m]. Let's move on then. Thanks for the nice discussion on this. 07:20:21 #topic state of harbour for app submitters (10 mins -- asked by thigg) 07:20:27 #info - Does jolla track metrics about the duration for approval/feedback in harbour? 07:20:35 #info - It seems sometimes approval takes multiple weeks, are there plans to change this? 07:20:40 #info - Is there any priority on / are there plans for the dashboard for app submitters in harbour? (e.g. history of app downloads). 07:20:47 #info - What would be the preferred way for third party tools to interact with harbour (e.g. do statistics ourselfs). They would need to authenticate. Is there an easy way for scripts to do that? 07:20:58 #link https://forum.sailfishos.org/t/harbour-dashboard-stats-data-gatherer/12520 07:21:05 Here's our prepared answer. 07:21:09 #info There are several parts to your question; I'll try to answer them all. 07:21:15 #info 1. We don't routinely collect app review statistics, but I did a quick check and over the last two months the average (mean) approval time for apps submitted to harbour was under four days. 07:21:26 #info 2. Typically we would expect app review to be handled within 1 week. If you've not received feedback after a week, please send me (flypig) a direct message via the forum and I will follow it up for you. 07:21:38 #info 3. There have unfortunately been a few apps which have taken longer than this. Your Flood to Java app was one of these, and for this we apologise. 07:21:46 #info This usually happens because of additional checks needed to review the app, or due to special requests made by the developer. 07:21:54 #info 4. Apps are reviewed in the order they're submitted, unless there's some specific reason why this can't be done (e.g. a special request from the developer). 07:22:03 #info 5. There are no current plans to make the harbour API available for wider use, but we note your request for one and will keep it in mind! 07:22:09 As a personal note, it would be great to have the API but that would require non-trivial technical changes. 07:22:15 However, I spoke to the testing team and they were very receptive to suggestions for improving the process. If you have ideas please share them. 07:23:05 Do we have thigg here this morning? If not, if anyone else has any suggestions or comments, please do share them. 07:24:34 Next :) 07:24:53 #info Anton Thomasson - very late, peeking in from hackathon at work 07:25:08 Ah, good that you could make it for your question attah_pine. 07:25:10 I think some suggested more involvement of the community obs in harbour. Running checks on the package during the obs stage can already help a developer (or inside the sdk). These checks also can be used in the initial ci checks for harbour apps. 07:25:29 I think flypig already has done that manually for one of my apps. 07:26:32 I've gotten more support in harbour than I deserve, frankly. 07:27:21 In my experienced the review team do a great job, but any opportunity for more automation, especially on obs/sdk as you suggest Thaodan, has to be good. 07:28:09 Yes they do however we can drop much of the work by automating a great chunk of the tasks. 07:28:23 the sdk already runs the checks during the building, if you build from Qt Creator 07:28:36 release time in harbour has been a major turnoff for many 07:28:43 ViGe, I guess you have to select the "Deploy" option? 07:28:45 ViGe: Mentioned above 07:29:06 Hey, sorry to be late. Thanks for the answer :) 07:29:17 poetaster: Yeah support in harbour is a thing, usually you should report to the projects page. 07:29:17 Ah, great that you made it thilo[m]! 07:29:22 Thaodan: Just wanted to make it clear, as what you wrote could be interpreted as a feature request 07:29:58 thilo[m], did you want to add/ask anything further? 07:30:00 ViGe: Hm you're right, I should have written that is already done. 07:30:05 flypig: Yes, the checks are part of deployment (and not building), that's correct 07:30:27 The sdk and obs already run those checks my point was about taking them into account during harbour validation. 07:31:00 Some of the checks are invalid for harbour apps as they validated packaging rules such as shared libraries in /usr/share 07:31:11 just take that into account as a developer. 07:31:19 2 more mins. I know poetaster is itching to move on ;) 07:31:29 How that is done in the ide could be a bit better explained in the docs iirc, i remember myself searching around 07:31:36 Thaodan, but obs is != sdk. there are gottchas. I didn't notice obs wasn't building translations, for instance. 07:31:44 #info fridlmue - community. Also very late... 07:31:59 Flypig: no, nothing to add 07:32:00 I'm fine with taking the time. 07:32:09 Nice to see you fridl. 07:32:14 poetaster: In case of checks are are equal 07:32:18 but yeah lets move on 07:32:40 Okay, we hit time, let's move on to the next question. Thanks all. 07:32:44 #topic Can libudev be allowed to be linked against for harbour apps? (10 mins -- asked by @poetaster) 07:32:51 #info In the main, to support building the Godot game dev env. 07:32:57 #link https://forum.sailfishos.org/t/harbour-is-libudev-really-need-to-be-depricated-in-harbour/13304 07:33:00 #link https://github.com/sailfishos/sdk-harbour-rpmvalidator/issues/168 07:33:07 #info I'm not sure how often it's used since the peripherals for cellphones are not that many :) Nevertheless, for game development and considering future, hopefully, tablets, conceivable for game controller support. 07:33:18 #info And Godot is fun. 07:33:27 Here's our prepared answer. 07:33:29 #info Thanks for your question poetaster, as well as for both your and sashikknox’s work on the Godot Engine. We love a bit of Flappy Bird (current best L3 51) and would love to see more native Sailfish OS games. 07:33:43 #info We’ve been considering allowing libudev linking for some time, but it requires more than just a new line in allowed_libraries.conf. 07:33:50 #info Controller access will need to be added to either an existing permission or a new permission. 07:33:58 #info At this point, we can’t guarantee that libudev can be supported, but we’d welcome input on its potential uses, and how best to handle permissions for it. 07:34:34 That's the prepared answer, but pvuorela has been looking into this and I'm hoping might be able to add a bit more. 07:34:48 It is an end run around Sailjail, then? Like dbus? 07:36:16 With the 3.5 release sashikknox put a lot of effort into 1. harbour validation and 2. sailjail support. which work really well (except on volla). So the motivation is there. 07:37:04 poetaster, is the libudev inclusion the only remaining change needed for harbour validation? 07:37:19 yes, I believe so. 07:37:29 yea, so on permission side there seem to be multiple problematic details. a) would need to get access to libudev, b) would also need access to the joypad devices but only them, c) the access to the /dev node could be problematic, there's resource policy handling, eg two apps wanting to use the same device. 07:38:24 compare to e.g. bluetooth keyboard which doesn't need any permissions and goes over wayland. might be nice if there was some similar wayland protocol for this. did find some preliminary work but that was getting old. 07:39:10 as sashikknox mentioned on the forum, there is the detour over SDL. 07:39:33 poetaster, would that avoid the resource sharing issues? I'm guessing probably not. 07:39:46 I think if there's no api for such a thing right now, just adding the permissions for these certain devices files might work. Newer Wayland APIs are another can worms as we are already quite behind in terms of Wayland APIs. 07:40:02 If I understood correctly, SDL would allow harbour validation, but wouldn't necessarily address the Sailjail question. 07:40:12 and that's a lot of extra labour. 07:40:15 hm, how was sdl2 handling this? 07:40:15 flypig, poetaster If sdl doesn't use the hardware directly. 07:40:46 My understanding is that SDL is a backdoor, but I'm probably wrong. 07:41:09 is the api called sdl joystick? 07:41:23 On the forum, sashikknox doesn't explain much about it, only that it's possible to "rewrite all udev part in godot to SDL2 ( long way, dont have resources for this )" 07:42:39 So pvuorela, is Sailjail access to /dev etc. a no go? And is there any realistic prospect of offering an alternative? 07:42:40 There is a joystick library in SDL, but I don't know the godot internals well enough to comment. 07:42:53 The api also needs to be packaged. 07:43:41 flypig: at least that doesn't sound nice and that longer term way would be more indirect access. 07:44:09 I think sdl also uses direct hardware access 07:44:10 https://wiki.libsdl.org/SDL_JoystickOpen 07:44:20 more prone to drift from upstream, more bugs, more labour. 07:44:43 2 more mins on this. 07:45:04 poetaster: Sure, just trying to see if sdl does it any different. 07:45:06 Thaodan: the api looks like something that could suit both direct or indirect access. 07:46:17 poetaster, we hit time. Do you want more time on this? 07:46:37 Or is there anything to add to the minutes? 07:47:17 It's fine, I'm proxy for sashikknox. I'll take into account a,b,c from pvuorela and take it up with sash 07:47:23 quite a bunch of different joystick drivers there https://github.com/bminor/SDL/blob/master/src/joystick/SDL_joystick.c 07:47:31 pvuorela: depends on he driver it seams https://github.com/libsdl-org/SDL/blob/a6018ae57fb26140777f43c13a6599d710183c86/src/joystick/SDL_joystick.c#L524 07:47:42 #info SDL info 07:47:44 #link https://wiki.libsdl.org/SDL_JoystickOpen 07:47:58 #info Joystick drivers 07:48:00 Please use the upstream repo for the links 07:48:08 #link hhttps://github.com/libsdl-org/SDL/blob/a6018ae57fb26140777f43c13a6599d710183c86/src/joystick/SDL_joystick.c#L524 07:48:12 https://github.com/libsdl-org/SDL/ is upstream 07:48:51 Okay, let's move on. This discussion suggests it might be good to discuss further on the forum. Thanks all. 07:48:55 In any case, I can't overstate the value of godot for attracting another strata of developers. 07:49:09 poetaster: good point. 07:49:16 #info I can't overstate the value of godot for attracting another strata of developers. 07:49:24 #topic Ownership and secret engineering project status (10 mins -- asked by attah) 07:49:30 #info How are things going? Should we be worried? Updates on ownership were estimated for summer i seem to recall, so i can't help to think it didn't go exactly according to plan. 07:49:40 #info On the upside, development activity on the open parts seems to have picked up speed again, so presumably the secret engineering project has concluded. Is there anything you can share about that? 07:49:51 We have a prepared answer of course... 07:49:57 #info It is true indeed that the ownership restructuring in regards the Russian-owned Jolla shares has taken longer than estimated previously but it is under execution and will be fully completed by the end of the year. 07:50:06 #info As a result, Jolla will have 0% Russian ownership. For clarity, business to Russia was already stopped in 2021. 07:50:26 That's the answer, but we have attah_pine here which is great. Would you like to add anything attah_pine? 07:50:47 That sounds great! 07:50:48 (or anyone else, of course) 07:51:10 And i guess we will find out exactly who the new owners are in due time? 07:52:08 And what about the secret engineering project? :) 07:52:18 ;) they dodged it 07:53:05 nah, that was godot :) 07:53:29 attah_pine, you can expect announcements about ownership in due course, but we'll have to wait and see what/how much detail is given. 07:54:44 attah_pine, I'm not exactly sure what you were referring to with the secret engineering project. I was hoping you could explain. 07:55:33 flypig: I think that wording was used by Jolla in some communication about changes in the release cycle. 07:55:40 Sure :) In about Q2 of this year SFOS development was halted in favor of some secret engineering project 07:55:46 (by the way, when I say "you can expect announcements", I'm making assumptions. I just imagine that there will be). 07:56:00 It is so secret, its even hidden from flypig 07:56:40 Popular assumption Qt6, my assumption Android support for Mercedes 07:56:48 attah_pine: there are always secret engineering projects :D 07:57:07 Okay, I think I know the communication you're referring to. I'm not sure the term "secret project" was mentioned there though! 07:57:21 flypig, no, that's almost a quote. 07:57:36 It stemmed from your pen, I believe. 07:57:43 Really? Oh, okay! 07:57:51 For sure... but here it was announced publicly that regular development would be put on hold 07:58:24 I think "engineering project" is a quote... secret is perhaps redundant 07:58:37 My recollection is that we said that there may be some reduction in community activity to focus on engineering work. 07:58:58 ...and SFOS releases 07:59:29 ah, I may have picked it up from attah :) 07:59:40 I asked once before in the same terms 08:00:39 I'm searching for the post, but I can't find it now... 08:00:48 I'm quite sure you guys are referring to https://forum.sailfishos.org/t/sailfish-community-news-24-march-games/10779/1 08:01:34 Oh, nicely found ViGe. 08:02:04 Yes, indeed 08:02:59 We hit time on this topic I'm afraid. Does that clear things up? Or do you want more time attah_pine? 08:03:44 Well, partly... :) 08:03:56 Thanks for your time 08:04:22 Sorry if we didn't clear everything up for your, but hopefully it was informative anyway! 08:04:53 We have just our usual PR, bug and General topics now. 08:04:59 #topic Open PR discussion (5 mins -- asked by jolla) 08:05:02 #info Any open PRs to discuss? 08:05:19 There weren't any submitted in advance, but maybe someone would like to raise one now? 08:08:27 Okay, I guess there are no PRs to discuss. That's not a bad thing :) Let's move on to the next. 08:08:33 #topic Untracked bug reports (5 min -- asked by pherjung) 08:08:40 #info The Community Bug Coordination Team have done a superb job once again this fortnight. 08:08:44 #info 7 high quality bug reports now recorded internally and tagged as "tracked". 08:08:49 #info 1 bug reports tagged as "fixed". 08:08:53 #info 2 marked as duplicates and closed. 08:09:03 #info 1 bug (12885) left unchanged in case there’s any more discussion on it. 08:09:34 Thanks for your amazing efforts once again. I think the report was submitted by you this time thilo[m]? 08:11:01 thanks a lot! 08:11:25 Alright, I guess we don't need to discuss further. Let's go on to the General discussion. 08:11:27 #topic General discussion (15 min) 08:12:04 I think Camera2 was a topic we said we'd return to. JacekJagosz[m] did you have further thoughts to share on that? 08:13:41 I found the answer to my question: look at: https://github.com/sailfishos/droidmedia/blob/master/droidmediacamera.h 08:13:42 poetaster: Regarding the joystick access in Firejail: there was a proposed wayland protocol for this, for flatpak it was suggested to add /dev/input/js* to the whitelist. 08:13:43 https://github.com/flatpak/flatpak/issues/7 08:13:54 https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/110 08:14:21 So just allowing the use of /dev/input/js* might be a good workaround for now. 08:14:37 poetaster, could you remind me of your question that this answers? Is this something to add to the minutes? 08:15:28 #info (re Sailjail joypad access) "just allowing the use of /dev/input/js* might be a good workaround for now." 08:16:15 Thaodan, thanks! 08:17:35 #info the Wayland protocol mentioned earlier the might exist, is wip right now: 08:17:46 #link https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/110 08:19:16 We are already on Telegram #sailfishos-fanclub :D 08:19:44 (And on IRC of course /JOIN #sailfishos-fanclub) 08:22:26 Thanks Thaodan for the links, and for flagging up the discussion on the Fan Club :) 08:25:07 2 more mins on General discussion, in case anyone would like to make any final comments, or have anything added to the minutes. I think we already had lots of nice discussion today. 08:26:24 Any of the https://github.com/sailfishos-community devs still around? 08:26:52 I figured out we could put some community infra/docs there such as notes for channels or irc bots. 08:27:15 I've looking into running a faq bot for the porters channels and maybe others as well 08:27:19 Looks like an interesting resource, but personally I wasn't aware of it until now. 08:27:23 such as #sailfishos-fanclub 08:27:47 Awesome-SailfishOS is a worthwhile read 08:28:07 I think Mikaela is still around... 08:28:17 That's worth adding to the minutes I think :) 08:28:17 #info "Awesome-SailfishOS is a worthwhile read" 08:28:24 #link https://github.com/sailfishos-community 08:28:35 #link https://github.com/sailfishos-community/awesome-sailfishos 08:28:45 #link https://github.com/sailfishos-community/equivalentsinsailfish 08:29:20 Okay, we reached time. Are we okay to wrap up Thaodan, poetaster? 08:29:27 y 08:29:50 sure thing, reading. 08:29:56 :) 08:29:57 Thanks for all the input! 08:29:57 #topic Next meeting time and date (5 min) 08:30:04 Proposing Thursday 10th November at 08:00am UTC. Note the UTC time change (so that it stays the same time in countries which are moving off Summer time). 08:30:20 Any concerns, objections, parrots? 08:31:31 Alright, silence is affirmation! 08:31:34 #info Next meeting will be held on Thursday 10th November 2022 at 08:00am UTC: 2022-11-10T0800Z 08:31:49 That's it. Thanks all for the nice discussion as always! 08:31:51 #endmeeting