07:00:57 #startmeeting Sailfish OS, open source, collaboration -- 7th July 2022 07:00:57 Meeting started Thu Jul 7 07:00:57 2022 UTC. The chair is flypig. Information about MeetBot at http://wiki.debian.org/MeetBot. 07:00:57 Useful Commands: #action #agreed #help #info #idea #link #topic. 07:01:04 #info Meeting information and agenda can be found here: 07:01:08 #link https://forum.sailfishos.org/t/community-meeting-on-irc-7th-july-2022/12069 07:01:16 I am the meeting's chairperson today, and will be doing my best to keep time and order. Please respect the timings and nose wearing. 07:01:16 #topic Brief introduction (5 min). Please prefix your name/handle with #info 07:01:23 #info Otto Mäkelä, community 07:01:46 #info David Llewellyn-Jones, sailor @ jolla 07:01:49 #info Carlos Orama, community 07:01:53 #info poetaster (Mark), community 07:01:53 #info Anton Thomasson, community 07:01:58 #info Nico, community 07:03:07 #info Raine Mäkeläinen, sailor @ Jolla 07:03:24 #info Damien Caliste, community 07:03:40 A nice community turn out today! 07:04:29 #info fridlmue, community 07:04:53 #info Simonas Leleiva -- privateer 07:06:24 The community have it at 7-3 today! It's great to have everyone here, and good to see we have at least one of our questioners too. 07:06:34 #info Björn Bidar - sailor @ Jolla 07:06:51 Ah, good company :) 07:07:07 Please continue to add your info during the meeting if anyone was missed. 07:07:19 Do we have remote though? 07:07:35 #topic PulseAudio update (5 min -- asked by remote) 07:07:42 #info I'd like to inquire into PulseAudio version bump. 07:07:47 #info Version 15 brings high quality codecs support for BT headphones. 07:07:51 #info Is it planned? 07:08:07 This is our first question. Here's our prepared answer. 07:08:12 #info For general information, the PulseAudio repo can be found here: 07:08:18 #link https://github.com/sailfishos/pulseaudio/ 07:08:24 #info As can be seen there, we're currently on version 14.2, updated with the following commit. 07:08:31 #link https://github.com/sailfishos/pulseaudio/commit/e0a3accd8840ecc728c32c1607c2950be5329eb8 07:08:36 #info Although it's no guarantee of future timescales, historically we've updated to even versions, with a roughly 18 month cadence between updates. 07:08:44 #info Theres no update planned per se now. It's highly likely that this will happen at some point. 07:08:51 #info Unfortunately our Audio expert is on summer holiday, thus we could not get his opinion for the PulseAudio update. 07:08:57 #info We therefore propose that we come back to this topic during the next meeting in two weeks' time. 07:09:26 That's the answer. Any comments/thoughts? Does returning to it sound okay? 07:09:41 Sounds good. 07:09:57 Where there any thoughts of switching to pipewire? :D 07:10:31 When reading the question, I thought about Pipewire as well. 07:10:32 We did discuss switching to pipewire last year but that's much bigger effort 07:10:40 But yeah, returning sounds okay :3 07:11:16 I'm still an alsa+jack guy :) 07:11:16 rainemak, that's useful info I think; I'll add it to the minutes 07:11:31 #info : We did discuss switching to pipewire last year but that's much bigger effort 07:11:32 rainemak, agreed. It is also still a bit crashy here, leaving me without audio after a week of runtime sometimes. But it has some nice audio routing stuff :3 07:12:00 Maybe in 18 months... 07:12:22 remote seems to be concerned about Bluetooth audio codecs. Would Pipewire help with that? 07:12:41 I think creating a PR is never wrong to push things along 07:12:53 That is my impression... as should PA 15 07:12:55 Yes, pipewire has those codecs too and better audio input handling too, afaik 07:13:12 (over bluetooth) 07:13:39 #info Both PulseAudio 15 and Pipewire would help with Bluetooth audio codec support. 07:13:46 pipewire is a whole different beast it would require to have new pipewire modules 07:14:03 I'm afraid PulseAudio on Sailfish has much more than one commit on top of master... 07:14:31 Well, you can use a lot of pulseaudio client stuff on top of pipewire, not sure about modules though 07:14:46 attah, your point is that creating a PR isn't straightforward? 07:14:59 And it is probably long term where desktop linux will move to 07:15:05 Precisely - still worthwhile... just hard 07:15:10 But there is certainly no rush 07:16:09 The info above sounded like it was only the one commit 07:16:31 attah, yes, sorry, that's poor communication from me I'm afraid. 07:16:51 No worries, good to have that cleared up 07:17:02 it's not just one commit but rebasign the current SailfishOS version can be still done. 07:17:42 pipewire would require porting all the routing and policy modules 07:17:59 jusa could say more 07:18:28 (by rebasing I meant rebasign sfos pa) 07:18:32 Not a small undertaking by the sounds of it. 07:18:34 I don't see the demand? 07:19:01 Other than broad scale abandonment of pulse.... 07:19:20 Well, it has lower latency and better sound quality here than pulseaudio 07:19:28 And better sound routing 07:19:41 And better bluetooth audio support 07:19:58 I.e. patapon is unplayable on my desktop with pulse 07:19:58 alsa + jack has both of those, but that doesn't make it appropriate. 07:20:09 what is patapon? 07:20:17 Well, PW has the jack API implemented! 07:20:24 patapon is a rythm game 07:21:21 Good discussion and useful info, but I let this run a bit over already. Anything to add to the minutes before we move on? 07:21:24 Sorry, my audio machines are stripped down for 24 channel (firewire, BOB) audio recording. I'm not a fan of pulse. just not sure the effort is called for now? 07:21:52 Well, maybe not, but it does everything slightly better than pulse 07:22:04 But yeah, we can also discuss that next meeting :3 07:22:12 cool. reading to do. 07:22:48 Nico, having a dedicated Pipewire discussion would be interesting for sure. 07:23:03 Let's move on, otherwise poetaster won't get his question answered :) 07:23:05 Expanding on PW. While I think it may not see use in phones, or at least as commonly; I'd like to point out that because of the better routing, it can open more possibilities, such as midi devices and maybe multiple mics, granted an application could make use of these. 07:23:09 Oh my apologies 07:23:28 Don't apologise; that's good info to have. 07:23:42 We can return to PW in the General Topic at the end too. 07:23:50 #topic RGB channel swapping in Canvas, Label, etc. (10 -- asked by poetaster) 07:23:55 I may mention https://forum.sailfishos.org/t/pipewire-replacing-pulseaudio/6196 for records also. 07:24:04 Oops sorry a bit too late… 07:24:12 #undo 07:24:12 Removing item from minutes: 07:24:23 #info https://forum.sailfishos.org/t/pipewire-replacing-pulseaudio/6196 07:24:28 #topic RGB channel swapping in Canvas, Label, etc. (10 -- asked by poetaster) 07:24:35 Hopefully that will work... 07:24:40 #info drawImage of Canvas renders images channel swapped. 07:24:48 #info To expand see the Stopmotion code which applies the same fixes I applied to "Paint". 07:25:00 #info Beside the RBG<->GBR swapping it also fixes the issue that image meta data (orientation for Portrait photos) is not being used when loading a url directly to Canvas. 07:25:11 #link https://github.com/poetaster/harbour-stopmotion/blob/ee5ab00ab4328151ba0d9d78f7bca192c493a15a/src/ImageProvider.h#L95 07:25:23 #info See also: 07:25:29 #link https://forum.sailfishos.org/t/canvas-requestpaint-renders-image-data-tinted-blue/11349 07:25:34 #link https://forum.sailfishos.org/t/qml-label-dispays-images-with-a-blue-ish-tint/8818 07:25:55 I had trouble summarising your question a bit poetaster; hopefully that makes sense. Here's our prepared answer. 07:26:03 #info Thank you poetaster for raising this, and also to gabrielg, Edz, attah and others in the forum for all your excellent detective work. This made understanding the problem a lot easier. 07:26:11 #info We explored this a bit more internally and think we've got a solution. The qtscenegraph-adaptation package is used to store large images in hardware textures to speed up rendering. 07:26:13 Thank you. I suck. Too many digressions. 07:26:26 #info On hybris devices, when such an image needs to be accessed by the CPU it's done via the HybrisTextureFactory::image() function, in which eglHybrisLockNativeBuffer locks the buffer and provides a pointer to data in the buffer's native format. 07:26:34 #info This can be either hybris's BGRA_8888 or RGBA_8888 and in this case it's hybris's RGBA_8888 format. Then the buffer is copied into a QImage(). 07:26:40 #info But when we copy the data out we only specify Qt formats which are equivalent to hybris's BGRA_8888 order, consequently RGBA_8888 images come out with their colours swapped. The following change should fix it: 07:26:53 #link https://github.com/sailfishos/qtscenegraph-adaptation/pull/2 07:26:58 #info However, we've not had a chance to test this across devices or use cases. 07:27:03 #info Before merging the change, it would be especially helpful if members of the community could test the change with various apps on different devices. 07:27:08 #info Testing it requires building and deploying the fixed branch of the qtscenegraph-adaptation package. Building can be performed using the application SDK. 07:27:19 Wow, that is nice! 07:27:22 Woohoo! 07:27:24 Great! 07:27:24 No worries poetaster, the question made sense, just a little hard to summarise. 07:27:51 Great investigations from both sides. 07:28:36 Yes, indeed, as I understand it the info everyone provided on the forum was a big help. 07:29:05 Would anyone be able to test it? 07:29:16 I'll try. 07:29:27 Super, thank you poetaster. It would be a great help. 07:29:57 If you run into any trouble getting the package built, message me (but I tested it out, and didn't hit anything unusual). 07:30:07 Cool. 07:31:06 Testing compilation at the moment in application SDK, everything run fine. 07:31:23 poetaster, you have two apps with this issue, is that correct? 07:31:47 Great, thanks for checking dcaliste. 07:31:51 3 07:32:15 Paint, Stopmotion and... ? 07:32:20 and tidings, presumably 07:32:38 Ah yes, Tidings. 07:33:01 What is the status of Stopmotion? 07:33:40 It's functional but still a 'raw' prototype. I did party tricks with it last weekend. people said 'ooh!' 07:34:07 I haven't had a chance to look at it, but am excited to see the results, I have to say. 07:34:14 that's the sound of success 07:34:27 Yeah, good to hear you got a good reception :) 07:34:38 a lot of ideas/inspiration from several sources. 07:35:02 That's the best way :) 07:35:08 True! 07:35:38 dcaliste, you've finished all the testing I presume :) 07:35:48 Haha :) 07:36:08 We're hitting time on this. Anything anyone would like added to the minutes before we move on? 07:36:33 just thanks! 07:36:52 indeed, nicely done 07:37:00 #info poetaster expressed his appreciation :) 07:37:15 I'm trying it on device. As far as I remember the issue was visible in tidings. But I've a slow data mobile connection at the moment. I need to get some example of RSS flux were it failed. 07:37:16 That's a rare achievement! 07:37:21 I'll pass the thanks on too. 07:37:57 dcaliste, we'll move to the next question, but let us know if you find out. 07:38:03 #topic Community Bug Coordination Team summary (5 min -- asked by pherjung) 07:38:07 #info Untracked bug reports 07:38:12 #link https://forum.sailfishos.org/t/community-meeting-on-irc-7th-july-2022/12069/2 07:38:28 This is our usual Community But Coordination Team summary. 07:38:33 #info The Community Bug Coordination Team have done a superb job once again this fortnight. 07:38:40 #info As a result of their work, we now have: 07:38:44 #info - 10 new high quality bug reports recorded internally and tagged as "tracked". 07:38:49 #info - 5 topics moved into different categories as they weren't really bugs. 07:38:52 #info - 4 bug reports marked as "pending" more info. 07:38:55 #info - 2 bug reports tagged as "fixed". 07:38:58 #info - 2 bug reports related to RBG-BBR swapping which hopefully we'll be able to mark as "fixed" soon. 07:39:12 Are there any CBCT members here? 07:39:43 Well, in theory I am in the same chat group, but not really active rn ;p 07:40:11 So no idea what there would be to add, apart from the usual thank you :3 07:40:58 Yes, thank you from me too. I'm continually impressed, especially with how the team tests the bugs and writes up notes for how to reproduce them. 07:41:31 I am too :D 07:42:42 Kudos from me too. I find it increases signal to noise a great deal. 07:43:16 Okay, we shall move on then. 07:43:32 #topic Open Pull Requests (5 min -- asked by rainemak) 07:43:39 #info During the last meeting we agreed to have a section on open Pull Requests, to help prioritise our work in reviewing them. What did we get? 07:43:46 #info We didn't get any submitted before the meeting, but everyone is welcome to raise interesting PRs for us to consider now. 07:44:12 So, this is a chance to raise any lingering PRs that aren't getting the attention they deserve. 07:44:31 Heh, I didn't open any, because I was busy with work and other stuff, I'll get there some day! 07:45:14 There is, of course, no obligation :) But for future reference, you can propose others' PRs as well. 07:45:37 Most PRs open right now are either wip or not so important to mention I think. 07:46:03 Okay, that's useful to know. Thanks Thaodan. 07:46:03 Yeah, I haven't seen any complaints about lack of review the last 2 weeks .-. 07:46:14 This is good. 07:46:26 Oh, maybe I can dig one up 07:46:47 Some are great as usual but are discussed elsewhere e.g. dcaliste 07:47:04 dcaliste has a few browser PRs: https://github.com/sailfishos/sailfish-browser/pull/957 07:47:38 https://github.com/sailfishos/sailfish-browser/pull/959 07:47:42 Indeed, pvuorela thanksfully took some of his time before summer vacations to give feedback on calendar stack ones. Thanks to him. 07:48:02 I wouldn't mind more testers on: https://github.com/sailfishos/droidmedia/pull/92 07:48:53 #info dcaliste has a couple of browser PRs that have been opened for some time. 07:48:56 #link https://github.com/sailfishos/sailfish-browser/pull/957 07:49:02 #link https://github.com/sailfishos/sailfish-browser/pull/959 07:49:07 I like dcaliste's PRs. The press and hold scheme sounds like good idea. 07:49:27 #info Thaodan requested testing on one of his PRs. 07:49:29 #link https://github.com/sailfishos/droidmedia/pull/92 07:49:32 I guess those might need feedback from whoever is responsible for design stuff 07:49:33 I almost tried to do this out of instinct. 07:49:47 Thaodan, same 07:49:57 Nico, about 959, I've used it for a while and while I like the consistency, I find it a bit cumbersome to have to long-press and them tap to close one tab, which is the usual action I'm doing. 07:50:02 rainemak, do you have any comment on dcaliste's browser PRs? 07:50:24 The target to close is easier to get though. But I'm not totally convinced of its advantages. 07:50:40 It requires more discussion and feedback from people who tried it I think. 07:50:50 dcaliste, sounds reasonable, maybe moing the close button to below the tab cards is bettter? 07:51:02 And not doing the long press 07:51:33 (I mostly close my currently open tab using the menu action instead of the tab view) 07:51:56 :) 07:52:24 Nico, I should change my habit for the menu then ;) 07:52:28 I think long-press is a good pattern here, that would allow slightly bigger buttons too... the current one is hard to hit 07:52:48 I think I'll just try out the long press patch for a bit 07:52:55 Just for records, one way to see all opened PR on SailfishOS is https://github.com/pulls?q=is%3Aopen+is%3Apr+org%3Asailfishos 07:53:28 Thaodan: how do we know if the pr is working? 07:53:33 "No results matched your search."? 07:54:04 flypig, maybe you need to get logged in ? Strange. 07:54:27 thilo[m]: nothing breaks and for affected devices the camera app doesn't freeze 07:54:58 Just tried that URL when not logged in, it gave the 404 page but allowed me to log in and then showed the PRs 07:55:12 Sounds doable ;) 07:55:57 I think it's an encoding thing. The following works for me: https://github.com/pulls?q=is:open+is:Apr+org:sailfishos+ 07:56:14 Hi, 07:56:22 I'm bit a late :) 07:56:33 Nice to have you here pherjung. 07:56:44 ditto! 07:56:51 :) 07:57:00 Welcome~ 07:57:01 #info Just for records, one way to see all opened PR on SailfishOS is 07:57:02 #link https://github.com/pulls?q=is%3Aopen+is%3Apr+org%3Asailfishos 07:57:06 We already did the bug roundup, did you have anything to add? 07:57:37 Let me move us on to General discussion; thanks for the info about the PRs. 07:57:38 #topic General discussion (20 min) 07:57:44 I have an question. Could it be possible to do the same as PR review, but with bug reports? There are some I have no idea how to reproduce them 07:58:15 Idea is to pick one and Jolla give some hint where we, the community, should look 07:59:00 It sounds like a nice idea to me. Did you have a bug in mind for an example? 07:59:46 Not now, but we can pick up one for the next meeting just as test 07:59:56 There's the very-long-standing issue of failing network switching with Android applications, here's one leaf of the tree: https://forum.sailfishos.org/t/android-connectivity-issues-analysis/12091 07:59:56 This hits me pretty much on a daily basis (home wifi / phone network while going to work / work wifi), is anyone working seriously on this? 08:00:51 I could try to give a look & comment to dcaliste's browser PRs during this week. 08:01:02 Are there any news on the ownership and "super secret engineering effort" front? 08:01:12 this one is already tracked, no? 08:01:43 Oh, what SSEE? I never heard of that, sounds interesting! 08:01:52 pherjung, I think that'd be a great idea. Could you flag it in advance in your bug roundup? 08:02:14 I'll do :) 08:02:30 Great, thank you. If it works, we can then make it a regular thing. 08:02:44 cool 08:03:17 rainemak, I will add your comment to the minutes, I think that's nice to hear :) 08:03:31 #info I could try to give a look & comment to dcaliste's browser PRs during this week. 08:04:04 #info all Sailfish OS PRs: click the "Pull requests" link at the header, enter "is:open is:pr org:sailfishos" to the search field, press enter 08:04:47 ExTechOp, I know there has been a lot of work on it (and fixes) in the past, and my understanding is that work continues, but I'd have to check to make sure. 08:05:29 #info ExTechOp raised the issue of long-standing network issues with Android. 08:05:33 #link https://forum.sailfishos.org/t/android-connectivity-issues-analysis/12091 08:05:55 As I said, still there on a daily basis. Occasionally Utilities / Restart network helps, sometimes the only thing to do is to reboot the phone. 08:05:58 Thanks rainemak, useful info. 08:06:07 flypig: here an example: https://forum.sailfishos.org/t/weather-display-on-events-screen-constantly-loses-network-connection/11962/ 08:06:43 could be great for instance to enable some logs 08:07:03 so we can find how to easily reproduce it 08:07:14 I know someone whose phone just doesn't connect to the mobile network at all after boot and needs toggling of the flightmode to connect... 08:07:43 I don't know if that was a bad flash or not yet though, but it worked at some point 08:08:03 So the question, pherjung, is how to help find a way to reproduce the issue reliably? 08:08:20 yes 08:08:28 ExTechOp: also try restarting App Support, as last resort before reboot 08:08:55 Sorry got side tracked. Thank you rainemak. 08:10:24 No company news to share then? 08:11:05 pherjung, I would say that some of these don't have an easy answer unfortunately, but if you could add it for the next meeting, that would be great. 08:11:33 attah, did you buy Jolla yet? :D 08:11:51 i wish 08:12:04 No news then... 08:12:09 attah, thanks for asking, but I think there's no additional news to share today. 08:12:25 okay, thanks for info 08:12:33 Do please ask again though :) 08:12:38 Does it smell like there might be news in the near future? 08:12:40 count on it 08:12:47 and about engineering work? 08:13:40 Anything particular related to the engineering work? 08:13:48 having installed scenegraph (hybristexture config), reboot? restart lipstick? 08:14:20 Since we don't know what it is... it is hard to be specific.... but how is it going? 08:14:27 well, community news has been reduce to worg on engineering stuff. So we are curious which part 08:14:28 poetaster, in my testing, just restarting the app was enough, but restart lipstick if you want it across the board. 08:14:56 I'm sure they are porting to qt6 atm, right? 08:14:59 pherjung, ah, yes, it's in relation to that. Thanks for clarifying. 08:15:18 sigh. 08:15:32 Oh, that's a bad sign. 08:15:34 Nico: that is what i choose to believe too :P 08:15:58 CONFIG += plugin verbose dither eglgralloctexture hybristexture ... no changes in app behaviour 08:16:16 attah: With a secret we can dream of anything we want! 08:16:55 poetaster, can you share a minimal working example of the problem? We used the one provided by Edz for testing, but maybe what you're seeing is something different. 08:18:38 And you changed to the pr commit, not just downloaded it? (definitely not speaking from experience) 08:19:07 pherjung, attah, Nico, concerning engineering work, there's nothing specific to report, but if you want the situation expanded upon, I think best to ask a question for us to prepare an answer for in a future meeting. 08:19:13 yup. But += could be wrong. 08:20:27 Well, I only heard about it just now! 08:20:33 poetaster, you shouldn't need to make any changes to *your* project to have the fix, you should just need to install the updated qtscenegraph-adaptation package. 08:22:03 Yes, sorry Nico, that's due to me following the conversation backwards poorly :) 08:22:10 you need to specify in the scenegraph pro which (hybristexture for instance) to build 08:22:39 poetaster, okay, then I'm misunderstanding. Apologies. 08:22:59 flypig, no worries. Just spill the beans to apologize! (That's how this works, right?) ;p 08:23:00 i just did "sfdk build" 08:23:26 Nico: :D 08:23:31 hey, a small comment about pulseaudio updating, it needs first moving to meson since autotools support was dropped in pulseaudio 15, this means changing the custom patches to work with meson, I have done that but no pull requests yet, might need some more testing first 08:24:00 Project MESSAGE: Enable the above features by adding them to the qmake config, for instance: 08:24:05 #info a small comment about pulseaudio updating, it needs first moving to meson since autotools support was dropped in pulseaudio 15, this means changing the custom patches to work with meson, I have done that but no pull requests yet, might need some more testing first 08:24:07 Mhm, meson, yummy. So much nicer than autotools 08:24:11 Thanks mal, nice info. 08:24:40 We are hitting time for today. Anything else to add to the minutes before we wrap up? 08:27:01 Alright, final topic then. 08:27:06 #info Next meeting will be held on Thursday 21st July 2022 at 07:00am UTC: 2022-07-21T0700Z 08:27:18 Oops, that was premature. 08:27:20 #undo 08:27:20 Removing item from minutes: 08:27:24 #topic Next meeting time and date (5 min) 08:27:28 Proposing Thursday 21st July at 07:00am UTC 08:27:59 20 seconds to disagree, otherwise it is accepted? :D 08:28:12 WorksForMe 08:28:14 ... 10 ... 08:28:16 Ditto 08:28:27 Excellent. Let's go with that then. 08:28:28 3 08:28:30 2 08:28:31 1 08:28:32 color fix works! 08:28:34 #info Next meeting will be held on Thursday 21st July 2022 at 07:00am UTC: 2022-07-21T0700Z 08:28:45 poetaster, fantastic! And just in time too :D 08:28:51 Nice 08:28:56 now for orientation :) 08:28:58 #info color fix works! 08:29:00 Thanks for hosting 08:29:06 ditto! 08:29:09 Thanks everyone! 08:29:13 And thanks everyone for joining and for the great discussion. 08:29:17 #endmeeting