07:10:56 #startmeeting Sailfish OS, open source, collaboration – 17th September 2020 07:10:56 Meeting started Thu Sep 17 07:10:56 2020 UTC. The chair is abranson. Information about MeetBot at http://wiki.debian.org/MeetBot. 07:10:56 Useful Commands: #action #agreed #help #info #idea #link #topic. 07:12:07 #info Meeting information and agenda can be found here: https://forum.sailfishos.org/t/community-meeting-on-irc-17th-sept-2020/1783 07:12:20 I am the meeting’s chairperson today, and will be doing my best to keep time and order. Please behave, respect the timings and be gentle. 07:12:30 #topic Brief introduction (5 min). Please prefix your name/handle with # info 07:12:30 #info Andrew Branson – Sailor @ Jolla 07:13:02 #info Jussi Maaniitty - Sailor @ Jolla 07:13:06 #info Otto Mäkelä - community 07:13:08 looks like the topic changing isn't really working 07:13:11 but never mind 07:13:14 #info Nico - community 07:13:16 #info Nikita Ukhrenkov (*Kit) - Sailor @ Jolla 07:13:34 #info David Llewellyn-Jones - sailor @ jolla 07:13:45 #info Björn Bidar - Sailor @ Jolla 07:14:14 #info Chris Josten - community 07:16:38 #info Simo Piiroinen - sailor @ jolla 07:17:40 #info Ville Nummela - sailor @ Jolla 07:18:46 #topic Separation or more freedom to the community for parts of the OS that are not relevant to the commercial clients. (15 mins by ApB) 07:18:59 #info Its kind of a weird topic since its not much of a question but more of a discussion around how to make the community work better with Jolla. The idea comes from a comment by @pmelas . Parts of the OS are probably (??) not that important to the commercial side of things. 07:19:05 #info Has jolla thought of separating (or maybe opening) those parts -apps mainly- from the parts that commercial clients care about so that the community can (potentially) improve them??? Or for more freedom in the store since in a commercial environment/work phone a user will be limited to what the company decides. 07:19:52 Sailbot woke up :) 07:20:04 I think this is more of a discussion point. 07:20:17 We're missing Sefriol I think. 07:20:30 We could start this answer by asking what are the most critical problems with contributions? Certainly the closed source trees are one aspect of a problem but what else is there? 07:20:31 reviewed. This collected information would definitely help us in trying to provide solutions for the problems in order of priority. 07:21:21 ^ Is the above a copy-paste with parts missing? "reviewed." ??? 07:22:09 which part? 07:22:41 I'm just wondering about the seemingly disconnected word "reviewed."? 07:23:01 I am wondering that as well. 07:25:37 The question is rather complex. My understanding of it was that there is a request to open source certain parts of the code that are currently closed source. While that is an ongoing task that we constantly discuss and work on. I'd like to understand better what are the more critical topics that are problems when contributing. 07:26:06 There's a process for allowing access to the closed-source part. Perhaps it's a bit unclear for many people what's involved with that? 07:26:06 Maybe the answer was not my finest work, but I'd like to understand this in bigger picture. 07:26:47 IRC sometimes chops off long strings without telling you 07:27:55 I guess there's sometimes a problem that the closed and open parts interact quite carefully, so it's hard to contribute to the open parts without having access to the closed parts? 07:28:13 I'm just thinking out loud. Maybe these aren't real problems? 07:28:21 yeah it's often the case that the backend is open while the UI isn't 07:28:28 Exactly. 07:28:33 I think the most critical thing to contributions is obviously Qt and that the whole UI is married to that. Currently the porting to Qt6 is already causing some issues with supporting sailfish as well for some libraries, since Qt6 removed some stuff, where alternatives were only added in 5.15. It would be nice if one could build a completely open source version of SFOS, that is just missing some enterprise features 07:28:33 instead of having to rely on official RPMs. (But I probably don't know enough about that matter) 07:30:19 Nico[m], you may have noticed that there's already been some community work to move components to the Qt6 api. E.g. dcaliste has been doing this for qmf. 07:31:13 flypig: Yes, I have. But afaik supporting Qt < 5.15 at the same time is complicated. 07:31:27 One of the things which I am slightly saddened by is for example the Messages app. It is a Telepathy frontend, which in my eyes is an hub for connecting a lot of IM accounts in one place. I've tried running and contributing to the Matrix Connection Manager, but the result is a bit dissapointing since the Messages app ignores most of the information the Connection Manager provides about the chats. It for 07:31:27 example ignores most group properties, the UI doesn't allow you to start chats with non-contacts and non-phonenumbers and so on. 07:31:56 Nico[m], I guess qmf doesn't have its own UI, so maybe that avoids this issue. 07:32:50 ahappyhuman, so that's a limitation of the front-end, which is mostly QML, but technically closed source? 07:33:36 ahappyhuman, what would you need access to, in order to help resolve that? 07:33:57 Mostly yes, and maybe as far as I know some of the glue between the front-end and back-end. 07:34:01 flypig: It seems like he found a workaround for the toList() and toSet() churn at least, yes 07:35:35 Nico[m], yes, there's a patch in gerrit related to that. 07:36:42 Think time's about up, if anyone has anything else to add... 07:36:45 ahappyhuman, would you consider going down the "signing up to closed source access" route to help with that? 07:36:59 So, could larger parts of this interface to Messages be made public or could it be expanded? 07:37:25 flypig: I believe the jolla-message is the main closed-source part 07:37:43 flypig: the rest of the components seem to be open source 07:38:32 ahappyhuman, so, if you had access to that, either in an open or closed way, would you consider contributing to fix the issues? I'm just wondering if your sadness is due to access, or because it doesn't just work already, if you see what I mean. 07:40:21 flypig: I would be able to fix certain parts, but I'm also not entirely confident in my UI-design skills. And there seem to be a lot of problems in the state of Telepathy right now and it it's not something I would be able to fix on my own. 07:40:36 i've checked that and it looks more open than some. the qml-plugin is OSS. 07:40:48 My sadness is mostly due it not working and not being able to fix that 07:41:15 i do think that's ripe for a community UI. The jolla one is really quite simple. 07:41:45 isn't most of the work going to be in https://git.sailfishos.org/mer-core/commhistory-daemon.git 07:41:50 If the UI was working, I would probably not care too much about the access 07:42:16 Do you want to contribute to the Jolla messages ui? 07:42:21 qml bits: https://git.sailfishos.org/mer-core/nemo-qml-plugin-messages 07:42:54 Yeah, there's probably a lot that could be done with even a simple open UI on top of all that. 07:43:42 Thaodan: I would like to, I'm mostly hesitating because Telepathy seems to be pretty complex and while I have a general idea how it works, I don't know all the fine details 07:44:17 ahappyhuman, could you list (elsewhere) the changes you think would be needed to get the messaging app to work well with Matrix? It'd at least be interesting to know how far off it is. 07:44:30 that's what piggz did with the camera. but it's still a lot of work 07:44:54 though I don't think the qml plugin for the camera is OSS, so he had less to start with 07:46:45 Ok, I'll try to sum up and we'll move on. I got the missing part of maajussi's answer 07:47:13 * ahappyhuman sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/CTaWwJwZrnXJpuilOHntTkvP/message.txt > 07:48:05 Hmm. The Accounts Settings is quite a topic. 07:48:13 Thanks for that ahappyhuman. 07:48:17 #info We could start this answer by asking what are the most critical problems with contributions? Certainly the closed source trees are one aspect of a problem 07:48:17 #info but what else is there? Do you have problems with missing guidance? Problems with an SDK's? Problems with contributing to git trees, e.g., creating pull 07:48:17 #info requests? Problems with getting things reviewed. This collected information would definitely help us in trying to provide solutions for the problems in 07:48:17 #info order of priority. 07:48:24 was that better? 07:48:56 At least I see the message as a whole now. 07:49:07 Thanks 07:49:07 account settings is a good point though - i don't think very much of that is open? 07:49:30 and that's a pretty important part of adding additional messaging accounts 07:49:52 Yes, it's closed, but a lot of it is qml, so can be viewed. 07:50:09 But it's quite challenging. 07:50:14 yeah. it's always the qml-plugins that end up the most opaque there 07:50:32 especially as there's no header or anything to see what they're made of 07:50:53 Is there some overview document, that shows which repo goes where in the stack? It's a bit hard to find the actual repo you are looking for, if you just have a general idea of what you want to do. But I can, you can always ask in that case? Also, there doesn't seem to be a good way to contribute to the documentation, which is usually the easiest way to start contributing. 07:51:03 Back in the day, connectivity (achieved through Messages) was really one of the nicest parts of Sailfish visible to the end user, unfortunately it's somewhat degraded since 07:51:25 ExTechOp, to be fair, that's as much to do with the messaging landscape changing, I think. 07:51:46 Many proprietary systems now, and more varied approaches. 07:51:53 yeah, though that's partly because the world moved away from intergrated messaging and into standalone apps. which was pretty terrible imho :) 07:52:12 Definitely. 07:52:23 flypig Indeed, but even so, the problem with Matrix integration shows there could be improvement here 07:52:26 And messaging seems to have evolved quite a lot, with server side histoory and what not. Telepathy hasn't been able to keep up 07:52:39 flypig: typed pretty much the same as you but slower sorry :) 07:53:02 :) 07:53:07 ahappyhuman: yes, though there has been work done on telepathy to catch up. 07:53:28 Related to problems with contributions - what ever they are, related to access to documentation, missing SDK features etc. Would it be possible to start creating a "wiki" kind of document in our new forum? 07:53:32 I think a lot of people were waiting for xmpp to catch up, but it got bogged down with too many extensions 07:53:34 ahappyhuman maybe ping Kaffeine on that, he did much work modernizing it. 07:53:57 yeah that's the guy 07:54:12 Move on? 07:55:12 To the next meeting item, that is. 07:55:16 Thaodan: I'll will 07:55:37 I think this set topic was the only one on the list. 07:56:05 I had a follow-up from last week about E2E testing. 07:56:11 hang on a sec 07:56:13 "last week" = "last meeting" 07:56:59 Weekly meetings confirmed 07:57:14 #info The Messaging app was pointed out as a good example of an app that is blocking community contributions due to its lack of support for group messaging, and there being no support for third party account plugins. 07:57:32 Nice summary. 07:57:41 I agree 07:57:51 +1 07:57:53 do we have any way forward on that? 07:58:15 I'll move on 07:58:29 #topic “E2E testing in Sailfish” brief followup (flypig - 5mins) 07:58:39 #info Last meeting @Sefriol asked about end-to-end testing 1 of SFOS components. This will be a very brief follow-up to that question, since I promised to do so. Here’s the discussion from the last meeting 3 for reference. 07:58:51 sorry for the 3 :) 07:59:01 "3 clicks" :) 07:59:02 and the one 07:59:58 the floor's yours, flypig 08:00:26 Thanks :) This is just a follow-up. Sefrial mentioned a paper that linked to a BitBucket repo that was inaccessible. 08:00:55 So one of the authors, Andrey Vasilyev, kindly went out of his way to put it back up in a repo. 08:01:08 Here's the link 08:01:10 #link https://github.com/yarfruct/sailfish-os-test-runner 08:01:38 Apparently this is an extended version of the original which they continued to work on. 08:01:58 They also mentioned this, which I thought might be interesting to some people: 08:02:01 "The main idea behind the tool is to automate the testing in the CI environment. It allows to build the RPM packages, install them onto the VM and then execute testing applications." 08:02:04 "Some of the tool features duplicate the features of the recently introduced sfdk tool that is shipped with the Sailfish SDK. Also there may be issues with the latest release of the SDK, the code is currently used along with 3.0.7 release. " 08:03:07 Andrey was really helpful, so for anyone interested, you could get in touch with him (but I wouldn't expect any support, of course). 08:03:33 I think the point about sfdk is interesting: there's much more scope to automate the build process now. 08:03:38 That's it. Monologue over :) 08:03:57 nice! 08:04:25 I really like the sfdk-tool. Made my life a lot easier :) 08:04:52 I know the SDK team put a lot of work into it; they'll be happy to hear it. 08:06:20 ok, shall i move on to general discussion? anything more to add? 08:06:37 Maybe I've overlooked it, but maybe it would be helpful if there was a quick guide on how to set up the required config before building. That was confusing for me when I first started using it 08:07:21 that sounds like general discussion to me :) 08:07:27 #topic General Discussion (20 min) 08:07:50 ahappyhuman: what do mean with "required config"? sfdk config target=something? 08:08:51 ViGe: yes, you can find it by running the sfdk --help a few times, but a quick start guide would be appreciated. 08:09:37 Has the SDK team considered providing official docker images, that you can use without installing the full SDK? Coderus provides some and they work really well, so it may not make much sense, but with the move away from OBS it would be nice, if there was something you can always depend on in CI and such, so it may make sense to make the official? 08:09:40 ahappyhuman: point taken 08:10:01 Can anyone give an estimate on when Sailfish 3.4 will come out, since it was bumped forward from "early September"? I'm a bit ticked that 3.3.0.16 broke eg. Android external memory device access and Google account calendar synch. 08:11:00 NIco[m]: the sdk already supports docker. 08:11:20 Can't say an exact date for the 3.4 release. Should be available shortly. 08:11:37 ViGe: Like something in the scope of explaining how to build your first hello-world application with sfdk. Otherwise the reference documentation is clear and useful. 08:11:38 ViGe, perhaps it would be worth converting your blog post into a post on the forum? 08:11:55 Thaodan: I know, but it is a bit expensive to install a full SDK inside a CI pipeline, pulling a docker image is simpler. 08:11:57 flypig: that would be a great idea imo 08:14:03 Or like a section on the Sailfish developer website 08:14:23 Nico[m]: some work is planned in that respect, but providing plain docker images is not the approach we would like to take. It doesn't fit the big picture well. 08:14:34 So the question is different. Its about could you run it inside a ci pipeline and distribute it independent of the rest of the SDK? 08:15:09 Thaodan: Yeah, I guess 08:15:33 flypig: maybe we should drop that awful blog comments thing and link every blog post to its own forum thread instead 08:15:51 I like the idea 08:16:11 martyone: Well, a plain docker image is very easy to use in a CI pipeline, but if you have a better idea, I'm all ears! 08:16:12 Yeah, that's be good. 08:16:34 "that'd to good", re blog comments. 08:18:23 I don't know much about Wordpress, but there must be a discord plugin for it. 08:18:39 Shall we move on the agree on the next meeting schedule? 08:18:47 Ah. Discord. Discourse. You know what I mean. 08:18:48 ooh that would be swanky 08:18:55 flypig: you meabn discourse not discord 08:19:11 Thaodan: I do :) 08:19:12 i was just thinking about disabling comments and sticking a link manually at the bottom of every blog post 08:19:27 imagine embedding discord on every blog post. it'd be carnage. 08:19:29 I thought you wanted to add yet another service :D 08:19:50 maajussi Let's. 08:19:52 Nico[m]: I am not ready to tell you more right now 08:20:38 For the next meeting I would ask, if this 7am UTC is ok time? Was the usual time something else? 08:21:09 I think it was an hour later? 08:21:20 Yeah, that's my recollection too. 08:21:33 maajussi the usual was later in the day and is the reason i missed my question. :/ 08:21:33 8am UTC? Would that be better? 08:21:38 appologies for that 08:21:43 ahappyhuman: "you can find it by running the sfdk --help a few times, but a quick start guide would be appreciated." - maybe you missed "sfdk --help-all"? 08:21:50 I'm fine with 7am UTC, although I have to say this is the first meeting I've been too 08:21:51 https://pl.wordpress.org/plugins/wp-discourse/ 08:21:51 should have looked into the time better 08:22:37 Yes, I think I made the mistake with the schedule and it resulted in the slow start today - I understood wrong something. 08:22:39 This time works for me, also later would work. Of course it could coincide with lunch time in Finland. 08:23:26 Yeah, the later meeting did tend to get cut off by Finnish lunch. 08:23:31 martyone: I think I missed that one, woopsie 08:24:18 This needs a poll tool for the schedule discussion ;) 08:24:44 ok, any more general discussion before we move onto that? 08:24:54 The problem of course is, we don't have here the people for whom this time is inappropriate. 08:25:05 @freenode_martyone:matrix.org: ominous :D 08:25:28 #topic Next meeting time and date (5 min) 08:25:52 Nico[m]: :) 08:26:17 Proposing 1st October at 7am UTC 08:26:31 I was about to also nominate 2020-10-01T07Z 08:26:49 +1 08:26:50 maajussi: +1 08:26:59 +1 08:27:29 #info Next meeting will be held on 1st October 2020 at 0700 UTC: 2020-10-01T07Z 08:27:36 gotta be precise :D 08:27:41 I think we have the majority vote! See you then! 08:27:50 =) 08:27:53 ok, thanks all! 08:27:55 #endmeeting