08:00:32 #startmeeting Sailfish OS, open source, collaboration – September 6th 2017 08:00:32 Meeting started Wed Sep 6 08:00:32 2017 UTC. The chair is Jaymzz. Information about MeetBot at http://wiki.merproject.org/wiki/Meetings. 08:00:32 Useful Commands: #action #agreed #help #info #idea #link #topic. 08:00:43 #info Meeting information and agenda can be found here: https://lists.sailfishos.org/pipermail/devel/2017-September/007989.html 08:00:57 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. 08:01:08 #topic Brief introduction (5 min). Please prefix your name/handle with # info 08:01:18 #info James Noori - Sailor at Jolla 08:01:33 #info Steph Gosling, community and porter 08:01:34 #info Fabio Isgrò - Community 08:01:37 #info Mister_Magister - Mad porter and community member 08:01:38 #info Leszek Lesner, Community Member and Developer 08:01:39 #info Kimmo Lindholm - Community 08:01:40 #info Leif-Jöran Olsson, community member 08:01:44 #info Simonas Leleiva, sailor at Jolla 08:01:48 #info Aleksi Suomalainen, community 08:01:50 #info Adam Pigg - chief xiaomi-mido porter 08:01:51 #info sugash community 08:02:21 #info nh1402, the reason for this meeting to be as long as it's going to be 08:02:40 #info Pami Ketolainen, sailor @ Jolla 08:02:45 #info Thomas, mer community member 08:02:47 nh1402: lol but even without you we've got many topics today 08:03:11 #info Lewis Rockliffe, mother of dragons, breaker of chains, and true king in the north 08:03:22 #info Martin Kolman, community & modRana 08:05:49 r0kk3rz: lolz 08:06:24 r0kk3rz: when you give birth to the next batch of dragons i want a spotted one 08:06:35 someone can spam the meeting out there ? 08:06:47 dr_gogeta86: people did, i saw :) 08:07:05 First topic,coming up 08:07:10 #topic Newer version of GCC (Asked by nh1402 – 10 min) 08:07:29 info New information has come to light that you can set a flag to use an older abi version which should make it easier to get a newer version of GCC working. If a more detailed explanation of how the community could build and test it if Jolla don't have the people power to do it, then that would be appreciated. 08:07:51 my beloved topic 08:07:52 #info eekkelund 08:08:27 Jaymzz: start ? 08:08:47 this is basically a follow up to last time, it's been mentioned that the community would have to do this if it was to move forward in Sailfish, but I was wondering what specifically would be needed to do, and how best to do it. 08:08:50 Sugash: What start? :) 08:08:53 sounds like a solution to me, i look forward to seeing gcc 7 in 2.1.3 ;) 08:09:45 yeah that sounds reasonable especially as newer software depends on newer gcc versions being available for compiling (or clang) 08:09:52 #info Damien Caliste, sorry being late 08:10:16 #info Andrew Branson, me too 08:10:17 is it impossible to have glibc4 and glibc 6 in the same time? 08:10:59 an approach like the old compat packages on rhel 08:11:00 we get latest GCC in Fedora all the time and we survive, so its doable 08:11:04 it's not that much rocket surgery to get things going. I'd propose to set up a test project and start by rebuilding core parts of mer. I'm sure we'll find some people in #mer to help (including me, time permitting) 08:11:28 A guy in italy is trying to port telegram desktop on sfos 08:11:34 just not sure about keeping binary compatibility 08:11:37 M4rtinK: Fedora rebuilded its whole repo though after the gcc5 switch 08:11:39 the show stopper is the gcc 6 problem 08:11:39 or during 08:11:57 -fabi flag should help ease the pain 08:12:18 I have made quick-and-dirty gcc 6.4.0 compile and packaging for sfos with switches that should make it compatible with gcc 4.8. if anyone interested, its at obs 08:12:41 rinigus, is public ? 08:12:46 #info Jonatan Walck - shamefully late community member 08:12:46 there is some gcc5 in openrepos dunno how it's working 08:12:47 -> url 08:12:59 yes, sure. https://build.merproject.org/package/show/home:rinigus:toolbox/opt-gcc 08:13:09 leszek: well, we generally rebuild everything once per release anyway 08:13:10 #link https://build.merproject.org/package/show/home:rinigus:toolbox/opt-gcc 08:13:22 M4rtinK: ah ok then this is no problem 08:13:34 abranson_: you were working on the gcc update? can you say anything about this? 08:13:35 I haven't tested it though. be warned 08:13:38 pain or not to recompile everything its a one time thing. and it will clear the repos from apps that are not maintained anymore 08:13:56 so not 100% applicable to Sailfish with all the proprietary stuff & apps mixed in 08:14:10 pketo: I didn't get very far with it unfortunately. I was still tracking down which of the patches were still required when I moved onto something else 08:14:25 3 miniutes left on this topic people, today we have a pretty tight schedule 08:14:39 i guess if you rebuild everything, with the abi downgraded, store apps will still work? 08:14:43 Jaymzz: meeting has time limit? 08:14:45 no 08:14:47 but as I said last time, the biggest danger is all store apps needing a recompile before they'll work again 08:14:54 rinigus: i'll give it a go tonight, i dont care if it breaks my tablet :) 08:15:03 #info Cedric Heintz, community member (late again...) 08:15:07 hello everyone 08:15:14 abranson_, why not put it in a container 08:15:14 abranson_: a proper packaging would be of great help! gcc is keeping us behind 08:15:27 abranson_: really? If its compiled with the compatible flag the binaries should be compatible and don't need recompile 08:15:28 * M4rtinK notes that apps using Flatpack rntimes would help there 08:15:29 I suppose we can move forward with the work rinigus has started 08:15:36 otherwise this flag does not make much sense 08:15:36 Mister_Magister: Well, today's meeting is going to be over 100 minutes! too much over that is not going to be feasible for us 08:15:57 just add a new runtime with new GCC - done! 08:15:57 testing it, patching etc. 08:16:08 leszek: indeed, although will the RPM deps be ok? 08:16:19 r0kk3rz: they should be ok 08:16:22 mostly 08:16:26 r0kk3rz: it should be possible to use it in parallel with the current gcc => shouldn't break anything, as far as I would see 08:16:32 otherwise ship meta packages :P 08:17:00 shall we move on to my next topic then :p 08:17:11 nh1402: we can do that :) 08:17:33 is there a definitive outcome and way forward? 08:17:48 +1 08:17:56 ie, if X works, will we see it in sailfishos Y ? 08:18:10 nh1402: Moving on in a second, seems like there still in some discussion left 08:18:31 piggz_: I think those that want to help with GCC update can work with rinigus repo, test, patch etc. 08:18:42 abranson_: [gcc] I had some errors with arm gcc compile when having lto enabled. maybe you could look into it as well? 08:18:51 and discuss in #mer channel for help 08:19:04 yes, and if that is successful, will sailors upgrade? 08:19:39 #info those who would want to help with GCC can do so by working with rinigus repository. More to be discussed in #mer channel 08:20:08 that sounds about right 08:20:11 note that its not proper gcc packaging. for proper upgrade, we would need to work on gcc as packages currently 08:20:24 ... with sfos 08:20:47 #info (from rinigus) note that its not proper gcc packaging. for proper upgrade, we would need to work on gcc as packages currently with sfos 08:20:57 3 min over time, we shall move on 08:21:06 #topic Picture in Picture (PiP) & Splitscreen Multitasking (asked by nh1402 – 15 min) 08:21:17 #info Splitscreen multitasking has been brought up before (by myself), Apparently the main thing to modify to support it would be Lipstick, PiP support would be nice, I'm not really sure which would require less work. I was wondering what specifically would need modifying (What in Lipstick), and how the community can help bring it if Jolla don't have enough resources to. 08:22:03 let's start with which would require less work, and if the work needed to implement it is *just* modifying Lipstick 08:22:34 Picture in Picture, or splitscreen multitasking. I'm guessing it's the former. 08:23:02 also, if anyone else is interesting in PiP 08:23:20 split screen would be awesome. and if not just 2 window but more that will be even more awesome 08:23:27 I would think that technically they are not much different, but I don't really know anything about this area :) 08:23:29 in general lipstick as wayland compositor is the only thing that needs change. Of course the applications itself need to handle resizing aswell so they don't appear broken with unreachable controls 08:23:30 not for phones ... but for the tablets out there 08:23:36 was an idiegogo perk 08:23:37 I am interested 08:23:41 yeah i'm talking about tablets 08:23:42 nh1402: making it work is one thing. the ui/ux to bring it up and make it easy for the user to use needs quite a bit of thought 08:23:42 *indie 08:24:07 tv's :) 08:24:16 locusf, also 08:24:16 ApBBB: yeah you are right 08:24:18 too lol :D 08:24:38 ApBBB: of course 08:24:50 on phones 2 windows can be usable still 08:25:23 4k 6" screens have plenty of realestate 08:25:30 If I understand well, PiP may solve the security issue of having app setting run from the system setting app. 08:25:30 Mister_Magister: more than 2 on any screen up to 10 inches sounds too much for me. 08:25:51 A swipe guesture from the bottom left or right corner is what I had in mind. It will then split the screen and allow the already open app to run in the left half of the screen. The right half shows a compact app drawer of the installed apps and you can click on the other app that you want to run there. Or choose even an already opened app 08:26:00 ApBBB: 4 windows ion 7" would do for me 08:26:12 but keyboard is the problem , thats wht Samsung had failure in grand series long before 08:26:12 ApBBB: one thing about splitscreen is that having two apps side by side, would stop "peeking" working on one of the apps. 08:26:55 the way Android handles it iirc, is that if the top app needs keyboard, it just basically draws the keyboard over the bottom app. 08:27:04 If keyboard size is modified it will bring huge success 08:27:07 leszek: or long tap on the edge to resize window 08:27:20 well the keyboard is an overlay in lipstick 08:27:21 there was a good poc on n9 08:27:24 from tmo 08:27:30 nh1402: thats why i said it needs a lot of thought. but from the top of my head if the half screen peek shows the whole screen underneeth having one app lose peek isn't much o f a problem 08:27:31 so it should do the same thing 08:27:47 Mister_Magister: also nice idea but I fear it might conflict with some applications that place their buttons at the edge (Webcat for example) 08:28:48 leszek: what i had in mind is a 3 finger gesture to split and bring up the app tray 08:28:50 locusf: i was thinking glacier could be a testing ground for this, how usable is that at the moment? 08:29:03 leszek: then bring app drawer from bottom and hold on app icon and move it to second half of screen 08:29:04 I wonder if jpetrell is free to reply something :) 08:29:07 that would be awesome 08:29:16 r0kk3rz: will relay to eekkelund :) 08:29:18 ApBBB: also a very good idea. Though 3 finger swipe is already used to change the volume 08:29:37 well I propose we organise a separate meeting just for multitasking discussion, put all the cards on the table if we do this, what would happen etc. Build up a plan, and then go from there. 08:29:51 We can make it fo r tablets .I think For now tablets are the good choice 08:30:05 lots of space 08:30:06 r0kk3rz: essentially it requires relayering of wayland surfacing 08:30:07 *surfaces 08:30:08 we need I think prototypes for the swipe gestures and such to try out what works best for the users 08:30:10 which is done in lipstick 08:30:10 and I have no idea how that could be done 08:30:38 locusf: yeah but having open chrome probably helps too, since its the same lipstick 08:31:04 krnlyng: might have a better idea of wayland surfaces and how they work 08:31:37 r0kk3rz: possibly, problem is that sailfish lipstick has different edge touch sensing from glacier 08:31:45 ah ok 08:32:05 technically lipstick draws already multiple surfaces in the home view, right? for android apps it shows their live application only minimized to the cover or am I mistaken? 08:32:17 also, is everyone happy with having a separate meeting for multitasking discussion, as in this meeting it could drag on and on. 08:32:28 i feel this meeting to be greatest since i joined meetings 08:32:29 +1 08:32:36 :) 08:32:38 nh1402: yup 08:32:41 Jaymzz: not sure what I should comment on 08:33:20 nh1402: +1 for a dedicated meeting 08:33:22 edge swipe areas are invisible and not discoverable, even the current Sailfish OS signature gestures require quite much teaching in Tutorial for people to learn 08:33:25 jpetrell: The topic is about split screen and PiP, and what the community could do to help cause we do not have enough resources at the moment to implement those features 08:33:27 +1 from me 08:33:39 locusf: I don't see any issue with wayland personnaly, but of course the compositor will have to be largely adapted. But nothing too technical, more design issues. 08:33:46 ah 08:33:53 leszek: not swipe 3 finger closing (like a 2 finger zoom) 08:34:04 (zoom out) 08:34:13 ApBBB: ah great 08:34:26 3 finger gesture is no good it requires to just use whole hand 08:34:38 unfortunately our homescreen and Silica components are not open source yet so developing split screen is not really possible by community 08:34:48 double tap in to a 3 finger swipe 08:34:52 Mister_Magister: yes 08:35:00 taaem: those are different surfaces, WindowPixmapItem 08:35:02 one point more to make it open source asap 08:35:07 #info (from jpetrell) unfortunately our homescreen and Silica components are not open source yet so developing split screen is not really possible by community 08:35:23 locusf: ah okay 08:35:26 except of course where it is if one goes with Nemo homescreen 08:35:54 for apps to change between full-screen and split modes would require more dynamic layouting (currently geometries are decided on app startup) 08:36:09 keyword being 'yet' so as leszek said when it could be open sourced 08:36:14 #info (more from jpetrell) for apps to change between full-screen and split modes would require more dynamic layouting (currently geometries are decided on app startup) 08:36:39 that is perfectly doable, but Silica needs to be reworked. then we need system UI for managing split screen and migrate all the system dialogs to support different state combinations 08:37:00 about layouting, I may be wrong, but geometry is already changed and app relayout accordingly when rotating portrait to landscape. 08:37:11 gotta love that elbowgrease 08:37:29 #info Silica needs to be reworked and we need system UI managing split sceen and we need to migrate all the system dialogs to support different state combinations 08:37:39 dcaliste: parent item dimensions yes, but not the geometries exposed by Theme object and icon sizes 08:37:45 the key components maybe closed, but we can discuss how to move forward, and then once we make a spec of how to handle it, then it comes to implementation eg in Nemo, wait for opensourcing etc. 08:37:52 nh1402: It's going over time again, but I'll give it a few more minutes 08:38:24 nh1402: meeting is going overtime because of you :D 08:38:40 It's all nh1402's fault XD 08:38:46 yes :D 08:38:58 just my 0.02EUR, but if I were interested in this going anywhere, I'd work on the open source replacements in nemo 08:39:00 Jaymzz: I still have *another* topic iirc 08:39:15 yup nh1402 have 3 topics 08:39:17 split screen prototype that uses wayland and lipstick would be valuable first start 08:39:25 nh1402: You do :) 08:39:26 jpetrell, mmh ok, I thought that these Theme values where only based on pixel density which is not changed by split screen, but you must be right, I'm not aware of all Theme values, I believe you ;) 08:39:37 tbr: I think we should discuss implementation details first, and then work on nemo 08:39:44 all the closed source and "wait for it to be open sourced" has been a major stumbling stone 08:40:03 also a reason why I stopped caring about the sailfish side of things 08:40:08 implementation details might take a while by itself. 08:40:16 thats true 08:40:27 nh1402: yes, but so far discussion is being squashed by "wait for xyz" 08:40:39 i'm again lazy at work because of meeting. gotta fix my multitasking lol 08:41:05 my patience for waiting on opensource is also running out. The end of the year is the deadline for me 08:41:13 next topic should be quick anyway :) 08:41:30 so plan is organise meeting to discuss implementation details, then work on nemo homescreen to implement it. 08:41:45 meeting(s) 08:41:49 wohoo \o/ 08:41:49 yup +1 08:41:54 leszek: why care. apps are the problem with SFOS. 08:41:59 this channel is generally available for meetings 08:42:18 also it's called *mer* meeting for a reason ;-) 08:42:24 ApBBB: you cannot improve the overall experience and convince people of SFOS if they always promise something but don't deliver it 08:42:26 #info Meetings will be organised to discuss implementation details. Work will then bein on Nemo homescreen. 08:42:33 more info about opensoircing progress would be welcome BTW 08:42:38 I'm fine with that, I'm not sure when or how to organise the meetings though 08:42:41 nh1402: Can we move on? :) 08:42:55 but this topic has been running over and should move on 08:42:58 nh1402: We can discuss it later, it's way over time now 08:43:08 quite a while since anything has really been said about that and no real user visible progress 08:43:15 nh1402, choose a first time and announce it on devel mailing list + irc here. 08:43:45 people who cannot attent can read the minutes later. 08:43:52 and participate async on the mailing list. 08:43:54 Alright moving on to the next topic. 08:44:00 #topic: Custom Gestures & Fastening UI. (Asked by Sugash – 10 min) 08:44:15 #info Currently we have limited user customizable guestures.which causes a problem either between left and right hand users . User definable guestures can add more customizability and it can showcase Sailfish Os as the best Customizable Os. 08:44:15 Fastening UI : 08:44:15 1.The animation speed is slow and it's seem to appeal as a Slow Ui or OS . ( The Warehouse patch is liitle laggy and it doesn't appear to be native ) . Speed can also be made User Customizable ! which adds more UI customizability ! 08:44:15 2.Remorse timer is the best of Sailfish . Few think why to wait for 5 sec for a single action ( probably Not true as we can continue with other activities ) . Many tend to wait for the completion if they work on the same page . So adding an immediate action button on the remorse action bar will make it easy & faster appealing . 08:44:18 i will merge 2 topics , its all about UI 08:44:19 leszek: M4rtinK: tbr: agree, we should add the opensourcing of sfos as topic for next meeting 08:44:40 jpetrell: I think you'd be needed on this topic as well :) 08:45:05 we need to add timely updates with new features 08:45:06 c_la: don't include me. I focus on mer and only that. I've long given up on sailfish. 08:45:34 Oh i want to tell something there. on my moto x force i have big.LITTLE architecture and when i disable bigger cores i have 4x1.5GHz but ui gots very very slow every animation is slow so i had to disable little ones. Is sfos rendering everything on cpu? on moto g2 4x1.2GHz animatinos are normal 08:45:51 c_la: good point! will you do that ? 08:46:25 we should leave this to general discussion, Sugash go on 08:46:30 Sugash: I think the sluggishness depends on the GPU. Take the Jolla 1 for example it handles animations much more smooth than for example the Jolla C which has a inferior GPU 08:46:43 since the launch sailfish 2.0 , UI has not xhanged much , Android is adding more features like multi tasking and improving UI 08:47:05 i wouldn't describe J1 as smooth. on occasion you can see hickups 08:47:22 ApBBB: its smoother than the Jolla C though 08:47:26 M4rtinK: no problem, but please do it if I don't do it in the next few days 08:47:30 Users are attracted to Only whats in UI , not every one is geek . Its the only way we can bring more users 08:47:53 c_la: ok :) 08:48:17 Sugash: there was an improvement with 2.1 which brings in a filter which smooths out stuff overall. Though yes it is just a trick as it makes the responsiveness slower in general 08:48:40 ya 08:48:53 Sugash: design issues are rarely getting "fixed" in sfos. there have been many complains about various stuff yet none has been handles by jolla. 08:48:56 though I like smoother animations more then responsiveness I have to say as long as it does not stand in the way 08:49:30 When it comes to design decisions. They are decisions you will always find people disaggreeing 08:50:02 sailfish is lighter and fast multi tasking 08:50:06 I agree with having an option to control animation speed, like in Android 08:50:20 leszek: you can always argue in a constructive way. its not like "i would like this white to be whiter" 08:50:21 but not everything is faster we need to make it faster 08:50:41 and point 2 I brought up before 08:50:42 jpetrell: Do you have any comments here? You're the UI guy :P 08:51:29 Jaymzz: we need to make it faster and Attractive 08:51:33 leszek: i wrote this > https://together.jolla.com/question/145880/gripes-with-2x-improving-the-uiux/ a while ago and havent received an answer or arguments against them 08:52:19 do everyone agree with custom guestures? 08:52:21 +1 for a 'do it naow' button on the remorse timer though 08:52:27 yeah 08:52:29 +1 08:52:40 +1 (both) 08:52:54 +everything 08:52:59 hmm... needs a mockup first 08:53:19 yes people will complain that you have to reach the top of the screen, but like they dont have to press it 08:53:21 another bad thing aboit remorse timers - they migh be canceled when you leave page they are on 08:53:30 2 minutes left Sugash, you need to wrap it up :) 08:53:38 seen that a couple of times, pretty bad UX 08:53:42 gow about fastening 08:53:49 ? 08:54:00 M4rtinK: sometimes it does the action, like with flight mode. 08:54:03 can't really happen wit yes/no dialogues 08:54:10 M4rtinK: where does that happen? 08:54:16 app caches are must for fast app openings , 08:54:21 Sugash: not sure how to do it more faster. 08:54:46 Sugash: there are caches for apps already just like for example the sailfish-browser 08:54:48 copping up with qt 5.8 08:54:50 IIRC system stuff is generally fine, but apps are buggy with this 08:54:51 5 9 08:54:52 Sugash: fyi fastening is affixing something in place :) 08:55:10 Sugash: thats a totally other topic to upgrade Qt and I think we discussed that in the past already 08:55:23 eq page/delegates probably getting garbage collected 08:55:30 in the animation i can see the delay and switxhings 08:56:05 Sugash: most probably a hardware limitation 08:56:11 Sugash: Time's up man, you asked for 10 minutes. Ready to move on soon? 08:56:11 the hardware cannot handle this or the drivers are crappy 08:56:14 may be 08:56:34 reproducer: open mail app, try to delete message, close app 08:56:34 I am not sure if SailfishOS uses EGL for all the UI stuff though 08:56:39 ok abyone giving +1 08:56:42 message will stay there 08:56:48 for te requeats 08:57:14 M4rtinK: yeah thats an overall problem I guess. As closing the app would kill the timer 08:57:15 Jaymzz: ok get votes and move on 08:57:18 there are still 3 more topics iirc 08:57:49 every topic will be over time today? 08:58:32 Mister_Magister: That's not really under my control, I can't just cut people while they're still discussing. That's why I always ask everyone to ask for more minutes than they think they need 08:58:40 Sugash: Okay moving on 08:58:47 Jaymzz: if mines next, just skip 08:58:52 Sugash: You can continue in general discussion 08:58:54 Mister_Magister: That's up to you now ;) 08:58:55 Jaymzz: it was just saying chill man 08:59:19 Jaymzz: ok thanks 08:59:20 'Mister_Magister I'm chill :D 08:59:37 #topic Update libgcrypt to at least 1.6.0 (Asked by Mister_Magister – 5 min) 08:59:39 M4rtink: I'll give look when I have time. Normally, it seems to me that the remorse is intersepting the Component.delete signal or equivalent. 08:59:56 #info I need updated libgcrypt to at least 1.6.0 for telegram-purple 09:00:11 well there is nothing much to say. i've compiled 1.8 on my obs and it's working great so i see no reason in not updating it 09:00:17 and one more thing will not be outdated 09:00:19 dcaliste: that could help, good point 09:00:27 i need it for telegram plugin 09:00:35 as long as it does not break anything else I am fully for it 09:00:58 leszek: i didn't have any problems so far 09:01:42 M4rtink: indeed, see last part of /usr/lib/qt5/qml/Sailfish/Silica/RemorseItem.qml the remorse should be executed on destroy. 09:02:00 Just don't know who i should ask about it so i'm asking here 09:02:01 Mister_Magister: file a bug in mer and create pull request? 09:02:31 pketo: i don't understand what you mean 09:02:41 Mister_Magister: the licence is still LGPL, I've checked, so creating a pull request in git.merproject.org should work ;) 09:03:12 Fork https://git.merproject.org/mer-core/libgcrypt 09:03:26 Update it with 1.8 and create a MR. 09:03:31 ah okay i get it now will do 09:03:34 https://bugs.merproject.org/buglist.cgi?component=libgcrypt&list_id=2255&product=Mer%20Core&resolution=--- 09:03:51 #link https://bugs.merproject.org/buglist.cgi?component=libgcrypt&list_id=2255&product=Mer%20Core&resolution=--- 09:04:09 Mister_Magister: 1 minute :) 09:04:32 Jaymzz: it's done i think 09:05:05 Alrighty :) 09:05:08 Jaymzz: skip mine 09:05:09 Mister_Magister: the main thing to watch out for is a license change. if it changes to GPLv3 then you have to use the earlier one. 09:05:28 abranson_: why? 09:05:30 abranson_: I've checked that licence is still LGPLv2.1 09:05:36 but as dcaliste said, that one is ok 09:05:39 nh1402: Okay, do you want me to delete it from TJC also or do you want it discussed next time? 09:05:40 its that still an issue ? 09:05:43 Mister_Magister: not compatible with sfos 09:05:47 okay 09:06:00 Jaymzz: next time 09:06:11 nh1402: You got it. 09:06:25 see https://gnupg.org/software/libgcrypt/index.html 09:06:28 I kinda seem to remember some change in this regard (GPLv3) but I could be wrong 09:06:44 So, moving on to the next topic 09:06:46 #topic QtWebkit Update to 5.212 version (Asked by Leszek – 20 min) 09:06:57 #info Make QtWebkit and Webcat great again by updating the out of date version of qtwebkit 5.6 with the newest 5.212 version. This one offers accelerated canvas, fixed css, new javascriptcore and webcore supporting more modern sites. I have already created test builds and integrated the Jolla/Mer patches so it can replace QtWebkit 5.6 directly: https://build.merproject.org/package/show/home:leszek/qtwebkit5212-wHtml5Audio+Video Video 09:06:57 Demo: https://youtu.be/HfzcJNgi9wY Git Repo: https://github.com/llelectronics/lls-qtwebkit 09:07:07 with all the new Qt 5 stuff being GPLv3... 09:07:07 #link https://build.merproject.org/package/show/home:leszek/qtwebkit5212-wHtml5Audio+Video 09:07:07 #link https://youtu.be/HfzcJNgi9wY 09:07:07 #link https://github.com/llelectronics/lls-qtwebkit 09:07:30 leszek: The stage is yours 09:07:31 For the benefit of users and developers we should ship the best technical solution as default WebView we can provide. even if a new WebView based on gecko is in the works we should still provide and support which comes as default now 09:07:43 QtWebkit 5.212 uses an updated WebCore, JavaScript and CSS Engine which provides vastly improved support for modern websites 09:08:00 it does not crash that often :) 09:08:03 +1 for whatever improvements in Browser 09:08:19 html5 audio video support still rudimentary 09:08:26 Sugash: it's for Webview, not the browser which is using gecko engine. 09:08:33 it works good for audio but video struggles to play anything smooth when its over 480p resolution 09:08:41 and sound is muted everytime media loads (some error in gstreamer or the SailfishOS package of it I assume) 09:08:53 I made a javascript hack that currently only seems to work for audio sadly not for video (in webcat) 09:09:01 leszek: is this the same situation as previous topic, ie. merge request or is this a completely different repo? 09:09:34 nh1402: I think I can do a merge request but the code is different. I did not take a look at the license yet 09:09:59 so its a completely other code base basically speaking. 09:10:49 fair enough 09:10:55 With WebCore, JavaScriptCore and other components developed in a different repo aswell on github (I think the qtwebkit official github repo just merges big chunks in their its own branch) 09:11:13 Another thing device pixel ratio hack (to correctly scale websites for mobile phones) should land in SilicaWebView as it makes the most sense here and benefits all applications using this WebView 09:11:40 In general I am still scratching my head why SailfishOS does not use the platform scaling option provided by Qt (its not just me everyone from Qt is scratching their heads aswell as they created it for that purpose) 09:11:48 this would make this whole DPR patch unnecessary 09:11:55 and make things a lot easier 09:13:12 what a loud silence.. 09:13:18 leszek: maybe Jolla wants some control over the app UI scaling 09:13:23 current SilicaWebView or qtwebkit in SFOS ships btw. with deactivated HTML5 audio+video support and deactivated SVG support. I think this should change. We are living in 2017. We cannot ship a default WebView without those features 09:13:56 Tofe: they could write their own platform specific stuff then which is compatible with the qt scaling 09:13:59 leszek: I agree 09:14:43 the other sad part is the outdated web technologies when it comes to security 09:14:52 leszek: it would mean rewrite a big part of the theming code, and it's probably impacting apps themselves too 09:16:22 good point about security 09:16:25 Tofe: that is true. One might workaround this whole thing by using QT_SCALE_FACTOR and then set the Theme.pixelRatio variable to 0.5 or something. To scale down the whole UI Theme part 09:16:40 it usually a mess with webkit using software 09:16:46 M4rtinK: yeah HTTP/2 is not supported until Qt 5.8 09:17:08 so qtwebkit 5.212 in general supports it but only if you compile it against 5.8 09:17:09 a lot of effort went into fixing that in Fedora recently 09:17:30 and qtwebkit 5.212 of course in general brings in more security due to its newer Webkit Core 09:18:10 My current implementation also implements the Jolla/Mer patches. Though one patch in particular the "offline" patch does not work as the underlying code changed too much 09:18:28 I would again advice to file a bug in mer 09:18:35 I would like request by the package maintainer of the current qtwebkit to solve this issue so we have a drop in replacement for the current qtwebkit 09:18:59 well, if it has a maintainer 09:19:00 +help 09:19:19 M4rtinK: several patches went in qtwebkit on mer so I guess it must have some maintainer 09:19:39 that sounds promissing :) 09:19:58 Also for better html5 audio + video support we need someone to write a platform specific video/audio player utilizing the native Silica components 09:20:06 that would make video playback much smoother I guess 09:20:26 leszek: +1 09:20:26 we could base it on the qtmultimedia platform stuff though this one is lacking Media Source Extension compatibility 09:21:08 Looking at members of qtwebkit5, https://git.merproject.org/mer-core/qtwebkit5/settings/members I see many sailors who usually discuss MR within days of submission. 09:21:27 leszek: there are some parts in a UI which should not be scaled, i.e. expressed in pixels on purpose. Separator lines, borderimages, things like that, which would definitely not look like before. 09:22:07 without MediaSource btw. html5test.com will not show anything. I updated the code yesterday to allow setting MediaSourceEnabled from the QML side so the new RPMs from yesterday will have that and it doubles the points on html5test.com in comparison to the current qtwebkit 09:22:08 Tofe: even on an 8k 5.5" display? :) 09:22:18 Tofe: I am not sure about that 09:23:36 leszek, r0kk3rz: I'm not sure either, but I'm just saying the effort might be not worth it, given the fact that the result will be ... exactly the same UI-wise 09:23:43