| schmittlauchhehim[m] | Hey, could it be that some Jolla repos are down right now? URLs like https://releases.jolla.com/releases/4.4.0.64/jolla-hw/adaptation-common/armv7hl/repodata/repomd.xml (for the adaptation-common repo) and https://releases.jolla.com/jolla-apps/4.4.0.64/armv7hl/repodata/repomd.xml (for the apps repo) throw a HTTP 403 error for me. | 00:42 |
|---|---|---|
| schmittlauchhehim[m] | Ah nevermind, poetaster had already mentioned it, sorry. | 00:43 |
| *** amccarthy is now known as Guest2164 | 02:16 | |
| *** amccarthy_ is now known as amccarthy | 02:16 | |
| dcaliste | Hello pvuorela, just a word maybe, if you still have the time for, about the process to be able to use Incidence directly in the calendar manager. | 08:31 |
| pvuorela | dcaliste: hey, sure. | 08:35 |
| dcaliste | Thanks, the current situation is: | 08:35 |
| dcaliste | - the worker in a thread contains Incidence in a Calendar, it is doing operations like converting these into EventData and listing Occurrences for each dates in the loaded ranges, | 08:37 |
| dcaliste | - the manager is taking a copy of these EventData and may fetch some additional info like attendees asynchronously, | 08:37 |
| dcaliste | - when we want to add support for a new capability from KCalendarCore (multi alarm support, todos…) we need to add these in EventData and add the copy operations from Incidence, while basically there is a 1:1 mapping between our EventData and Incidence. | 08:39 |
| dcaliste | The open question I'm wondering about is "can we somehow avoid the duplication, while still delegating DB operation and other costly operations to a worker thread ?" Now that's where I'm working: | 08:40 |
| dcaliste | - implement behind the ExtendedStorage API an asynchronous access to the DB with a worker thread. | 08:41 |
| dcaliste | But it is raising some difficulties: | 08:41 |
| dcaliste | - some operations done by the calendar worker are not DB centred but still costly, like the computation of all occurrences for given dates, or sending email for meeting events. | 08:42 |
| dcaliste | - some operations done by mKCal are not related to DB access neither like the alarm handling with timed, but requires some additional calls to the DB (like when a notebook is nodified, one need to get a list of all DB event if the visibility is switch on). | 08:44 |
| dcaliste | I've now an implementation of the ExtendedStorage API done async that is working fairly well, but I'm wondering how to solve the two mentioned difficulties. This is not a question I'm waiting an answer for, don't worry, but mostly to expose the problem and trying to make it clear. | 08:45 |
| dcaliste | During the long break to come from your side, pvuorela, I'll try to address these issues and see if I can go somewhere with these. | 08:46 |
| dcaliste | So it was mostly to keep you inform about the process and open questions at the moment. | 08:46 |
| pvuorela | alright. sounds like there is need for some threading anyhow. | 08:47 |
| dcaliste | The code is currently in a branch on my fork in Github. The branch is misleadingly named "cleanup". | 08:47 |
| dcaliste | Indeed, and I'm wondering at the moment if the alarm code for timed for instance should run in the worker thread or in a dedicated thread by itself. | 08:48 |
| dcaliste | The first is obviously better, but it means that the DB thread operation is not an implementation detail anymore, but should be exposed in a way. | 08:48 |
| dcaliste | The second option solves certain problem, but requires more synchronisation between the DB worker thread and the Timed thread, which may be fragile and requires much more copies. | 08:52 |
| dcaliste | So that's the kind of issues I'm facing at the moment, but I'll have time to experiment different solutions ;) | 08:53 |
| pvuorela | alright :) | 08:53 |
| poetaster | piggz[m], can you think of any reason why sailfish-components-webview-qt5 would no longer render mobile? I have a feeling like I downgraded something, but I checked and re-installed latest. | 09:08 |
| *** fifr[m] is now known as Guest2227 | 11:29 | |
| henk | huh, I think I have another idea what might be causing my "swiping right in an sms message shows the wrong contact"-issue. I sent and received sms messages with a contact yesterday (let’s say "adam") and kept them open. then today I received a sms message from a different contact (let’s say "bob") and "opened it" from the notifications screen. swiping right to left opened the | 12:53 |
| henk | contact screen for "adam", though. switching back (swiping left-to-righ) to the sms messages overview, selecting "bob’s" messages, then again swiping right-to-left, then opens "bob’s" contact screen. | 12:53 |
| henk | bit hard to try to reproduce, I need two other people (or phones) to write me messages … | 12:53 |
| piggz[m] | <poetaster> "piggz, can you think of any..." <- What have you done? 😂 | 12:55 |
| poetaster | that's a good question. but it seems odd that it could be the same on two different phones! I'm guessing that I overwrote something in .local/share while recovering stuff. | 12:56 |
| poetaster | but I cherry pick, or so I thought. Now I just flashed and discovered the august image is still a broken gs290 . damn it, dman it, damn it. | 12:58 |
| poetaster | I don't get it. that was the latest image I downloaded for sfe-yggdrasil/Sailfish_OS/ | 13:00 |
| henk | does anyone happen to have two phones and can help me test whether this behaviour is reproducible? | 13:02 |
| direc85[m] | The repos seem to be back up again | 13:05 |
| *** amccarthy is now known as Guest2246 | 13:23 | |
| *** amccarthy_ is now known as amccarthy | 13:23 | |
| poetaster | henk, I can write you a message | 14:50 |
| poetaster | but only from one number ;-( | 14:50 |
| henk | poetaster: when my GF is available, she can be the second … currently when I swipe from her message r2l (right-to-left), I get to her contact. can you /query me your number so I can create a "contacts" entry for you? I guess that’s required for it to work … | 14:52 |
| poetaster | +491799051735 | 14:53 |
| poetaster | heh. | 16:11 |
| poetaster | piggz[m], I'm not mad. Fresh install. ssu to .72 , no funny b'ness. install a webview equipped app (or two) and watch them fail. that's on the GS290 .... | 16:12 |
| poetaster | I should have stayed at Akido and done an hour of rolling around ... ;) | 16:13 |
| piggz_ | poetaster: so, i need to understand of this is a port or an app issue ... does the browser work ok? | 18:17 |
| poetaster | piggz_, the browser seems fine. but apps that use and embedded webview are not. Well, the old QtWebkit ones behave as expected :) | 18:18 |
| poetaster | But I don't see how it could be the port per se if it's volla and the 22? | 18:19 |
| poetaster | just booting my smashed, reference device where this all 'just worked' | 18:20 |
| poetaster | also 4.4.0.72 GS290 and there, in fact, every thing works as expected. I'll try to isolate the difference. The GS5 threw me. | 18:22 |
| poetaster | don't get it. these devices contain the same foo. Ok. I'll go through all the packages first, but I think there identical since I ref/up all the time. | 18:26 |
| poetaster | piggz_, the webview and the browser are two different beasts. for 9 months I've been posting bug reports about stuff that works in browser but not in webview (audio, certain JS, incorrect inner-width, blah, blah). | 18:28 |
| piggz_ | poetaster: are you using sailfish.webview? | 18:44 |
| poetaster | piggz_, yes, I've been fixing various things in apps with that. for instance, login on tooter, webviews of articles in tidings, etc. | 18:55 |
| poetaster | piggz_, on the GS5, the godot 3.2 builds which work on the GS290 do not work. but that's another wrinkle. | 18:58 |
| poetaster | piggz_, maybe I'm on developer instead of testing. I was dead certain, but? | 19:01 |
| piggz_ | i dont think it should matter | 19:02 |
| poetaster | piggz_, yeah, it's kind of standard. one thing I noted on the last flash, when updating from .68. .72 I got a number of hybris related downgrades which looked odd. | 19:04 |
| poetaster | The following 16 packages are going to be downgraded: | 19:05 |
| poetaster | libargon2 libargon2-tools libgbinder libKF5Archive5 libsfosdevenc libtls-padding pulseaudio-modules-droid-jb2q pulseaudio-modules-droid-jb2q-common sailfish-device-encryption-community | 19:05 |
| poetaster | sailfish-device-encryption-community-droid sailfish-device-encryption-community-generator sailfish-device-encryption-community-service sailfish-device-encryption-community-settings | 19:05 |
| poetaster | sailfish-device-encryption-community-wizard sailfish-fpd-community systemd-ask-password-gui | 19:05 |
| poetaster | The following package is going to be reinstalled: | 19:05 |
| poetaster | lipstick-qt5 | 19:06 |
| poetaster | The following package is going to change vendor: | 19:06 |
| poetaster | qtscenegraph-adaptation meego -> chum | 19:06 |
| poetaster | Must get child to bed. | 19:06 |
| poetaster | not hybris related. | 19:08 |
| poetaster | piggz_, malfunctioning have stuff like: i|adaptation-community|droid-config-yggdrasil-preinit-plugin|0.0.2-1.1.4.jolla|aarch64 | 20:06 |
| poetaster | funtioning: i|@System|droid-config-yggdrasil-preinit-plugin|0.0.2-1.1.1.jolla|aarch64 | 20:06 |
| poetaster | working, and not working: | 20:13 |
| poetaster | but.txt:i+|@System|droid-config-yggdrasil|0.0.2-1.1.1.jolla|aarch64 | 20:13 |
| poetaster | not.txt:i+|adaptation-community|droid-config-yggdrasil|0.0.2-1.1.4.jolla|aarch64 | 20:13 |
| piggz | HengYe[m]: PP keyboard works on the the pro, just need to make it work on the original | 21:25 |
| *** amccarthy is now known as Guest2296 | 22:24 | |
| *** amccarthy_ is now known as amccarthy | 22:24 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!