| *** amccarthy is now known as Guest1716 | 02:51 | |
| *** amccarthy_ is now known as amccarthy | 02:51 | |
| rinigus | poetaster: we have been discussing these requirements issues most of Saturday. I have a feeling you should read up on how rpm requirements are set, including manual and automated way. | 04:32 |
|---|---|---|
| *** amccarthy is now known as Guest1728 | 06:03 | |
| *** amccarthy_ is now known as amccarthy | 06:03 | |
| dcaliste | Good morning pvuorela, how are you ? | 07:00 |
| pvuorela | dcaliste: good morning. i'm fine, thanks. how are you? | 07:02 |
| dcaliste | I'm good thanks. Moving our meeting last week (and making an exception in my calendar), I noticed a bug that may exist for a while, but I've never encountered before, see https://github.com/sailfishos/buteo-sync-plugin-caldav/pull/23 | 07:03 |
| pvuorela | alright. i'll check. | 07:04 |
| pvuorela | good thing that it got moved :) | 07:06 |
| dcaliste | Indeed ! I'm still surprised, I didn't notice it before, or anyone… I still wonder if it's coming from very recent changes, because I'm running currently nemo-qml-plugin-calendar with the uniqueId PR. But looking at the master code, I don't see how URL can survive the dissociation -> Data::Event -> KCalendar::Event mechanism. | 07:09 |
| dcaliste | It could be interesting to move the dissociation code into the Worker:saveEvent() routine, but since it can fail, I'm not sure it's a good idea. Even if early failure when presenting the modification page to the user as we have it now is not handled either. | 07:11 |
| dcaliste | Besides, as mentioned, I've been running https://github.com/sailfishos/nemo-qml-plugin-calendar/pull/54 for several days, and I'm very happy that nothing breaks. I've removed the WIP tag accordingly. | 07:39 |
| dcaliste | Of course, you also need jolla-calendar#335 and jolla-email#539. | 07:40 |
| dcaliste | By the way, what do you think, pvuorela, about the solution I propose for the alarm compatibility ? Alarm would still store the triplet notebookUid, Uid and recurrenceId. And CalendarUtils in jolla-calendar implement a routine to get the uniqueId from the triplet. | 07:41 |
| pvuorela | nice. i'll check again those too and let's hope to get them in. | 07:41 |
| dcaliste | I would have prefered to have this convenience routine in nemo-qml-plugin-calendar since all the code related to the uniqueId is there, but no library is exported there, just the QML bindings. I could create a singleton there, taking the CalendarUtil code from jolla-calendar though. What do you think ? | 07:43 |
| pvuorela | trying to find the jolla-calendar part | 07:43 |
| dcaliste | It's in compànents/plugin.cpp | 07:44 |
| pvuorela | not spotting changes on that and the jolla-calendar PR is wip. is that up to date? | 07:45 |
| dcaliste | Oops, I forgot to push the commit add this. | 07:46 |
| pvuorela | guess it's now better. at least the snippet doesn't look too long so maybe just good enough for now. | 07:51 |
| dcaliste | Ok, nice then. I will require a commit in jolla-alarm, to add the conversion from the triplet to the uniqueId when calling the (new) DBus eventView() API. See /usr/share/jolla-alarm-ui/pages/CalendarAlarmDialo | 07:52 |
| dcaliste | s/I will/it will/ | 07:52 |
| pvuorela | on that i'm still pondering a it whether dbus api should be more like uid+recid. notebook maybe as hint. | 07:53 |
| pvuorela | dunno. | 07:53 |
| dcaliste | If we accept different events sharing the same uid, notebookUid must be added to the viewEvent() API, which is an API break anyway. Thus changing it to use the uniqueId make the API simpler. | 07:56 |
| dcaliste | Of course, at the moment, notebookUid is redundant with the uid argument since we have a 1:1 notebookUid <-> uid relation. | 07:56 |
| pvuorela | i'm not sure if that's entirely mandatory. sure the uid+recid will not necessarily be unique without the notebook information, but on the common case there should be only one instance to be fetched | 07:57 |
| pvuorela | just that uid+recid type of api is nice while instance identifier gets a bit into qml and implementation specifics. | 07:58 |
| dcaliste | Exact, instanceIdentifier (at the moment) and uniqueId (later) are implementation details of the QML bindings. The solution could be to keep the triplet in the DBus API and then, use the converter to the uniqueId there for internal business. | 08:00 |
| pvuorela | yea, something like that. | 08:01 |
| pvuorela | and guess there could be some lookup method to search if some notebook has specific uid+recid instance. | 08:03 |
| dcaliste | This will require to add a viewEventByIdentifier(uniqueId, startDate) API to the DBus interface, because there are cases where we just have the uniqueId at hand, See components/CalendarWidget.qml for instance. | 08:04 |
| pvuorela | sure. adding such should be ok. | 08:07 |
| dcaliste | Ok, then, I'll amend the commit accordingly and remove the WIP tag. | 08:08 |
| dcaliste | While running the tests for the uniqueId PR in nemo-qml-plugin-calendar, I noticed that the tst_calendarsearch was segfaulting sometimes on quit. This is due to the model trying to disconnect from an already delated manager. see https://github.com/sailfishos/nemo-qml-plugin-calendar/pull/56 | 08:11 |
| pvuorela | looks good. just added some minor remarks. | 08:17 |
| dcaliste | Thanks, I've addressed your remarks. | 08:25 |
| pvuorela | thanks. | 08:27 |
| dcaliste | And to finished, importing a train travel ics file this week-end, I notice a mistake I introduced in the ImportPage, see jolla-calendar#336, in case there is no attendee to the event :/ | 08:30 |
| pvuorela | keep 'em coming :) | 08:35 |
| dcaliste | I've a meeting starting now with our PhD students. Thank you for the discussion and the PR reviews. Enjoy your week ! | 08:38 |
| pvuorela | thank you too! see you. | 08:39 |
| stboom | Hello Together | 10:44 |
| stboom | I was wondering if anyone has experience with the sony xperia 5 III and salfish os. I was wondering that the 10III is supported and the 5III is not on the list. | 10:46 |
| poetaster | rinigus, I'm still unable to get opt-kf5-libqxmpp symbols working. I hacked in a provides to test and that does work for 32 bits, but I don't get it: https://build.sailfishos.org/package/binary/sailfishos:chum:testing/opt-kf5-libqxmpp?arch=i586&filename=opt-kf5-libqxmpp-1.5-1.42.1.jolla.i486.rpm&repository=4.5.0.19_i486 | 12:13 |
| poetaster | rinigus, the other qxmpp I built uses different library names and it works for compiling shmong. | 12:14 |
| poetaster | rinigus, never mind. I'm giving up the opt version. | 13:45 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!