Tuesday, 2020-07-28

*** zbenjamin_ is now known as zbenjamin01:03
*** frinring_ is now known as frinring03:11
dcalisteHello chriadam, how are you ?07:06
chriadamhi dcaliste, I'm well thank you.  how are you?07:07
chriadamhow is your hand?  improving, I hope?07:08
dcalisteI'm fine, thank you. My hand is verry slowly improving. I still cannot move much my third finger though.07:09
chriadamhopefully it continues to improve07:10
dcalisteDoctors are optimistic, so why not be myself ;) On another subject, I've encountered a strange behaviour recently having two alarms for my events with exceptions.07:11
chriadamI saw that you created an mkcal PR for something related but didn't have a chance to review it yet07:11
dcalisteIt's related to mkcal MR4307:11
dcalisteBut I don't know if it's due to various changes we discussed last week in caldav sync plugin.07:12
dcalisteIn particular the fact to remove the copy function in incidencehandler and replace it with kcalcore operator=07:12
chriadamah, our hand-rolled one might duplicate alarms also in some case where it shouldn't?07:13
dcalisteThe issue is from timed where the recurrence id is stored as a string and sometime expressed in my time zone or in UTC.07:13
dcalisteNo, no, it's the recurrenceId representation that is failing.07:13
dcalisteIn the handler in mkcal, I still cannot figure out exactly why my exceptions are sometimes represented in my time zone, and sometimes (after sync) are represented in UTC.07:14
dcalisteSo, when the mkcal handler try to match an existing alarm to remove it and readd it, the match is failing and it is adding a second with the other representation.07:15
chriadamtz stuff is always tricky :-/07:17
chriadamI had hoped we had squashed all those possible cases in the past, but I guess not ;-)07:17
chriadamso your PR does what - checks whether it matches in device_tz OR utc_tz representation?07:17
dcalisteThe MR, instead of relying on the string representation, is converting back the string to KDateTime, and then, I use operator==, which does the job to compare times whatever the TZ.07:18
chriadammakes sense07:19
chriadamthanks for looking into that.  I also wonder why the representation might be different in different circumstances, but perhaps it doesn't matter07:20
dcalisteI still need to test it properly. On paper, (besides implementation mistkes) it looks alright, but I need to use it for some days in real life to be sure. About why the representation is changing, this is my open question at the moment ;)07:20
chriadamnot something as simple as "the representation we get from the server changes between syncs" due to server doing weird things?07:22
chriadammost likely it's something tricky on our client side, though, depending on device tz, the phase of the moon, etc :-P07:22
dcalisteI think, it's indeed due to the server changing the representation of recurrenceId, but it was not happening before. So looking at the removal MR of the incidenceHandler::copyIncidence() func, it seems that the recurrenceId was not copied there.07:24
dcalisteSo, the original recurrenceId from the creation of the incidence was still used, whatever the server was doing to it.07:24
dcalisteAfter the MR removing this function, the recurrenceId it indeed changed to the value of the server.07:25
dcalisteThat may be the best explanation.07:25
chriadamah.  I wonder if we didn't update it because it matched (kdatetime comparison) even though the representation itself was different07:25
chriadameither way, probably best to always keep in sync with server side07:26
dcalisteYes and it simplifies the code much, avoid exception for this or that. But it requires to patch mkcal to avoid doubling alarms for exceptions.07:26
chriadamsounds reasonable.  thanks.  unfortunately I probably won't get a chance to review it this week, but hopefully next week.07:27
chriadamI'm currently deep in qtcontacts-sqlite / qtpim things.  finally making the robustness improvements to allow properly separated addressbooks.07:28
dcalisteNo problem for the timeline, it will be for version > 3.4.0 anyway, so I have more days to use it and discover minor issues like that.07:28
dcalisteI've seen your branch in git indeed. That's very good, thank you.07:29
chriadamthank you for your patience, and your efforts as always ;-)07:29
chriadamI poked Martin about the jolla-calendar changes, yesterday07:29
chriadamnot sure if he has had a chance to look at those yet (I guess not, he just got back from vacation)07:29
dcalisteThank you. I'm proposing also jolla-calendar!267 to change the focus behaviour when we have finished with the text editing of an event but the keyboard pops up nonetheless.07:30
chriadamgreat - mentioned that one to martin too07:31
dcalisteOk thank you very much. Of course he will be very busy returning from vacations, so no hurry.07:33
*** vilpan is now known as Guest4024807:33
*** Guest40248 is now known as vilpan07:33
chriadamI didn't have anything more to discuss - due to contacts stuff I haven't had a chance to progress any reviews on our side, unfortunately.  similarly, I didn't get a chance to look into the TJC issue with syncloop triggered between contactsd and carddav (TJC#229248), but I'm hoping that will be resolved automatically due to the changes I'm making for addressbook separation etc07:34
chriadamaside from those we already covered, did you have anything else to discuss?07:35
dcalisteNo problem with time, you're very busy for sure. Just one stupid question, with the new forum based on discourse, what is the best way to mention a topic in a commit ? (like old TJC#xxxxxx)07:35
chriadamI have no idea :-(07:36
chriadamgood question...07:36
chriadamone sec07:36
chriadamno response is forthcoming just yet, I will get back to you on that one ;-)07:39
dcalisteThanks, it's not a big deal, but citing the TJC id was convenient, and there was almost always a TJC entry for something ;)07:40
chriadamdefinitely an important thing to be able to accept community contributions for issues raised on the developer forum, so I'm certain there is a process / tag format, just not sure what it is07:40
chriadamthank you for raising it07:41
dcalisteThat being said, thank you for the meeting. I'll continue testing the changes in caldav plugin.07:41
chriadamthanks very much.  have a great week!07:41
dcalisteYou too.07:41
mad_devdoes anyone know how images are served in settings | accounts | service(twitter - google)?10:25
flypigmad_dev, I'm not sure I understand; could you explain a bit more what you're trying to do?10:52
mad_devflypig, I am working on RTL support. I came across an issue in the account settings page where the mirroring declaration did not affect the icon.10:54
mad_devI checked /usr/share/accounts/ui/twitter.qml | twitter-settings.qml | twitter-update.qml | TwitterSettingsDisplay.qml all in ui/accounts10:55
mad_devI also checked /usr/lib/qt5/qml/com/jolla/settings/accounts/ nothing stood out10:55
flypigAh, okay, now I understand. You want the QML page which draws the icon.10:55
flypigLet me have a dig around.10:55
mad_devflypig, yup. I did find /usr/share/lipstick/eventfeed/twitter-delegate.qml and will test now. However, I don't think it's it10:56
flypigIt looks like the actual icon is shown by /usr/lib/qt5/qml/com/jolla/settings/accounts/AccountSummary.qml10:59
flypigWhich is included by /usr/lib/qt5/qml/com/jolla/settings/accounts/AccountMainSettingsDisplay.qml10:59
flypigSee lines 46-51 of that second file.11:00
flypigCould that be what you're after?11:01
mad_devhmmm, it could be. I think my mirroring declaration might be the issue. I am using this https://github.com/Logic-gate/unofficial-jolla-language-pack-ar/blob/master/tests/rtl._qml11:03
mad_devI tested AccountSummary with the declaration, but nothing happened.11:03
flypigThe whole issue of RTL is new to me, so I'm not familiar with LayoutMirroring I'm afraid. What would you expect to happen to the icon and text in the summary?11:06
mad_devto be mirrored to the opposite declared direction11:08
mad_devI'll test different approaches and report back. Cheers for the help11:09
flypigReading up on LayoutMirroring (here https://doc.qt.io/qt-5/qml-qtquick-layoutmirroring.html) it looks like the issue is that mirroring applies to rows, but this has been implemented by setting anchors directly.11:11
flypigAlthough the docs do also say it reverses the anchor settings.11:12
mad_devYessssss..it did in fact work, I forgot about whole numbers...0 == 1. https://twitter.com/m4d_d3v/status/1288069942957481984/photo/111:13
mad_devmy line numbers were wrong11:13
mad_devcheers flypig11:13
flypigmad_dev, that's great :)11:14
flypigmad_dev, out of curiosity, what do you mean by whole numbers here?11:15
mad_devMy approach to RTL support is to simple append the declaration at certain line. I use cat -n which starts count from 1. But sed counts from 0 as usual. So every time I see the place where the declaration needs to be, I need to subtract one...so, a simple mistake from my part11:18
mad_devA rudimentary approach...but it works11:19
mad_devwhole numbers as in 0, 1, 2, 311:20
mad_dev...and yes. This is done manually :) I go through each file, find the most suitable line, then append.11:23
flypigmad_dev, thanks for the explanation; that makes perfect sense then!11:36
flypigIt sounds like a lot of work for you, but sometimes the simplest solutions work best, especially when you're figuring things out still.11:37
mad_devYup, the other solution is sounds cool but it's not practical. An initialization wrapper which starts apps, if the system locale is set to an RTL language, insert RTL declaration at the beginning of each file. If not, launch normally.11:45
mad_devJust a thought, never explored it more.11:45
zotanwith rpmbuild is there an easy way to stop it rebuilding everytime? My install step is failing, but the build step takes 20 minutes :-(16:59
flypigzotan, apparently you can do `rpmbuild --short-circuit -bi /path/to/specfile`17:01
flypigI've not tried it, but someone said it here: https://serverfault.com/a/21756117:01
zotanflypig: That seems to work, thanks!17:05
flypigGreat :)17:05
piggzflypig: anything to report on network first boot issue?21:51

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!