Tuesday, 2021-07-20

dcalisteHello chriadam, sorry to be late today.07:10
chriadamhi dcaliste, no problem at all07:10
chriadamI hope you had a nice vacation!07:10
dcalisteYes thanks, the outside temperature was very hot, but overall it was enjoyable.07:11
chriadamnice one07:12
chriadamregarding the open PRs: before Pekka left to vacation, he did a quick round of reviews on the open PRs and had some various comments07:12
chriadamso I didn't end up merging those ones - figured it would be better to wait until you can respond to those review comments, at least07:12
dcalisteIndeed, I noticed his comments, but I still haven't got the time to reply. I should be able to do this week, or even in the coming days though.07:14
chriadamNo rush - Pekka is on vacation for a bit longer anyway07:14
chriadamso if there's something which needs further discussion, it will not happen until he gets back I guess07:14
chriadamaside from that, was there any specific PR which you think can be progressed on our side, currently?  if so, I will put that on my list for tomorrow07:15
dcalisteI didn't touch my computer during vacations, and I came back some days ago, mainly working for my job actually. But before leaving, I worked on this one : https://github.com/sailfishos/messagingframework/pull/207:16
chriadamouch, you lost your emails running the test :-(  ugh07:17
chriadamwe should insulate that better in qmf, I guess, using a test-specific database or similar, I think :-(07:18
chriadamso, which two tests are still failing?07:19
chriadamis it tst_QMail_ListModels::test_qmailaccountlistmodel and tst_qlogsystem ?07:20
chriadamI will run the test suite using the current qmf code and see if the tst_qlogsystem failure occurs currently - I think it probably will, honestly.07:21
chriadamwell, I will have to do that tomorrow unfortunately, not tonight07:22
attahSince this is in your area; i just wanted to check if i was understandable, and that you think my comment on this PR will eventually be looked at https://github.com/sailfishos/nemo-qml-plugin-email/pull/207:24
chriadamlet me check07:24
chriadamyou are absolutely correct, I think07:26
attahgood, that's one bug i don't need in my life :)07:27
attahI always thought *this* was the point of the silly folders...07:27
chriadamthe reasons are probably lost in the mists of time07:28
chriadamI can ask Matt tomorrow07:29
chriadamthank you for pointing out the potential issue07:29
dcalisteAbout download location, I agree with all the points here. But I would like to add that it would be good to tell QMF to actually use a given location. Because at the moment, the attachments are stored twice, once under .qmf and then copied to ~/Downloads/ But it's a large amount of work...07:30
attahhappy i could be of any help, even a blind squirrel and all that :)07:31
dcalisteAbout the failing tests, chriadam, you'll see that the account one is still failing. It is what I was testing when I lost all my emails. And I stopped digging further at that moment.07:32
dcalisteThe isolation issue is not in QMF actually. Because I setup a different HOME via environment variable and this did the trick for QMF and accounts.07:33
chriadamwell, I will have a look into that one tomorrow07:34
chriadamanything related to a&sso integration is likely to have some issues, honestly.07:34
dcalisteBut the problem in some watchers over DBus I guess that noticed that the account xx was deleted (by the test), and decided to actually delete all related data to account xx. Since account ids are not unique within two databases, this also deleted my real account that shared the same id.07:34
chriadamouch.  yes.07:35
chriadamwell, thanks for looking into the other issues and fixing those.  I will add some comments to the MR tomorrow based on my findings07:36
chriadamdid you have anything else to discuss today?07:36
chriadamor flypig did you have something to discuss with dcaliste?07:36
dcalisteNo that's all, maybe looking for flypig, if you know if he's around ?07:36
chriadamhe was earlier, at least07:37
dcalisteIt's about this thread : https://forum.sailfishos.org/t/calendar-entries-scheduled-time-one-hour-later-than-set/6480/07:37
dcalisteAn issue in my opinion in EAS sync plugin when the server and the device are not using the same timezone.07:38
flypigHi, sorry for not joining this morning.07:38
flypigI'm glad you had a good vacation dcaliste.07:38
dcalisteHello flypig, thanks. I just would like to bring to your attention, when you have time, a possible bug in the sync plugin for Exchange server running with a timezone different than the one on device.07:40
flypigOkay, so the situation arises if the server timezone and device timezone don't match.07:40
dcalisteWhen creating an event on device in a different timezone, after it is send to server and come back, the event on device is using the time in device time zone but with the server timezone.07:41
flypigI'm not aware that anything specific changed in this regards. It must have been like this for some time.07:41
flypigIt certainly looks like a problem.07:42
chriadam(is it just: on the device, we assume that the incoming datetime is in LocalTime spec rather than UTC spec, if no specific timezone specification is received with the timestamp?)07:42
flypigGood question chriadam.07:42
dcalisteThe symptom, is that indeed the localtime is used but the timezone of the server is applied on it, instead of the event timezone.07:43
chriadamsorry, what do you mean by "the timezone of the server is applied on it" ?07:44
attahAlso note that the tz of the event was set to UTC by default when creating it... shouldn't input always be in local time?07:44
chriadamlocal time certainly sounds more reasonable to me07:45
chriadamI might be misunderstanding something in that case07:45
dcalisteSo an event that is sent as 14.00 UTC+1, is read back as 14.00 UTC. I suppose that the server is sending back 14.00UTC+1 in a UTC calendar, and the sync plugin is confused by using the calendar tz with the event time disregarding the event tz itself.07:45
chriadame.g. it might be shown in the UI in local time, but stored to the db in UTC07:45
chriadamah I see07:45
dcalisteattah : actually no, creating an event on device always propose by default the local zone, not UTC.07:46
dcalisteThis was not clear with Steve first post, but was cleared when he described what happened in detailled steps.07:46
flypigI'm just checking whether we have an internal report lodged about it... I don't see anything.07:47
attahAh, my bad... so that first screenshot was having re-opened the event, and not creating it?07:47
dcalisteattah: indeed. It was not clear. But the first screenshot is a modification.07:48
attahsorry for the noise07:48
chriadamflypig: fwiw my tz is very different to our internal server's tz, so if you need me to test something, give me a yell haha.07:49
dcalisteattah, no problem.07:49
flypigThanks chriadam, and also attah (it's always useful to discuss to get clarity).07:49
chriadamdcaliste: thank you for bringing our attention to this07:49
flypigYes, agreed, thank you dcaliste. We can try to reproduce it.07:50
dcalisteI've no idea, if it's simply because of different tz for server and event or if it's important that the server must be UTC...07:50
flypigI don't see any internal bug. I thought there was something similar but can't find anything right now. If there isn't I'll create one and post a note on the forum.07:51
chriadamthanks flypig07:52
chriadamwas there anything else to discuss tonight?07:52
dcalisteNo that's all, thanks.07:52
chriadamgreat, in that case, I hope you have a great week.  I will test the qmf PR and also vanilla qmf unit tests tomorrow.07:53
* chriadam -> home, gnight07:53
dcalisteThank you, good night.07:53
devnull-Hi there, maybe I missed a information. But why are packages for broken due to the fact, that sailfish-weather is not available anymore ? for example when I try to install patterns-sailfish-store-applications in the emulator I get this error notice:09:17
devnull-Fatal error: nothing provides sailfish-weather needed by patterns-sailfish-store-applications-1.0.25-1.16.1.jolla.noarch09:17
ViGedevnull-: That's interesting. What does the output of command 'ssu lr' look like?09:27
devnull-moment please09:35
devnull-[root@SailfishEmul pkgs]# ssu lr09:36
devnull-Enabled repositories (global):09:36
devnull- - adaptation0    ... https://store-repository.jolla.com/releases/
devnull- - customer-jolla ... https://releases.jolla.com/features/
devnull- - hotfixes       ... https://releases.jolla.com/releases/
devnull- - jolla          ... https://releases.jolla.com/releases/
devnull- - sdk            ... https://releases.jolla.com/releases/
devnull-Enabled repositories (user):09:36
devnull- - store ... https://store-repository.jolla.com/SDK Emulator Wayland/i486/?version=
devnull-Disabled repositories (global, might be overridden by user config):09:36
devnull- - apps ... https://releases.jolla.com/jolla-apps/
devnull-Disabled repositories (user):09:36
devnull- - home ... https://download.jollamobile.com/home:/honeybadger/latest_i486/09:36
ViGeright, somehow I missed the fact that you are talking about the emulator09:43
ViGeAs you can see the apps repository is disabled on the emulator09:44
devnull-@ViGe: thanks for the hint, missed that ... so I try to enable it and then SFOS is nice to me again ?10:07
ViGeSomehow I doubt that10:08
ViGeI mean, sure, you can install sailfish-weather after enabling the repository, but why would you even want to do that?10:08
devnull-@ViGe: You're right I've told only a half of the story ... Lets come from the real problem side. I've got a backup of my device here done, tar(ing) the home-dir to a file ... then I missed somehow to create a SF Backup with onboard tools ... After upgrading to SFOS 4 I found that my contacts and calendar is not useable anymore due to the fact, that at least the sqlite-db schemas have changed.10:15
devnull-@ViGe: Now I try to restore my contacts in the emulator, what have worked already but I've not found a working way to export that to a vcard or somewhat else like Webdav Server or simply do a backup with vault ...10:16
ViGeI see10:19
*** pa- is now known as pa15:04
fridllbt: thanks for the Info with the right API. I do check the logs usually after asking a question. So no info got lost :-)15:40
fridlbut I think my python dependency tree is to long to get it realistically resolved.16:19
piggzPSA: https://forum.sailfishos.org/t/announcing-sailfishos-chum/738117:53
fridlGreat Idea! Nice!17:57
fridlthis will make things easier in the future and uses all the nice possibilities coming with obs :-) Now I finally need to get anything working on OBS myself...18:02
rinigushopefully we can get all those aarch64 builds appear easier18:22
fridlwhat am I doing wrong when obs says "excluded" for aarch and arm on my app? (in my home repo?)18:27
fridlthere is the rpm dir in the repo and it is building in the SDK (as far as i know)18:28
rinigus... checking ...18:34
rinigusfridl: service should be not webhook but tar_git18:35
rinigusline 2 in https://build.sailfishos.org/package/view_file/home:fridlmue/harbour-avarisk/_service?expand=118:36
fridlOk, i'll change. I was trying to follow this. https://forum.sailfishos.org/t/community-obs-refurbished-and-re-floated/6556/22?u=fridlmue18:38
fridland it is 'url' then in line 3, not 'repourl'18:41
fridlfor the webhook working i need to add that line as well, right?18:42
rinigusI don't use webhooks myself, so don't know much about them. looks to be right...18:45
piggzi also find it easier to not use webhooks now...18:52
piggzthe service trigger button works these days18:52
rinigusfridl: thanks for submission! just replied to it and all what's needed is a specific tag or commit ID in _service19:39
fridlrinigus: Ok, I'll check that out!19:58
rinigusfridl: you could just tell me the version/tag and I will set it in the _service19:59
fridlrinigus: the version would be the release tag. The one from github is used then, right?20:00
rinigusyes, from github.20:01
fridlyeah, that would be nice.20:01
rinigusso, which one - V0.7-4 or the one with pre-release?20:01
fridlShould I respond in OBS as well, to test something?20:01
rinigusyou could write the version there, if you wish :)20:02
phoenix616hello, I'm looking for some documentation on keyboard layouts, especially what the possible objects and their fields are for the qml files. I tried looking online but the links to the maliit api reference were broken :S20:06
phoenix616although I guess the docs would need to be for the jolla keyboard plugin? not sure what exactly specifies the possible options in the design20:07
rinigusfridl: as far as I remember, OBS doesn't like to have '-' in the version info of the tag. hence the strange version of your package.20:12
rinigusOBS rewrites the version by tag. can you try to tag as 0.8.3 instead of V0.8-3?20:14
rinigusotherwise all looks to be fine...20:14
rinigusI wonder if we can force version to be kept as in RPM SPEC by OBS. well, except the release bit20:16
fridlrinigus: i changed that in github20:16
fridland will keep that versioning then.20:16
fridlphoenix616: Sorry, we are not ignoring you. I just can't help!20:17
phoenix616fridl: don't worry, I understand :)20:18
phoenix616I also just found that the base qml files from the jolla keyboard plugin are just in a folder so maybe I can find my way from that already20:18
rinigusfridl: can you try without V in front? it still is somehow unhappy :(20:19
fridlrinigus: done.20:19
rinigusfridl: and then the old tags are kept - maybe that is also confusing. see that the same commit has 3 tags assigned now20:20
rinigusignore it, it worked it out20:20
rinigusfridl: testing build was fine and its building in chum now. thank you!20:24
fridlwow, very nice! :-)20:34
fridljust for the procedure: When I update the version on github, is chum then automatically updating, or do I need to trigger it there manually or do I need to ping s.o.?20:35
fridlor submit the updatet version to testing?20:35
fridlrinigus: nice, and installation on my devel Xperia X worked fine from Chum with Zypper :-) Thumbs Up!20:59

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