Tuesday, 2019-03-26

*** frinring_ is now known as frinring06:01
*** ecloud_wfh is now known as ecloud06:38
dcalisteHello chriadam, how are you ?07:52
chriadamhi dcaliste07:52
chriadamI'm well thanks!  how are you?07:52
dcalisteI'm fine, thank you.07:53
chriadamI saw that Pekka just merged his refactor PRs to jolla-email + nemo-qml-plugin-email07:53
chriadamare you willing to rebase your signature ones on top of current master branches?07:54
dcalisteYes, I'm on it.07:54
chriadamif so, I will then merge those (probably tomorrow as it's getting late for me tonight)07:54
chriadamfantastic, thanks very much07:54
dcalisteRebasing went well (some minor conflicts in .pro file and in enum declaration), I'm compiling now.07:55
dcalisteIt will take a bit of time, I've updated the SDK yesterday, so I need to build QMF first.07:55
chriadamah cool, np07:55
chriadamregarding the wiki on tjc, Joona mentioned that he sent you an email with some information about the direction Jolla wants to take for the secrets UI07:57
chriadamnot sure if that information was somewhat orthogonal to what was discussed in the tjc or not07:57
cleeI've gotta say07:58
cleeI just got Sailfish up and running on an XA2 I picked up from Amazon07:59
cleeand it's... surprisingly not bad so far07:59
dcalisteYes, I've read the email. I was waiting for the MER<->SailfishOS merge to happened, to fill up bugs regarding the various issues and improvements he mentioned.07:59
albertuxHi to all...............I'm again afflicted to high drain-battery... on Xperia X... contactsd raise to 70% cpu usage and rss to about 1 Giga (withe crest monitor) any IDEAS??07:59
chriadamclee: great!  if you have suggestions for improvements, don't hesitate to send me an email, and I will make sure they get forwarded to the appropriate people internally.  of course, cannot promise that any specific changes will be made.08:00
cleechriadam: well, I was actually thinking I might like to try writing an app or two08:00
cleeI used to use Qt, back in the KDE1-3 days08:00
chriadamalbertux: I will try to help you08:01
dcalisteI see more the TJC page as a wide brainstorm while the specific points addressed in jpetrell's email are more fitted as bug issue to be tracked separately in my opinion.08:01
chriadamalbertux: basically, contactsd handles a couple of things: 1) xmpp roster synchronisation, 2) sync between the privileged and non-privileged database.08:01
dcalistechriadam: ok, I've pushed the rebased version of the nemo-qml-plugin-email MR.08:02
chriadamalbertux: it may be that a duplication event could have created far too many contacts internally due to a bug08:02
pvuoreladcaliste: hey, wonder if you would mind merging my calligra changes if it's not too much trouble? i've tried to ping #calligra a couple times but no response so far. maybe i'll apply for that developer account myself but to be seen how and how fast that would proceed.08:02
chriadamalbertux: this could cause the various operations which contactsd does, to balloon improperly.08:02
chriadamclee: that'd be great!  have you tried the application developer SDK?08:03
cleechriadam: not yet, I downloaded it and installed it but I haven't started learning everything that's changed since I wrote Qt code08:03
dcalistepvuorela: no problem, can you git format-patch those and put them in the rpm dir of a branch in MER-calligra ?08:03
chriadamclee: Sailfish OS applications use a specific style and component set called Silica: see https://sailfishos.org/wiki/Application_Development and https://sailfishos.org/develop/docs/silica/08:04
pvuoreladcaliste: well they are in phabricator.08:04
albertuxchriadam: ok, I have a multiple account on Conversation (xmpp) it's a possible problem?08:04
dcalisteYeh, but just as patches, without commit comment and the authorship will be wrong ;)08:04
chriadamalbertux: well, it is a potential cause of high cpu usage in contactsd.  but it shouldn't bring the device to its knees - that's most likely indicative of a bug in contactsd or qtcontacts-sqlite backend08:05
pvuoreladcaliste: would that command line tool do the trick?08:05
cleechriadam: sweet, thanks08:05
chriadamalbertux: (basically, depending on network connectivity, you may have multiple things which trigger roster change handling / update)08:06
pvuoreladcaliste: i can also send you formatted pathces of course, though then those won't have the approval metadata things i've seen in git history.08:06
chriadamalbertux: also, if you have duplicated contacts due to bugs in contactsd/qtcontacts-sqlite/sync, then any operation will take significantly longer.08:06
chriadamyou can investigate that with sqlite3 on device, checking /home/nemo/.local/share/system/privileged/Contacts/qtcontacts-sqlite/contactsd.db08:06
chriadamyou can investigate that with sqlite3 on device, checking /home/nemo/.local/share/system/privileged/Contacts/qtcontacts-sqlite/contacts.db08:06
dcalistepvuorela: oh, they are added by hand anyway by the commiter… as far as I know.08:07
pvuoreladcaliste: right, strange workflow.08:07
dcalisteI agree!08:07
pvuoreladcaliste: oh well. maybe i'll email you?08:07
dcalistechriadam, pvuorela: rebase in jolla-email went without conflicts, great.08:08
dcalistepvuorela: ok for email also.08:08
albertuxchriadam: there is also /usr/lib/buteo-plugins-qt5 with 95% CPu Usage....08:08
chriadamalbertux: which process?  msyncd?  or a child sync process in particular?08:08
albertuxchriadam: child08:09
chriadamcan you determine which one?  e.g. is it a google one, or?08:10
dcalistechriadam, pvuorela: ok, I've force pushed the rebased versions of signature MRs in nemo-qml-plugin-email and ui-jolla-email. I'm going to test on device…08:11
albertuxchriadam: no, I can't determine... it is out of my skill08:11
chriadamalbertux: np08:11
chriadamalbertux: you can try "ps axu | grep carddav" / "ps axu | grep caldav" / "ps axu | grep google" to see if it's one of those common ones08:12
dcalistepvuorela, chriadam: ok signature status is still there ;)08:13
chriadamso far so good08:13
dcalistechriadam: I've investigated the TJC issue https://together.jolla.com/question/201611/caldav-calendar-msyncd-fails-when-updating-event-with-exceptions/08:16
dcalisteThe one about broken upsync of recurrent event with exception on a open-xchange server.08:16
chriadamah, yes I remember08:17
chriadamit was different etags reported for exception vs series?08:17
dcalisteFor this etag strange problem, I've opened a bug on the open-xchange bug tracker: https://bugs.open-xchange.com/show_bug.cgi?id=6408608:18
merbotdcaliste: Error: Error getting bugs.open-xchange.com bug #64086: NotPermitted08:19
dcalisteI don't know what will come out of it.08:19
chriadamthanks.  does it only affect openxchange as far as we know?08:19
dcalisteIn the mean time, I've looked at the GNOME implementation and OutlookCaldavSynchronizer one regarding the calendar-query request.08:19
dcalisteYes it affects only open-xchange as far as I know.08:20
dcalisteBut I don't konw how many web services are actually using this backend…08:20
dcalisteFor the request, GNOME and the Outlook stuff are doing the same than us, as far I as understand their code.08:20
dcalisteAnd Sabre is advertizing the same request on their web page: http://sabre.io/dav/building-a-caldav-client/08:21
dcalisteCurrently, I've not much any idea how to nicely workaround this open-xchange issue. But…08:22
dcalistethis problem shed highlight on another issue we have in the plugin.08:22
chriadamI've been using the sabre dav docs for carddav implementation08:23
chriadamagh, I have to run - need to pick gf up from airport08:23
dcalisteThe last commit fixes spurious detection of local changes because of down-synced exception being detected as local modifications.08:23
albertuxchriadam: I will try, but not now...I'm on work I'  will sends you a feedback, thanks, very kind ... ;)08:23
chriadamsorry - will merge those signature PRs tomorrow08:23
chriadamalbertux: no problem, thank you very much08:23
dcaliste;) sure, we can continue discussing this another time.08:23
chriadampvuorela and jpetrell should still be here08:23
* chriadam -> away08:23
*** haasn` is now known as haasn08:30
dcalistepvuorela: ping me when your calligra patches are ready.08:37
pvuoreladcaliste: sure.08:38
pvuoreladcaliste: ping08:53
dcalistepvuorela: got them.08:56
dcalisteI'm going to push them now.08:57
dcalistepvuorela: ok, I've pushed the commits, sorry for the delay, I've been occupied at work.09:46
pvuoreladcaliste: yay, thanks!09:48
pvuoreladcaliste: created mer-core PR.10:24
dcalisteDoes anyone know how to mount a directory in SDK? I've added in the VirtualBox UI a shared folder with read/write access like the home and the config already there. Then in the virtual machine, I've run sudo mount -t vboxsf -o rw,exec,uid=mersdk,gid=mersdk,dev,fmode=0755,dmode=0755 developpement ~/toto13:57
dcalisteBut when trying to create a link (ln -s foo), it is complaining with System Error: Read-only file system13:57
dcalisteWhile, creating files is fine.13:57
*** Milamber_ASF_ is now known as Milamber_ASF16:24
tortoisedochello hou17:46
tortoisedocI am bundling libcontentaction5 with quickbar (as it's not allowed in harbour for some obscure reason)17:46
tortoisedocbut im getting a weird error17:46
tortoisedoc:-1: error: RPM Validation: Libraries: [/usr/share/harbour-debota/lib/libcontentaction5.so.0] Cannot link to shared library: libQt5SystemInfo.so.5 [/usr/share/harbour-debota/lib/libcontentaction5.so.0] Please see our FAQ here: https://harbour.jolla.com/faq#Shared_Libraries17:46
tortoisedoci have set the rpath17:47
tortoisedocbut maybe the name of the bundled so17:47
tortoisedoc no the name is correct17:49
tortoisedoc%define __provides_exclude_from ^%{_datadir}/.*$ %define __requires_exclude ^libcontentaction5.so.0|libQt5SystemInfo.so.5|libQt5SystemInfo.so.5.*$17:50
tortoisedocthis are my spec entries17:50
tortoisedocmal : any suggestion?17:50
maltortoisedoc: I have very little experience with store validator17:53
tortoisedocthat thing really is a blocker :D17:53
tortoisedocmartyone : any help?17:54
tortoisedocit might really be the name..18:21
*** SpeedEvil is now known as Guest2573818:25
*** BitEvil is now known as SpeedEvil19:20
coderustortoisedoc: you need to build qt5systeminfo19:32
coderustortoisedoc: qt5systeminfo not in https://github.com/sailfishos/sdk-harbour-rpmvalidator/blob/harbour-qa/allowed_libraries.conf19:32
tortoisedocmartyone : :-1: error: RPM Validation: Libraries: [/usr/share/harbour-debota/lib/libcontentaction5.so.0] Cannot link to shared library: libQt5SystemInfo.so.5 [/usr/share/harbour-debota/lib/libcontentaction5.so.0] Please see our FAQ here: https://harbour.jolla.com/faq#Shared_Libraries19:59
tortoisedocmartyone : so dependencies of side-shipped libraries are of course restricted too20:00
tortoisedocsounds legit20:02
tortoisedocmartyone : btw my requires_exclude is %define __requires_exclude ^libcontentaction5.so.0|libQt5SystemInfo.*$20:48
tortoisedocwithout the libQt5SystemInfo, the errors about libQt5SystemInfo are there too (not allowed to require)20:49
tortoisedocmartyone : https://pastebin.com/pLAb6P2n20:53
tortoisedocthis is my validator log20:53
tortoisedocI wonder, conisdering I am linking against a library which is OK if whitelisted, the library which links should be whitelisted as well (assuming it's the only linking dependency)20:56
tortoisedocso in this case, as I have whitelisted libQt5SystemInfo (as the tool correctly validated by not displaying the errors), libcontentaction5 should perhaps not have been triggering a linking error (as libQtSystemInfo is, in fact, whitelisted)20:57
tortoisedocbut hey20:57
tortoisedoccant be holier than the pope, right?20:58

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