Sunday, 2020-08-30

Techgreedpiggz need help, could you share latest build for mido, the link on XDA is not working.07:58
riniguspiggz, r0kk3rz : wrote down how to use _service on suse obs (will hopefully get merged into nemo main docs repo)07:59
rinigusexample project
r0kk3rzcool thanks08:00
rinigusin short, we can get _service using upstream services working similar to what we have08:00
r0kk3rzi wonder if we can copypac everything over to suse obs08:13
rinigus_service would not work, unfortunately. those have to be adjusted08:15
rinigusand there is no sailfish over there :)08:15
riniguspiggz r0kk3rz : doc got merged, new link
T42<adampigg> So, what would we be using as an OS ?08:29
rinigus@adampigg: depends on what will be obs in future. iff obs will be provided for sfos and bumped to the latest obs software, that would allow us to convert _service files accordingly08:34
rinigusand a bump in obs version maybe a way to reduce maintenance by keeping it as close as possible to upstream08:35
r0kk3rziirc theres a lot of patches to get sb2 to work08:47
r0kk3rzwhich with newer gcc just might not be required08:47
rinigusr0kk3rz: as far as I have seen, suse obs is using kvm. let's see if we get issues while pushing packages from nemo into it08:55
T42<adampigg> Rinigus: worth posting that info on post as info for possible way forward?09:20
rinigus@adampigg: it is mentioned above in my post already, let's get replies to what we have already...09:32
T42<meierrom> Techgreed: Most maintained ports are now found here:
mal@meierrom about fp3, except for a graphics bug everything works (small part near the bottom of the screen is not visible), I have no definite plans when I will make it public11:48
*** Sellerie7 is now known as Sellerie11:52
T42<NotKit> @rinigus [r0kk3rz: as far as I have seen, suse obs is us …], KVM on aarch64 server?12:15
rinigus@NotKit: can't tell much about it, see line 3 in
T42<meierrom> mal: Sounds great! :)13:03
T42<techone %lastname%> @meierrom thanks15:55
T42<Max %lastname%> Does anyone know when version 3.4 will been published16:10
T42<Max %lastname%> Thanks16:10
riniguslooks like is now limiting number of repos that you can have. just managed to hit max while fixing packaging issues encountered while compiling against fedora_3216:14
rinigusinteresting that few required a fix like - otherwise was getting errors with debug packages having no files16:16
rinigusjolla devs - should I push those fixes as PRs on or should we wait now for the move?16:17
T42<adampigg> @rinigus [looks like is now limiting …], it has been like that for some time, i hit the limit a while back16:17
rinigus@adampigg: wasn't aware of it. strange limit unless the plan was for a while there16:18
malrinigus: fixes to repos are welcome16:26
malrinigus: hmm, will that debug_package disable debug stuff completely?16:28
malin sfos libglibutil has content, does fedora do something differently16:30
malthe debug package of that16:30
rinigusmal: don't know whether it will disable it fully. let me generate the log, so I can share it17:14
malI can make a test build on OBS17:14
rinigusmal: great! I assume on SFOS side with debug_package disable?17:15
malthat branch you linked17:16
rinigusthanks! I just started a build against fedora17:18
rinigususing master branch17:18
malrinigus: that seems to disable debuginfo completely so we can't have it like that on sfos17:18
rinigusmal: ok, then some other solution will be needed. let me get the logs17:19
malthe question is why doesn't fedora generate debug stuff correctly17:19
rinigusmal: good question,log is
rinigusso, I get these empty debugsourcefiles.list for few packages.17:22
rinigusI though it was related to "release" option in make of that SPEC. but if it works on SFOS with the debug symbols, I don't know17:23
rinigusmal: if nothing strikes you immediately, I will ask around17:24
rinigusmal: turns out that debug info is searched in fedora only for libs and exes with executable permissions set.18:20
rinigusso, fix was
deathmistmal: what's up with hybris-17.1? last time I heard there was just one issue left: linking error related to TLS stuffs that get patched during build18:34
malrinigus: hmm, are libs really supposed to be 755?18:37
rinigusmal: I think so. checking on PC seems to be the same, at least most of them18:38
rinigusfedora docs:
malneed to check on sfos18:39
malrinigus: looks like there are some libs with 644 on sfos, most have 75518:44
malrinigus: that seems to be distro-dependent, on ubuntu most are 64418:49
rinigusmal: could be. I have mostly 75518:52
rinigusmal: any idea why mce-headers are not noarch? fedora is again rather upset it cannot find any debug symbols over there :)18:53
malmaybe mistake?18:53
rinigusmaybe, I will submit PR for review then18:55
malrinigus: not sure what defines the permissions usually, I heard that on fedora libs are 64418:55
rinigusmal: let me check. I have it on virtual machine18:56
rinigusmal: no, fedora has 755. I use gentoo and that's 755 too. ubuntu was 64418:58
rinigusmaybe rpm/deb difference?18:58
malwondering about that too18:58
rinigus(just maybe debian made a different choice)18:58
malyeah, seems like it18:59
rinigusmal: suse is also 75519:01
rinigusdon't have access to debian, so cannot check if it is from there19:01
malrinigus: my ubuntu has 64419:03
rinigusactually I have debian x86 in one vm, let me check19:04
malrinigus: I will make an internal bug about that and we'll look into that19:05
rinigusdebian is 64419:05
rinigusmal: I can submit a host of PRs already with 644->755 . or is it early?19:06
rinigusalso, seems like there are more packages with the same issue. can you make internal bug report for making more repos possible for eager bugfixers?19:06
malrinigus: note that some of those packages have also debian packaging19:07
rinigusmal: that may explain it19:07
malthe bug reference is needed anyway for merging (no need for you to know it though)19:08
riniguscoming from early days. but I suspect that with the updates of rpm, we may get the same issues later on sfos19:08
malI found about 20 libs on my device with 64419:09
rinigusYeah, quite a number19:10
malnot much compared to the whole number of libs19:11
malrinigus: you can make PRs, maybe make one first and show it so I can have quick look that it looks good19:28
rinigusmal: submitted libglibutil (permissions fix) and noarch fix for mce-dev20:33
malrinigus: how will that libglibutil now do with debian packaging? one option would be to have a permission override in rules file20:42
malrinigus: or maybe debian build automatically fixes the permissions, would be nice to test somehow20:44
rinigusmal: maybe. I wonder what is an expected permission - canonical for linux - and then set it that way20:44
rinigusdebian could set permissions separately. do we have these packages packaged by someone using debian?20:45
rinigusI'll be off to sleep, let's continue tomorrow20:47
malrinigus: I made a test build in ubuntu and permissions seem ok, based on build log there is a permissions fixing step in the build20:52
piggzmal: rinigus: worried we;re jumping the gun a little as still no direction from Jolla, or, are we trying to push them to new OBS ?20:56
malrinigus: before you make new PRs I need to check one thing tomorrow21:26
T42J %lastname% was added by: J %lastname%22:35
T42<J %lastname%> 724222:36
r0kk3rzpiggz: if its possible to use vanilla obs then that will remove some of the maintenance burden23:25
r0kk3rzwe also might have to migrate all ports from patterns, but thats not so bad23:27

Generated by 2.17.1 by Marius Gedminas - find it at!