Tuesday, 2021-05-04

*** nyov is now known as Guest8094804:00
ViGeThaodan: about that thing which you hate - I copied the idea from one of your PRs from another project ;)04:27
ThaodanViGe: Oh eh well than its my fault and I just fixed that.06:43
ViGeThaodan: How is /usr/include/KF5/KCalendarCore there before it's installed?07:33
ViGeas it's installed by the BuildRequires07:33
ViGehint: It's not07:36
ViGeThaodan: /etc/sailfish-release on the other hand is installed in target images, so it should be safe to use it even before BuildRequires are installed07:37
rinigusThaodan: looking at that kf5calendarcore_found - that is again asking buildreq to be installed before you can determine buildrequirement07:38
rinigusor assuming it is always installed on SDK if available.07:39
ViGeIn any case I think it would be best if piggz started using tbuilder :-P07:39
ThaodanIt should work for sdk07:39
rinigusactually, we should have sailfish_version macros defined for specs07:39
Thaodanmkcal-qt5 should add KF5/KCalCore  to -I anyway07:40
rinigusthen developers could adjust the requirements accordingly.07:40
rinigusThaodan: I don't know the context of it and on why that extra dependency was added. for some reason, piggz decided to add KF5CalendarCore - I guess it wasn't pulled in with mkcal-qt5 in some cases07:42
rinigusre macros: https://forum.sailfishos.org/t/add-sailfishos-and-sailfishos-version-to-development-environment/2979 - I think Thaodan noted that having just `sailfishos_version` is sufficient07:46
ViGeIf I recall correctly, Thaodan tried to get those, but the PR was rejected because reasons07:47
rinigusbut taking into account our expertise in SPECs (quite low for me), _major _minor variants or some demo how to use it in the form 4.0.1 to strip version components would be nice07:47
rinigusViGe: there was something, but I cannot recall much regarding it07:48
rinigusViGe: re tbuilder - piggz tried it last night and it worked without shared_build setting. he has somewhat different filesystems arrangement in his PC and looks like something broke in determination of full path. we didn't debug further as I couldn't reproduce it on my pc08:19
rinigusand in this case, piggz didn't use tbuilder to build dependencies - just for amazfish itself. so, early days with it08:20
ViGeSounds promising08:23
fridlin the last days I have a very increased battery drain on my XA2 at home. As well I have the Problem, that I can't sometimes make or receive calls and SMS. I changed to 2G only for testing. At the Moment Sailfish shows wireless network connection, but i cant receive calls or sms.09:15
fridlcan I somehow debug that? It could be a provider or a system Issue. Deutsche Telekom told me on request there was no change in the 2G/3G architecture recently and they're not aware of a problem.09:16
*** vilpan is now known as Guest2153114:44
*** Guest21531 is now known as vilpan14:44
piggzrinigus: ViGe: Thaodan: tbuilder is -a- solution, but the full solution should be able to build on a remote server, either OBS, or some gitlab/hub ci setup15:22
riniguspiggz: as a replacement for OBS, yes, probably. not sure how to handle CI in this context. ideally, you would rebuild only the packages that have to be rebuilt and avoid rebuilding ones which are ready. which is simple to do with the local repo, as on PC15:34
rinigusmaybe some CI expert can chip in and implement it.15:35
piggzrinigus: for example, when updating a port mw packages, i noramlly just trigger the update on obs adn let it do its thing .... now I would have to dedicate my PC to endless hours of local builds for new versions15:36
riniguspiggz: not sure it is endless hours. you just trigger the build of corresponding package and it will build it as well as all others depending on it. that is for handling packages within one project.15:38
piggzrinigus: pinephone kernel would be endless hours ;)15:38
riniguspiggz: but that is not rebuilt if you change something in libhybris15:39
piggzwould be interesting to see tbuilder build that actually ... it uses lots on obs cross compiler gic15:39
piggzrinigus: no, but it does update regularly, and i just tell obs to do it15:39
riniguspiggz: see results at https://sailfishos-sony-tama.s3-website.pl-waw.scw.cloud/15:39
piggzrinigus: looks good15:40
riniguspiggz: this is sony tama aarch64 port done via tbuilder. droid-hal are just put into RPMS folder and the rest is done automatically15:41
rinigusif I rebuild hal and push new rpms into RPMS/target folder, run tbuilder, it will just rebuild all bits depending on it15:41
AjollaXA2Hello, I found a directory /home/defaultuser/.local/share/system/Contacts/qtcontacts-sqlite/ and a directory .local/share/system/privileged/Contacts/qtcontacts-sqlite/ on my old Xperia X with Sailfish OS v3. I want to transfer the databases directly to my new Xperia XA2 Plus - DoubleSIM, so I copied the directories' contents to the new phone while15:42
AjollaXA2preserving the ownerships.15:42
AjollaXA2But no luck, the 'People' App in Sailfish OS doesn't see the contacts that I intended to transfer. Am I forgetting something? (Clearly I do. :)15:42
rinigusre pinephone kernel: how do you build it now? is there a reason for it to build longer than some android device kernel?15:42
piggzrinigus: obs build take about an hour ...15:43
piggzmy laptop would be similar, 5th gen i515:43
piggzbut, the build uses cross-compilers from obs to do it automatically15:44
piggzthey could probably be installed in a terget tho15:44
riniguspiggz: pinephone kernel you maybe able to offload to ... suse obs. they use aarch64 machines15:45
rinigushmm, they probably even have it available there15:46
Thaodanpiggz: obs is build with a master and some workers as slaves. Ideally we have a master that is the brain and slaves that are the hands of the master.16:17
ThaodanThey key is that building and thinking is separated16:17
ThaodanIn a way it would be possible to build on top of the obs builder and set a master that runs somewhere and can trigger jobs in the ci16:18
ThaodanI don't think githubs ci allows for non js dependencies to be resolved16:19
piggzattah: you use amazfish?19:06
attahpiggz: was just about to poke you :O19:07
piggza saw u mention it on forum :)19:07
attahNo, honestly I don't... but that's more a not being the smartwatch kind of guy than anything19:08
attahI did try it for a few weeks, and i do think it is an important contribution19:08
piggzwhich watch?19:08
attahAmazfit GTS19:09
piggzdecent choice, thats what i have19:09
attahBut i'm way too vain, and not sporty enough (:19:10
piggzheh, i couldnt do without gps tracking19:10
attahAnd i use it like twice a year19:11
piggzits a shame, the one thing i need it to do well, i still struggle with corner case bigs .... time!19:11
piggzthe huami watches take a time, including the TZ offset ... but have no concept of DST19:12
attahSo my Seiko automatic got to come back, although now i'm betraying it witha Braun BN0119:12
piggzso, im always fighting DST bugs19:12
attahBut i guess that can be abstracted with syncing the offset pretty m uch always?19:13
piggzyes, but then weird things happen at dst change, like, syncing data back, but then it gets stuck syncling the same block of data over and over!19:16
piggzbecuase the watches send back activity data as a stream, in blocks, not always the full amount19:16
piggzim currnetly trying to ignore DST, as its not visible on the watch anyway, and converting that to local time in the app as requried19:17
attahOh... so it uses the *resulting* time for things?19:17
attahUTC everywhere is indeed the natural solution19:18
piggzyeah, i was considering removing all tz fromt he watch, and converting what the watch send s back to local time19:19
piggzi already store the db info in utc19:19
attahAnd what it displays for the actual watch time is different still somehow?19:20
piggzyeah, if it was as easy as just telling the watch the time, it would be fine .. but, i have to use the value the watch tells me to gather data back off it19:20
piggzno, not the displayed time, what time it sends back with data19:20
attahso the displayed time is different still somehow?19:21
attahthat's super weird19:21
piggzwell, it doesnt display TZ info at all, so its easy to get the displayed time correct19:21
piggzthe watch just displays what you tell it19:22
attah...but that's separate from the time used for timestamping data, which is what i find odd19:22
piggzbut, i might ask for activity data from 2am on saturday morning ... the watch will say, here is data from 1am up to 2am .... i will store the 2am value as last sync time, then ask again, and the watch will report back for 1am, not 2am19:23
attahdealing with broken protocols is no fun19:25
piggzyeah, if you could just say, this is the utc time, and this is the tz, it would all work, we would both be talkng the same19:26
attahspeaking of weird protocols... i meant to check with you if you'd be interested in pitching in with IEEE11073 libs/support19:27
attahLike for blood glucose meters and all manner of medical end-user devices19:28
piggzin what way?19:28
attahNot sure on how that project would be set up, thus me not having asked yet19:29
attahBut i was thinking your BLE expertise comes in handy19:29
attahas that is one of the more common modes of transport19:29
piggzsure, how does the protocol work19:29
attahSupposedly it is suuuper generic and the initialization sets the reporting format almost arbitrarily with ASN.119:30
attahbut in reality i suspect nobody bothered implementing much beyond the sample/basic format19:31
attahThe specs are paywalled unfortunately, and quite expensive at that... (on top of being quite silly)19:33
attahBut hopefully we would find someone that cares slightly more deeply about this problem domain, and just help out with the technical bits to get it off the ground19:35
attahi.e. a diabetic19:35
piggzthere is HDP in BT GATT, eg https://link.springer.com/chapter/10.1007/978-3-030-51517-1_419:35
piggzso, reading them is dead simple19:36
attah...but are the message formats?19:37
piggzif theyre just gatt characteristics, you just read the value19:38
attahinteresting... that would indeed by quite nice19:38
attahfrom what i found they were huuge async blobs coming in somehow19:39
piggzwell, iee11073 may be different to bt HDP19:39
piggzmaybe i need one of those cheap blood oxygen sensors19:40
attahI took the opportunity to pick a bt enabled blood pressure monitor when i needed to get one19:40
attah...and promptly realized i knew nothing of how to start talking to it, even if i could decipher the messages19:41
piggzis it ble?19:42
attahi believe so, let's double check19:43
attahYes, the manual claims it is19:44
attahOmron X7 for reference19:44
attahso i take it you are not completely against the idea?19:51
attahso do we try and swing some kind of PoC and then go fishing for collaborators maybe?20:14
attahWe should also poke these people https://openrepos.net/content/pisarz1958/glukometr-diabetes-diary20:15
attahBecause it does just that, but specific to glocose meters, and seems to not be getting updates20:15
attahand i think only full-fat BT20:16
piggzargh, trying to fix sync bug (time issue) ... i fixed for europeans, but broke for mysel f¬20:18
attahi know that feeling20:19
attah...all too well20:19
attahat least now i'm mostly out of the printer swamp (:20:20
attahand i must say i admire your perseverance20:21
piggzi can see the problem ... its working out the fix!20:24
attahi need to sleep, gn!20:24
piggzc ya20:24

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