Monday, 2018-10-01

*** ChanServ sets mode: +v T405:58
spiiroinmal, elros1: simply using zero for max "delay" at hal level yields sensorfwd "interval" range of [minDelay,minDelay] -> The way sensorfwd interprets interval requests is: requests that are not in range are ignored -> asking for anything except that device specic minDelay value would be ignored06:15
spiiroini.e. need to provide some kind of sane-enough guess to use if hal does not provide maxDelay06:16
T4<adampigg> spiiroin, /any plans to open up access to fingerprint scannee ;)06:22
spiiroinT4: while developing open component *can* be engineering decision, open sourcing something that started as closed component *is* legal & business thing, i.o.w. i.d.k.06:48
T4<adampigg> hmmm06:51
T4<adampigg> so, what help can be given to porters with such snsors06:52
T4<adampigg> is what im really sure we can organise around a solution06:52
T4<adampigg> spiiroin, btw, t4 is a telegram bridge, @adampigg will alert me ;)06:53
T4<adampigg> spiiroin, even just documenting the architecture and giving us pointer would be good06:54
spiiroinreview/testing appreciated:
spiiroin^ should fix those build failures for devices with older android sensor device api08:41
spiiroin^ mal elros1?08:41
spiiroin@adampigg basically atm: if the prebuild fp sw for xperia x + necessary config changes does not work, then it is going to need sw changes to currently closed sw08:44
r0kk3rzit'd be nice to have fingerprint support on community devices08:46
spiirointo have any changes for that to work: the device needs to have 32 bit fingerprint hal, e.g. must have something at: /system/lib/hw/*finger08:46
spiiroinr0kk3rz: I've not yet seen a device that would have worked out of the box despite having android fp hal/api "supported"08:48
T4<adampigg> +108:49
r0kk3rzspiiroin: understood, but imo thats a main benefit that we provide for you guys08:51
r0kk3rzwe investigate all the weird and wonderful hardware quirks08:51
Mister_Magisterjusa: ping08:58
T4<m_aurel> Interesting discussion about contributing deveopers and devices. We are working on a pre-installed alternative mobile OS. This is important to leverage the market for end users and get the necessary awareness. The challange is to find a mature and stable OS, that supports all the features a consumer expects. Currently we think, that only08:58
T4Sailfish OS and UBports fullfils this requirement, but we think of a step by step migration to Nemo Mobile, startting with a modifierd Glacier UX.08:58
Mister_Magisternemo ftw08:59
T4<adampigg> i have a fingerprint.searchf.so09:01
Mister_Magister@adampigg: i hope it will work on more devices that alien does09:07
T4<adampigg> is searchf a placeholder ... lib64 has that and goodix.so09:14
T4<adampigg> mal, strangely, mido claims to support white/blackboard effects, but they dont seem to do anything11:15
T4<adampigg> there are some additional ones not exposed in qt/gst i will have to try and get to work11:15
T4<adampigg> the other backend i looked at, android, also doesnt support the vendor vairables11:15
Thaodanspiiroin: at least a doc how this cs component works and how to implement a free one would be great.11:18
spiiroinreview/testing appreciated:
spiiroin^ sensor aliasing changes in android headers = breaks compilation for some devices11:49
elros1spiiroin: compile and works good so far
b0n0Guys i took a dmesg of wlan failing to start. Driver seems crashing as i dont get the wlan interface and there are some things like 'wcnss triggered by userspace' error15:18
b0n0Ill upload it, sec..15:18
b0n0Having troubles uploading :/15:22
malb0n0: log has a message "wlan: module is already loaded"15:34
malso are you accidentially trying to load the module a second time15:38
b0n0mal: no15:42
b0n0Thats strange15:42
elros1it can be loaded by systemd-modules-load. IIRC you can blacklist it in /etc/modprobe.d/15:47
malb0n0: are you modprobing it manually or via systemd service?15:47
b0n0mal: systemd service15:56
malb0n0: I saw no wlan related errors in the log15:57
b0n0What about wlan that got triggered by userspace?16:21
malafaik that is what is printed when you modprobe it16:22
malor use the touch command in the service16:23
b0n0Hmm, thats weird, wlan had always worked16:23
b0n0Maybe wifi changes got applied in 2.2.18?16:23
malso you can't see the interface in output of ifconfig -a?16:24
b0n0Only "lo" interface is present16:24
malI highly doubt anything in sailfish side could have caused the issue16:24
maldid you rebuild hybris-hal?16:24
b0n0I repo synced and built yesterday16:25
b0n0Maybe something in lineage side :/16:25
malare you sure you didn't lose any of the local changes you had made to kernel or somewhere else?16:25
elros1b0n0: try rfkill16:26
b0n0Yes, pretty sure16:26
b0n0elros1: why rfkill?16:26
elros1why not?16:26
malyes rfkill is one option but UI should trigger rfkill unblock16:26
malb0n0: so you had made sure kernel is pointing to your own github instead of upstream?16:27
b0n0Yes, theres only my github :/ i tried git config and the branch is correct  as well16:27
malb0n0: that issue just doesn't make any sense16:34
malb0n0: try masking the wlan service and run the commands in it manually16:34
b0n0Ok ill try16:35
piggzr0kk3rz: mal: so, it would seem to me that fb support would be something jolla would want to support on ports....cant understand the current situation of having a closed source bit....17:11
piggzcan someone with an X grep the fpd-slave package and post it to see whats in there?17:12
malelros1: the rotation doesn't work on my device17:13
malelros1: which rotation did your device have in initlal startup wizard after the fix?17:14
malok, I would want to have landscape17:15
malbecause that the logical rotation on my device based on hardware buttons17:15
b0n0I tried that and  wlan still doesnt work. In dmesg it says -Iwlan option in init.qcom.rc is forbidden and if i delete it, wifi still doesnt work17:15
b0n0-iwlan is a wpa_supplicant service param anyway17:16
elros1mal: what about sailfish-browser?17:16
malthat is doesn't weem to have anything to do with your issue17:16
malelros1: I can't get anything on screen if I set HYBRIS_HAL_TRANSFORM_ROT17:17
elros1mal: but did you also set QT_QPA_EGLFS_WIDTH ?17:17
elros1like for native protrait screen17:17
r0kk3rzpiggz: who knows, im sure theres a reason somewhere17:18
elros1mal: to be sure what value you use for HYBRIS_HAL_TRANSFORM_ROT17:18
piggzlikely, but im sure harnessing our collective createiveness we can get a solution17:19
piggzany idea why the qtmultimedia repo on mer git contains ab old copy of the source, plus a submodule with a link to the actual version?17:20
r0kk3rzgit tree?17:21
elros1@piggz: make sure you are on correct branch17:21
malelros1: no help17:21
malelros1: still black screen17:21
elros1mal: you did not use 90 for env?17:22
malI tried both 90 ans 27017:24
mali.e. 4 and 717:24
* vknecht just found looking for ril/qmi answers...17:26
elros1so I have no idea why it works for me17:26
malelros1: I mixed QT_QPA_EGLFS_WIDTH and QT_QPA_EGLFS_PHYSICAL_WIDTH but even after fixing that I only get corrupted content on screen17:29
elros1I did not set PHYSICAL_WIDTH only EGLFS_WIDTH/HEIGHT, I will try it on photon (portrait screen) now to see if there is any difference17:31
malelros1: i meant that I had only set physical size before17:37
*** ChanServ sets mode: +v T417:38
elros1my hwcompositor Version: 1010001 (interpreted as 1010001) if that makes any difference17:48
malelros1: I have 105000117:56
piggzelros1: yeah, wrong branch *holds head in shame*18:09
T4<chronosmsx> where i can find local manifest of officially supported by linageos device?18:10
malyou create it based on the lineage.dependencies file in the device repo, also check the repos listed as dependencies from further dependencies18:12
T4<chronosmsx> what is the base of xml file?18:39
T4<chronosmsx> With hybris dependencies18:39
T4<chronosmsx> I found lineage.dependencies, only one repo18:39
T4<chronosmsx> What another i should include18:40
malit's the same format as the main manifest, here an example for some random device
malyou add only the repos you have as dependencies, you might also need the vendor repo from
T4<chronosmsx> If it's only one dependency, i should write only one <project?18:48
malof course you add also repo that has the dependency file18:48
malat minimum you have two repos, device repo and kernel18:49
T4<chronosmsx> I should also include in append to device repo and kernel Dependencies from lineageos file and proprietary vendor folder?18:52
T4<chronosmsx> Components included in xml file:18:55
T4<chronosmsx> device repo18:55
T4<chronosmsx> kernel repo18:55
T4<chronosmsx> lineageos.dependencies18:55
T4<chronosmsx> and proprietary vendor from TheMuppets's account18:55
r0kk3rzadd those, sync, build18:55
r0kk3rzif it errors, go hunting for more18:56
mal@chronosmsx come on, think a little, I said you check what repos you need from the dependencies file not add the file to local manifest18:57
malhow about you just make a file and show it18:57
T4<chronosmsx> Revision is a branch?19:03
T4<chronosmsx> If submodule that i wrote gas lineage.dependencies file, it will synced automatically? Or i should write them in xml file?19:10
T4<chronosmsx> It's about dependencies19:16
malwhat file is that? the one you just created?19:17
T4<chronosmsx> Yep19:17
r0kk3rzcopy them to your local manifest xml19:17
T4<chronosmsx> Should i include device tree and kernel first?19:17
malorder is irrelevant19:18
T4<chronosmsx> but device tree and kernel is required to be in this file?19:21
malyes, and those are already in the file you gave19:21
T4<chronosmsx> There's common cpu tree19:22
T4<chronosmsx> That's enough?19:22
malor is the device repo missing19:22
T4<chronosmsx> There is device and kernel repo, not included to this file19:23
malthat seems to only have the common device repo, not the actual device repo, so add one more repo19:23
T4<chronosmsx> There is specifical device tree and kernel19:23
T4<chronosmsx> Common repos is dependencies19:23
malthe kernel was already in the file19:24
T4<chronosmsx> Okay, process of download was started19:31
T4<chronosmsx> I really need to fetch subodules?19:33
malsubmodules are not really used in android I think19:34
T4<chronosmsx> Sometimes i gave errors19:34
T4<chronosmsx> error: Server does not allow request for unadvertised object 0e4f36e95456d47bfa46c1557dc0f02a480648ae19:34
T4<chronosmsx> I should create ssh key?19:35
malno idea19:35
T4<chronosmsx> error: Cannot fetch LineageOS/android_libcore19:38
maljust to be sure, did you use correct branch when you did repo init?19:40
kimmoliok. mer hybris got hacktoberfest spam?
kimmoliis this worth of a free t-shirt ?19:41
malnot the first time19:42
kimmoliwas it that should be labeled invalid or something19:42
malor just closing is one option19:49
kimmolithen spammer gets score...19:49
r0kk3rzsurely an unmerged pull doesnt count19:51
r0kk3rz"Tag any spam or irrelevant PRs with the "invalid" label to disqualify them."20:00
r0kk3rzthanks digital ocean, i'll be sure to never buy anything from you20:01
malnot sure if that was related to hacktoberfest or not, I think I saw some odd PRs before also, some months ago20:03
T4<chronosmsx> mal: you should strongly not recommend to newcomers use —depth option20:46
T4<chronosmsx> Without depth option it works very well, with depth - errors20:46
T4<chronosmsx> It in relation with get option, depth can be fetch metadata only, i think it can be MERGE commits20:48
T4<chronosmsx> But idk20:48
malwell you were asking about how much space it needs so I assumed you had limited amount disk space, I usually only tell about the switch if someone asks20:48
piggzmal: running a patched qtmultimedia, and patched gst-droid which allows selecing hdr scene mode20:48
malpiggz: cool20:48
malpiggz: if the scene mode enough or do you have to also set the other values?20:49
T4<chronosmsx> Now i performing actions on server, i have 80GB 8 ram and 8core 2.8ghz20:49
piggzmal: not sure yet ... in jolla-camera app, it seemed to be enough, but that was a hack using quirks ... in my own app, it seems to fail capturing the image, and the scene-mode resets back to auto .... doesnt just happen with hdr, but some like sepia work20:50
r0kk3rzpiggz: nice hack20:53
piggzmal: r0kk3rz: oh nice, fixed my app, it now takes hdr photos and doesnt reset the scene21:04
piggzit was my quirk hack interfering21:04
T4<adampigg> (Photo, 1280x720)
piggzMister_Magister: ^^ mido exclusive :D21:07
r0kk3rzis that just a qtmultimedia camera app?21:07
Mister_Magisteri want that too21:07
piggzr0kk3rz: yeah, just something i knocked up21:07
r0kk3rzyeah cool21:07
Mister_Magisterpiggz: gimme that21:07
piggzthe app is on my github .... but plenty more is needed......21:08
piggzpatched qtmultimedis21:08
piggzpatched gst-droid21:08
Mister_Magisteri want all of them21:08
piggzdroidmedia built with hal1 camera21:08
r0kk3rzfull stack hacks, thats how we roll21:09
piggzi hvent started on detecting what is actually supported yet, its all hardcoded models21:09
r0kk3rzmagic config somewhere maybe21:10
piggzr0kk3rz: no, full stack patches!21:11
piggzmal wont let me use a magic config21:11
r0kk3rzyeah well, thats mal21:11
malpiggz: about the vendor filter, the Qt documentation said QCameraImageProcessing::ColorFilterVendor is the base value, so I understood that the custom values would be larger than that, so for example QCameraImageProcessing::ColorFilterVendorHDR wold be QCameraImageProcessing::ColorFilterVendor+121:16
piggzmal: yeah, that ocurred to me21:16
malso just change the condition to the >=21:17
mal*to be21:17
malah, it's like that already21:17
Mister_Magistergood night everybody piggz r0kk3rz especially mal21:17
T4<chronosmsx> There's not my $DEVICE (in breakfast)21:17
Mister_Magister@chronosmsx DEVICE is what you set it to be so its only your fault21:18
malyou ran breakfast $DEVICE? what did it print21:18
Mister_Magisterconsult yourself21:18
piggzmal: no, youre right, should be > only21:18
malpiggz: maybe even allow the limit value, it doesn't hurt21:19
piggzabout to add the same to the effect mode21:19
mal@chronosmsx if you did the .env file correctly "echo $DEVICE" should print the correct device codename21:19
maland also the -profile files need to be correct21:19
malshow your local manifest21:20
mal@chronosmsx sorry, no need21:20
mal@chronosmsx that says clearly that vendor repo is missing21:21
mal@chronosmsx so how did you add the vendor repo to your local manifest21:21
T4<chronosmsx> I doesn't have any vendor repos in manifest :D21:22
T4<chronosmsx> I should add them21:22
malI said to add the suitable repo from
malonly one should be enough, in your case the samsung repo from there and the local path for that is then vendor/samsung21:23
malpiggz: which effect modes are missing?21:24
piggzmal: emboss,sketch,neon21:27
piggzaccording to my props21:27
piggzunfortunatley, the coder for the effects chose a different style than the exposure coder!21:28
malok, maybe those could be added to Qt as defines and not vendor extensions21:28
malpiggz: how did you handle the gstreamer side of that?21:29
piggzmal: still, theyre not in GST either21:29
piggzmal: sorry, gst side of?21:29
malpiggz: I mean how did you pass the hdr value to gst-droid from Qt?21:30
piggzmal: the value gets passed fromt he plugin, using my hack....21:31
malyes, thought that some kind of hack is needed21:32
malpiggz: other option would be to patch gst we have to include a proper value for that, this is all very hacky now21:33
malwondering if upstream gstreamer would like to add some define for that, then we could backport the patch21:34
malsame with Qt maybe they could add HDR mode to filters, not sure if that is usually considered a filter though21:37
piggzmal: hdr is an exposure, not an effect21:37
piggzit would help if qt and gst terminologu aligned21:38
T4<chronosmsx> Mal: what directory I should specify? vendor/samsung/ha3g?21:38
mal@chronosmsx I said it already, you should read my messages better, vendor/samsung21:39
malpiggz: I think we should ask abranson what he thinks of trying to get those enums to upstream21:40
malboth gstreamer and Qt21:41

Generated by 2.17.1 by Marius Gedminas - find it at!