Thursday, 2019-10-03

T42<adampigg> Mal, apologies for dissapearing, went to sleep05:49
T42<adampigg> Yeah, the portrait layout needs a little tweaking, but atleast it works05:50
Mister_Magister!seen masha1106:20
merbotMister_Magister: masha11 was last seen in #sailfishos-porters 1 day, 23 hours, 25 minutes, and 6 seconds ago: <masha11> Hello. When I try to call or send SMS, I get the message "no network coverage."The inclusion of mobile data gives a message: "SIM card is not used". Any ideas how to fix this?06:20
Mister_Magisterah it's just slow06:20
Mister_Magistermal: in free time please create motorola:griffin on obs thanks06:24
T42FlorianPropst was added by: FlorianPropst07:25
sledgesPSA: Sailfish OS community collaboration meeting taking place in the internet IRC channel called #mer-meeting in just under 10mins07:52
masha11Mister_Magister: Hello10:06
masha11Mister_Magister: I checked the performance of SIM cards in android. They do not work in LineageOS (no network coverage), they work in CyanogenMod.10:08
masha11However, the sound does not work in SFOS during a call.10:11
Mister_Magistermasha11: finally got you10:11
Mister_Magisteri'm building cm1410:11
Mister_Magistermasha11: do you have audioflinger-glue?10:11
masha11Mister_Magister: yes, I have10:12
Mister_Magistermasha11: have you uploaded repos to github?10:13
masha11Mister_Magister: not yet. I wanted to take everything to the end first.10:14
Mister_Magistermasha11: don't do that! upload asap10:15
Mister_Magisteropensource is there to help you10:15
Mister_Magistermasha11: otherwise i would have to do everything from beginning10:16
Mister_Magisterfor now i'm building cm14 to have same base as source we are using10:16
masha11I'll try. To be honest, for the first time I will upload something to
Mister_Magisterit's pretty simple10:17
Mister_Magisterinstructions are in hadk10:17
masha11I will try to do it today/tomorrow.10:18
Mister_Magisterkk i have to wait for cm anyway10:20
T42<srLuxint> Is anyone familiar with nfcd daemon? I'm trying to grasp the idea of 'target' class and its 'transmit' method in particular. The thing is I'm trying to make some adaptation-specific plugin and I can successfully get the tag and it's content within 'adapter' class, I have 'nfc_adapter_add_tag_t2' method called and then I have this 'tr12:29
T42ansmit' method failing on timeout. What does it represent though? How can I pass my data to 'neard' plugin?12:29
malwhich android base are you using?12:35
T42<srLuxint> Android 6.0.12:38
malmaybe check what was done in
maland implement same things using the hal api you want to use in android side12:39
T42<srLuxint> Yeah, I've seen it. I want to use libnfc from NXP. So binder has no help at all. And I've managed to get it running almost but can't yet get the right sequence for passing the retrieved  NDEF content to system. And I'm sorta stuck with it.12:42
maldo you have the code somewhere?12:44
T42<srLuxint> Not yet I'm afraid. I guess I'll post it later somewhere. The thing is nfcd seems to be written really well in terms of code readability and its whole architecture. Yet not a single piece of documentation to be present.12:50
T42<srLuxint> I'll try then to mimic nfcd behavior from working device based on logs.12:50
T42<adampigg> looks like mesa will build with gcc6 i think13:21
Mister_Magister@adampigg this gcc is faulty still so if any problems you will be back to square one13:22
T42<adampigg> yeah, still worth a shot tho13:22
rinigusadampigg, Mister_Magister: main issue with opt-gcc6 is that it doesn't like to be started by sailfish boosters. other than that, it has been working well for years13:26
rinigus... just start the app without booster and it will work13:26
T42<adampigg> rinigus, its not an app, its whole mesa :)13:27
Mister_Magisterrinigus: well not really13:27
Mister_Magisterrinigus: i've built tdlib using gcc6 all fine tests passed and when you try to compile basic example using it it fails13:27
masha11Mister_Magister: today I don’t have time to commit everything. I will write how to finish.13:28
rinigusMister_Magister: you have to be sure that you link to opt-gcc6 stdc++. for now, you can use static linking with it. let me find an example13:29
Mister_Magisterrinigus: ofcourse i'm doing it13:29
rinigusMister_Magister: like ?13:31
Mister_Magisteri'm using cmake13:31
Mister_Magisterbut yes, like that13:31
Mister_Magisteron second thought i have that in my qt app aswell so yeas like that13:32
rinigusadampigg: shared libs could be tricky. mainly have to ensure that stdc++ is the correct one.13:32
rinigusall I can say that it worked for mapbox gl and, from august, with valhalla.13:33
rinigusthose are libs that refused to work with the gcc available at that time.13:33
rinigusbut proper gcc packaging should be better. and as far as I understood, its on the way13:34
mal@adampigg only thing left in that video gallery PR for cam app is that do we want that much debug logging?14:42
vknechtmal, hello :-) indeed the logging is a remnant of debugging what was going on, guess it's not so useful now (in btnPlayPause at least)14:53
malvknecht: maybe check if some of those are still relevant and remove the rest14:53
mallooks like mostly button clicks which are not really relevant14:54
T42<adampigg> cheers both14:55
T42<adampigg> merge and release later14:56
vknechtok will remove the click logs14:56
T42<adampigg> mal, feel free to raise issues for formatting that bugs you, and anything else14:56
malvknecht: is the filename logging useful?14:56
vknechtguess not14:57
malvknecht: if you want to keep onRecorderStatusChanged logging then make that a bit more clear what it logs14:58
malvknecht: similar to what onImageSaved has14:58
vknechtok will remove it too14:59
malvknecht: might be good idea to check the whole code later for excessive logging, not now though15:00
deathmisthey, this has been a long standing issue: BT status is "tied" to WLAN and instantly breaks when toggling off WLAN, chipset is Qualcomm Atheros, base lineage-15.1 and SOC MSM899815:07
deathmistI also see a lot of "hal_bluetooth_lock" wakeup sources in dmesg too even tho it shows as off in GUI & rkill says BT is soft blocked15:07
T42<adampigg> rinigus:what package contains the stdc++ libs?15:08
rinigusadampigg: it should be with the compiler itself15:10
rinigusAs packaged (wrongly) opt-gcc615:11
T42<adampigg> rinigus, ok, so i need the compiler installed on the runtime system too15:24
maldeathmist: I have never heard that bt and wlan would conflict like that15:27
maldeathmist: which bluebinder version do you have installed?15:27
deathmistmal: 1.0.7-1 and BT related defconfig changed at (some probably unneeded)15:31
rinigusadampigg: it's not really made for it. I don't know how well would libs mix. For apps, we just bundle them. For mesa it would be trickier. You can try to put libs under/usr/lib and see how it will go. But be ready to relay everything. In optgcc package libs are under/opt15:37
rinigus... sorry for missing whitespaces15:38
rinigusadampigg: actually, check how i linked mapbox gl qml plugin. Maybe it's closer to mesa in this respect. I can check in 1-2 hours15:41
T42<adampigg> rinigus, ill be trying shortly15:44
deathmistmal: ok maybe I overexaggerated the "instant" cut-off, but it takes maybe 20 seconds for BT to time out after WLAN is toggled off (tested using BT headset w/ music) here's a journal log after I connected the headset, played some music and toggled WLAN off after a bit:
maldeathmist: what does rfkill list?16:20
deathmistmal: OH I see the issue: there is a seperate "bt_power" thing under rfkill which is seperate from "hci0" and it shows as soft blocked as well as WLAN16:23
deathmistinterestingly enough it doesn't get changed when I use neither the WLAN or BT toggle now16:25
maldeathmist: I think connman messed up things if there are too many rfkills
maldeathmist: try that16:27
deathmistmal: awesome! I'll report back in a bit, thanks!16:29
T42<srLuxint> Up to my NFC issue: couldn't figure out how to pass any data directly to nfcd, but at least I could figure out the exact meaning of the 'transmit' method which appeared to be, which I could never believe (lol), a method to transmit raw data to tags, which also seems to be the only way to pass any data to nfcd. And the libnfc library16:55
T42has the exact same API to transceive raw data. So basically wrapping nfcd's transceive CB with that API worked perfectly fine, at least I've finally succeeded in reading tags content with the system feedback received (i.e. browser opening link).16:55
rinigusadampigg: when you compile and link mesa into .so, add -static-libstdc++ if you use opt-gcc6. this is the way mapbox gl qml plugin is compiled and I hope it will work with Mesa as well. then you shouldn't need c++ libs from /opt/gcc617:01
deathmistmal: now I get a "warning: 'bluetooth_power_rfkill_probe' defined but not used [-Wunused-function] error, forbidden warning: bluetooth-power.c:392" which makes it error :/17:14
deathmistI know the fix is probably "-Wno-error=unused-function" but where exactly do I put it?17:15
maldeathmist: so comment out the function obviously17:30
deathmistmal: I already tried it, when I do it leads to and the mentioned file brings me to , now what?17:37
maldeathmist: add back bluetooth_power_rfkill_probe but make that function dummy, i.e. so it won't do anything17:41
maljust returns success17:41
deathmistmal: no luck, with "static int bluetooth_power_rfkill_probe(struct platform_device *pdev) { return 0; }" it still gives me the same "dev_attr_extldo" error shown in the previous pastebin17:52
vknechtif you really want to add the flag, wouldn't it go here ?
vknecht(not saying it's the thing to do ;-)17:58
vknecht*best thing17:59
malvknecht: did you add back everythin?18:08
maldeathmist: ^18:08
maldeathmist: come on, then you use it in some dummy way18:08
mallike (void)foo18:08
deathmistmal: I "fixed" the build by appending "-Wno-unused-function" to drivers/bluetooth/Makefile's ccflags-y line and btw the patch works! thanks again for it, as I use a BT headset this was a very annoying thing to deal with <319:03
piggzmal: rinigus how can i be sure gcc6 was used for the build?19:03
rinigusmal: its not in the path - its installed at /opt/gcc6. So, you have to configure make system to use /opt/gcc6/bin/g++19:05
piggzu mean piggz ;)19:05
piggzyeah, i know that, and did it, but the log shows useing ccache to build19:05
rinigusindeed - :)19:06
riniguspiggz: is it possible to disable ccache?19:07
piggzits a shame the live log doesnt work, i have to wait 15 minute for the build to finish to see!19:14
riniguspiggz: yep, and for a while already19:17
malI need to complain about that again19:28
malpiggz: you can merge the cam PR if you want19:28
piggzmal: yep, on it19:31
*** electro is now known as Guest7456119:47
vknechtthanks :-)19:58
piggzmal: vknecht: release done....20:32
piggzrinigus: i still cant be sure the right gcc is being used20:32
vknechtpiggz, congrats :-)20:36
vknechtjust sent a translation update, but not a big deal if it has to wait20:39
piggzi think it would work best if i did 0.x.1 versions for transations20:50
malpiggz: probably a good idea21:22
piggzmal: any thoughts on disabling ccache on obs build?21:23
piggzto be sure of compiler used21:23
malpiggz: it seems you messed up tagging somehow21:27
piggztagging of mesa?21:28
malpiggz: cam21:28
malpiggz: 0.5.0 commit is not in master branch21:28
piggzit is now21:30
malthat tag is pointing to some strange commit21:30
malnow it looks ok21:32
T42<Umeaboy> Any new devices that have been unofficially ported but not added to the adaptations wiki?21:36
malmight be21:48

Generated by 2.17.1 by Marius Gedminas - find it at!