| rydare | tom | 01:23 |
|---|---|---|
| rydare | oko | 01:23 |
| rydare | a | 01:23 |
| rydare | best | 01:23 |
| rydare | where r u | 01:23 |
| rydare | ooops sorry wrong channel | 01:24 |
| *** ChanServ sets mode: +v T4 | 02:53 | |
| MeowDude | I forgot how to install packages straight to the target in the platform sdk. sb2 something -m I think. Checking faq now but if someone is online and happens to know, that'd be great! | 04:52 |
| MeowDude | okay nevermind, found it surprisingly fast. for anyone looking through the logs: "sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -m sdk-install XXX" | 04:53 |
| T4 | <samyy_010101 (MISSING @USERNAME! telegram.org/faq#usernames-and-t-me)> @Zosenko [Hi everybody, one question, which cloud for a …], nextcloud at your home..m | 06:29 |
| MeowDude | well a 3.0 build has been made for hammerhead. If only my new Nexus 5 were in the mail I could personally test it. | 06:39 |
| MeowDude | mal: any news on that port for g4 play/harpia? A guy on xda pm'ed me interested in building it themselves and I said it was a private build by someone else (I briefly mentioned it). How bad were the camera issues? | 07:11 |
| T4 | <adampigg> Maybe a Xiaomi mipad 2 cherrytrail could replace my tbj | 07:53 |
| T4 | <eugenio_g7> I have one, haven't managed to get libhybris working yet | 08:00 |
| T4 | <adampigg> Maybe not then! | 08:01 |
| T4 | <eugenio_g7> Heh :D I guess the Mi Pad 4 would be perfect, pretty cheap and with a snapdragon. Too bad Xiaomi still hasn't released the kernel sources | 08:02 |
| T4 | <adampigg> Maybe aigo | 08:41 |
| T4 | <adampigg> Just a little pricey for older hardware | 08:41 |
| T4 | <adampigg> But at least will be compatible | 08:42 |
| *** ChanServ sets mode: +v T4 | 08:42 | |
| T4 | <eugenio_g7> Yes, tbj firmware should work fine on the aigo | 08:48 |
| r0kk3rz | geez piggs, and you just spent all that effort on the tablet too | 09:02 |
| r0kk3rz | scorpion tablet is nice ;) | 09:03 |
| T4 | <adampigg> R0kk3rz I know :( 😭 | 09:14 |
| *** ChanServ sets mode: +v T4 | 14:37 | |
| r0kk3rz | https://git.merproject.org/mer-core/nfcd | 17:46 |
| r0kk3rz | i wonder if it uses kernel or libnfc | 17:49 |
| rinigus | looks like cpu sleep is significantly longer for 3.0.0.8 than it was before (https://gist.github.com/rinigus/15d5d8b23b8738f1d80a63f81a86f33e ). I suspect it may result a missing alarm clock (have to check). have seen that emails were not checked during a day, probably due to longer sleep than deltaT specified for emails (5 min in my case) | 18:01 |
| rinigus | has anyone observed something similar or is it limited to onyx? | 18:02 |
| T4 | <Zosenko> Juste one thing, is it possible add a alien dalvik into one port of SFOS3? | 18:11 |
| T4 | <benzandsBot> yes, but no :D | 18:15 |
| T4 | <benzandsBot> you *might* be able to do that by using files from official jolla devices/ports, but it's not really legal | 18:15 |
| T4 | <benzandsBot> instead we hope that Anbox will work properly on sfos | 18:16 |
| T4 | <Donreddy> May u guys help me to get sfos in oneplus3 | 18:23 |
| T4 | <Donreddy> Where to get this | 18:23 |
| T4 | <benzandsBot> only if you already read HADK and followed all steps in it :) | 18:23 |
| T4 | <Donreddy> Idk where it is.. Hadk ? | 18:24 |
| T4 | <benzandsBot> @Donreddy [Idk where it is.. Hadk ?], https://sailfishos.org/wp-content/uploads/2018/08/SailfishOS-HardwareAdaptationDevelopmentKit-2.2.0.2.pdf | 18:24 |
| T4 | <elros34> what? https://forum.xda-developers.com/oneplus-3t/development/sailfish-os-oneplus-3-3t-t3724264 | 18:24 |
| T4 | <Donreddy> @benzandsBot [https://sailfishos.org/wp-content/uploads/2018 …], Thanks sir. | 18:25 |
| T4 | <Donreddy> @elros34 [what? https://forum.xda-developers.com/oneplus …], Thanks but it is 2.* sfos is showing 3.0 in there site. | 18:25 |
| T4 | <benzandsBot> @Donreddy [Thanks but it is 2.* sfos is showing 3.0 in th …], you can update to sfos 3 | 18:26 |
| T4 | <benzandsBot> manually | 18:26 |
| T4 | <Donreddy> Ok thanks for clarification. | 18:26 |
| r0kk3rz | rinigus: im sure my jc missed an alarm the other day | 18:29 |
| rinigus | r0kk3rz: I have seen missed alarms earlier on onyx as well, before 3.0.0.8. but these sleep times seem to be excessive, much longer than before. | 18:30 |
| rinigus | ... so I wonder if something is wrong with the port or its common for sfos | 18:31 |
| mal | rinigus: r0kk3rz I have seen random alarm issues on fp2 at least, maybe c also | 18:44 |
| rinigus | mal: did it increase with 3.x? | 18:46 |
| mal | rinigus: I haven't used alarms on 3.0 fp2 | 18:48 |
| T4 | <adampigg> Mal, tried new camera stuff? Abranson has had a go today | 20:02 |
| mal | @adampigg I tried those yesterday | 20:03 |
| T4 | <adampigg> Good,.work ok? I dont think asd is working, havnt cheched why yet | 20:03 |
| T4 | <adampigg> Also, in hal3, redeye IS a flash mode, so works for abranson | 20:04 |
| abranson | I think it might vary by device. I think on the Jolla1 it was too. | 20:35 |
| mal | abranson: should we add some quirk for that | 20:39 |
| *** ChanServ sets mode: +v T4 | 20:39 | |
| abranson | mal: yeah. should use the existing hdr quirk too | 20:39 |
| mal | abranson: do we need the hdr quirk, or does hdr scene mode set it automatically | 20:43 |
| abranson | mal: would have experiment | 20:44 |
| abranson | I noticed that some devices do asd with a 'scene-detect' param too | 20:45 |
| abranson | also very interestingly, hdr scene mode is failing on the xa2 in hal3 mode, because there's a raw_image_notify callback coming through that droidmedia doesn't handle. | 20:54 |
| piggz | abranson: handle it then ;) | 20:55 |
| piggz | error on this line https://git.merproject.org/mer-core/qtmultimedia/merge_requests/19/diffs#867d87c02b9fb577614781b0e2c319b268e31673_50_69 | 20:55 |
| abranson | piggz: easier said that done... | 20:55 |
| mal | piggz: is it ok that two values are mapped as auto? | 20:59 |
| piggz | mal: its an error, shoud be ASD | 20:59 |
| abranson | i'm a bit confused as to the differences between auto, manual and asd. seems like there should only be two of those. | 21:01 |
| mal | abranson: I thought we checked earlier that auto is not auto | 21:03 |
| abranson | yeah last I gathered 'auto' should map to manual i.e. no scene mode, and 'asd' is the real auto | 21:04 |
| abranson | not sure though | 21:04 |
| abranson | but that would solve that problem | 21:04 |
| mal | piggz: https://github.com/LineageOS/android_hardware_qcom_camera/blob/cm-14.1/QCamera2/HAL/QCameraParameters.cpp#L684 | 21:11 |
| mal | piggz: and https://github.com/LineageOS/android_hardware_qcom_camera/blob/cm-14.1/QCamera2/HAL/QCameraParameters.cpp#L669 | 21:11 |
| abranson | mal: yeah that's pretty convincing | 21:13 |
| mal | abranson: so now we have a useless enum value in gst-plugins-bad | 21:16 |
| abranson | mal: good job it's only just gone in :P | 21:17 |
| mal | abranson: I should have checked that before it went in | 21:18 |
| mal | abranson: but we did talk about earlier already | 21:18 |
| mal | abranson: I just have been so tired that I didn't pay attention to the new code in those | 21:20 |
| T4 | <akaWolf> We're all just a humans | 21:28 |
| piggz | abranson: mal: so, in that case, ExposureASD can be removed? | 21:35 |
| mal | piggz: in my opinion yes | 21:41 |
| abranson | Yeah I don't think it's a big deal to remove it. | 21:44 |
| abranson | These changes aren't straightforward. I probably shouldn't have merged it so quickly, but I thought it was a lot simpler than it turned out to be. | 21:45 |
| piggz | so, in gst-droid, map gst_manual to "auto" and gst_auto to "asd" .... and in qt, ExposureManual = gst_manual, and ExposureAuto to gst_auto | 21:46 |
| mal | yes | 21:46 |
| mal | no idea why no scene mode is named auto in android | 21:47 |
| abranson | So the default setting should be manual then, because not all devices have the asd mode | 21:47 |
| mal | abranson: don't we use auto now as default? | 21:48 |
| abranson | We do | 21:48 |
| mal | then we need to change the default | 21:48 |
| mal | if we remap those | 21:48 |
| piggz | does the current camera even use those modes? | 21:49 |
| mal | piggz: maybe directly? | 21:49 |
| piggz | well, the mode would still be 'auto' in android land, but be called manual instad of Auto ... which is a current bug | 21:50 |
| mal | piggz: even reference says it mean off https://developer.android.com/reference/android/hardware/Camera.Parameters#SCENE_MODE_AUTO | 21:52 |
| mal | piggz: I think it's auto in the sense that paramaters are completely free to change | 21:52 |
| mal | piggz: whereas asd selects which scene mode to use automatically | 21:53 |
| mal | piggz: so is manual the correct name for that is a good question | 21:53 |
| piggz | well..... | 21:54 |
| piggz | decide befoew I change it agian! | 21:54 |
| piggz | :D | 21:54 |
| mal | abranson: any comment? | 21:55 |
| piggz | mal: in the gui, ExposureManual string could be "Off" | 21:57 |
| mal | piggz: I think yes | 21:57 |
| mal | piggz: gst does say GST_PHOTOGRAPHY_SCENE_MODE_AUTO: Choose scene mode automatically | 22:00 |
| mal | piggz: so that matches asd | 22:01 |
| mal | so GST_PHOTOGRAPHY_SCENE_MODE_MANUAL: Set individual options manually | 22:01 |
| mal | maybe that is logical then for android "auto" mode | 22:02 |
| piggz | yes | 22:03 |
| piggz | mal: see the specs of this ... any reason why libhybris shouldnt work? https://www.gsmarena.com/xiaomi_mi_pad_2-7770.php | 22:06 |
| piggz | intel aton + gfx | 22:07 |
| mal | piggz: does it have lineage? of course with atom you could do native graphics | 22:08 |
| piggz | mal: https://github.com/latte-dev | 22:10 |
| piggz | unofficial 13 https://forum.xda-developers.com/mi-pad/development/unofficial-lineageos-13-0-xiaomi-mi-pad-t3565760 | 22:10 |
| piggz | mal: how does native work? | 22:10 |
| T4 | <eugenio_g7> there is CM12 and LOS13 for the mi pad 2, but they don't work fully | 22:10 |
| T4 | <eugenio_g7> I haven't managed to get libhybris working yet on the mipad 2 (both on CM12 and LOS13), now trying the mesa route | 22:11 |
| mal | piggz: assuming the kernel has suitable drivers it could be possible | 22:11 |
| piggz | well ... if native was possible, and the pad was cheap enough, may be worth a shot | 22:11 |
| mal | @eugenio_g7 how does libhybris fail? | 22:11 |
| T4 | <eugenio_g7> segmentation fault when loading the hwcomposer plugin, but it's been a while :) I should give that a shot again (and perhaps ask for help here, since I mostly lurk) | 22:13 |
| piggz | well, there is one on ebay atm | 22:13 |
| T4 | <eugenio_g7> I guess TheKit was trying to make libhybris work on cherry trail tables, but not sure how far he has got | 22:14 |
| T4 | <eugenio_g7> piggz: hardware wise is nice, perhaps a bit warm, but I guess that's the norm with atoms | 22:15 |
| T4 | <eugenio_g7> screen is pretty good though | 22:15 |
| piggz | @eugenio_g7 it unlocks ok, bootloader etc? | 22:16 |
| T4 | <eugenio_g7> yup, it always asks for confirmation (boot / go to fastboot) when the bootloader is unlocked | 22:18 |
| T4 | <eugenio_g7> but nevertheless it unlocks ok :D You don't even have to ask xiaomi for the unlock code like it happens for their other devices | 22:19 |
| piggz | might give it a go | 22:19 |
| TheKit | piggz, Intel gfx blobs have some unpleasant thing with inline TLS access code in them, but could be worked around in theory | 22:20 |
| T4 | <eugenio_g7> piggz: I'd say that if you find one < 100€ it would be a good buy | 22:20 |
| T4 | <eugenio_g7> otherwise I'd go for a Mi Pad 4, but only when (if?) xiaomi releases the kernel sources | 22:21 |
| piggz | @eugenio_g7 k....i wont be spending loads ... otherwise id get an aigo! | 22:22 |
| T4 | <eugenio_g7> TheKit: have you got mesa working on cherry trail tablets? | 22:22 |
| TheKit | yes, MESA is not an issue with them | 22:23 |
| T4 | <eugenio_g7> do you remember which mesa version worked? 9.2.5 (the one in mer) doesn't support cherryview, I'm trying to upgrade it to 10.2.9 which should work in theory | 22:25 |
| piggz | mal: how does a native port work? | 22:26 |
| piggz | any guide? | 22:26 |
| T4 | <eugenio_g7> piggz: the cheaper mi pad 4 is 153€ which I'd say it's a pretty good price | 22:28 |
| piggz | hmm, yes, only 134 on gearbest | 22:34 |
| piggz | but no lineageos | 22:34 |
| T4 | <eugenio_g7> no kernel sources yet :/ | 22:35 |
| mal | piggz: ask locusf, I think he did something for raspberry pi | 22:35 |
| piggz | mal: abranson: qtmm pushed with ASD removed, and #if guards for new enums | 22:36 |
| piggz | mal: abranson: and gst-droid pushed | 22:45 |
| piggz | mal: and before moaning any firher about my code, just look at this... | 22:48 |
| piggz | https://git.merproject.org/mer-core/qtmultimedia/merge_requests/19/diffs#c0bce27e58e4d9b97ee5525edce2a5b059e26928_119_150 | 22:48 |
| piggz | that is, a switch, on a QFlag ... with cases being the QEnum the flag is based on... | 22:49 |
| piggz | i think its a typo in the api, and the enum should be passed into the function, not a flag of the possible enums | 22:49 |
| piggz | !!! | 22:49 |
| merbot | piggz: Error: "!!" is not a valid command. | 22:49 |
| piggz | in anycase, you cant switch on a qflag ... it doenst make sense :D | 22:49 |
| TheKit | @eugenio_g7 mesa-12.0.1 | 22:52 |
| TheKit | I didn't try 10.2.9 | 22:52 |
| T4 | <eugenio_g7> thanks! | 22:52 |
| T4 | <eugenio_g7> do you have the spec at hand by any chance? | 22:52 |
| piggz | well, if i pick up a tab for 50 quid, i'll chip in with the effort ;) | 22:53 |
| mal | @eugenio_g7 there is also the surfaceflinger plugin as an option to hwcomposer plugin | 22:53 |
| mal | @eugenio_g7 some devices have buggy hwcomposer and surfaceflinger method works better | 22:54 |
| T4 | <eugenio_g7> yup, tried that too but no dice. Sorry for being vague but it was a while ago :) But I guess it's worth a try again | 22:54 |
| mal | piggz: how does that switch even work? | 23:14 |
| mal | piggz: maybe some suitable overloaded operators in QFlag | 23:17 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!