| *** ubuntu is now known as Guest51680 | 01:47 | |
| *** Guest51680 is now known as zhxt | 01:59 | |
| *** mgrover is now known as ghosalmartin | 07:27 | |
| r0kk3rz | https://arstechnica.com/gadgets/2018/06/acer-chromebook-tab-10-review-a-good-but-not-great-first-chrome-os-tablet/ | 07:58 |
|---|---|---|
| r0kk3rz | would make a great sailfish os tablet | 07:58 |
| Nokius_ | r0kk3rz: pure port =) | 08:53 |
| Nokius_ | chromeos is a fork of gentoo | 08:53 |
| r0kk3rz | yeah | 08:59 |
| r0kk3rz | it'll be cool if they make more chromeos tablety things | 09:00 |
| T4 | <NotKit> somewhat expensive though, I would prefer something with rk3388 | 09:40 |
| T4 | <NotKit> well, that mystical "OP1" is rk3399 though | 09:41 |
| r0kk3rz | ah rockchip thingie | 09:51 |
| elros | anybody was able to use compass with latest sensorfwd? Magnetometer get values but compass stay at 0 | 10:05 |
| mal | elros: yes, works fine for me | 11:37 |
| mal | elros: is your hwsettings.ini ok? | 11:38 |
| elros | now it works :) need to find out why | 11:40 |
| elros | ah it starts when I have active rotation test page in messwerk | 11:41 |
| mal | elros: I had some issues that it didn't start on the first attempt after boot, after that ir worked fine | 11:46 |
| elros | hmm thats not the case for me. There is no compass data unless I start HybrisOrientationAdaptor (via messwerk) | 11:50 |
| elros | if anyone have idea: sensorfwd log https://pastebin.com/WCnPFrHi | 12:07 |
| mal | what app did you use for compass? | 12:08 |
| elros | it's Orienteering Compass | 12:09 |
| mal | I'll check that log later | 12:11 |
| elros | thx | 12:12 |
| T4 | <adampigg> wondering if i should take a screwdriver to my tbj | 16:02 |
| T4 | <adampigg> now both top and bottom of screen are unresponsive, making landscape usage almost impossible | 16:03 |
| T4 | <adampigg> only thing keeping me sane is that the Home button does exactly that! | 16:04 |
| vknecht | abranson, hello... tried following up on http://merproject.org/logs/%23sailfishos-porters/%23sailfishos-porters.2018-06-18.log.html#t2018-06-18T20:56:44 but I don't get any additional output | 16:05 |
| vknecht | to build "android libs" I re-made libcameraservice_32 libdroidmedia_32 minimediaservice minisfservice libminisf_32 | 16:05 |
| vknecht | then pacakged droidmedia-local and installed it on device with pkcon local-install then reboot | 16:05 |
| vknecht | does that need rebuilding the whole hybris-hal too ? | 16:06 |
| vknecht | (also uncommented the define in other files https://pastebin.com/j5zVrmYM, and rm'ed -rf out/ before rebuilding droidmedia) | 16:10 |
| abranson | vknecht: yes those are the right lines to uncomment. you need files like libstagefright.so and friends. they're carried with the /system, not droidmedia | 16:14 |
| abranson | you can just copy them over directly to the device. the Android.mk files nearby will tell you which targets you need to build for each source file | 16:15 |
| abranson | vknecht: looks like you need to build 'libstagefright' | 16:18 |
| T4 | <adampigg> abranson, how can i help debug color issues with droidmedia | 16:18 |
| abranson | ^ like that | 16:18 |
| T4 | <adampigg> eg, youtube in browser | 16:18 |
| vknecht | abranson, I have out/target/product/idol3/system/lib/libstagefright*, I should copy them in /system on device ? | 16:18 |
| abranson | adampigg: if you turn on debugging in those android files, you'll see a lot more of what's going on with the colour formats etc | 16:19 |
| abranson | vknecht: yes. same place. back the old ones up. | 16:19 |
| abranson | vknecht, adampigg: it sounds like you've got similar problems | 16:19 |
| T4 | <adampigg> also have an issue with no thumbnails being generated, and built in media player not working | 16:20 |
| T4 | <adampigg> ill see what i can do in te coming days, i have a few days off | 16:21 |
| T4 | <adampigg> mal, and compass thoughts? | 16:21 |
| T4 | <adampigg> s/any | 16:21 |
| mal | @adampigg is that all in the logs? btw, about thumbnails, check /tmp permissions | 16:26 |
| T4 | <adampigg> yes, thats all i could see in logcat | 16:27 |
| T4 | <adampigg> mal, abranson, missing thumbnaiils is only for newly recorded videos. if i clear the tracker cache, thumbnails come back, but not for the new videos from the camera | 16:30 |
| T4 | <adampigg> so, possibly same issue as above, jut different symptoms | 16:32 |
| mal | ok, any errors in log from tracker or such | 16:32 |
| T4 | <adampigg> i was just looking | 16:33 |
| T4 | <adampigg> tracker doesnt seem to log to journal... | 16:33 |
| mal | only errors afaik | 16:33 |
| T4 | <adampigg> nothing in logat | 16:35 |
| T4 | <adampigg> ill look later when im more comfortable than on a bus! | 16:35 |
| T4 | <adampigg> ashandy as this tbj is, its not my laptop and a usb cable! | 16:35 |
| vknecht | abranson, here's the new output https://pastebin.com/gELj6gXD ; note that I had more libstagefright_whatever_.so files that what I replaced, here's the list : https://pastebin.com/8NULmUh1 ; also I have this, if it matters : https://github.com/vknecht/droid-config-idol3/blob/master/sparse/etc/gst-droid/gstdroidcodec.conf | 16:41 |
| elros | adampigg: I gave you the link few days ago to solution. iirc ffmpeg cause it, removing it helps | 16:43 |
| abranson | vknecht: lots of codec format in there :D | 16:48 |
| vknecht | abranson, in gstdroidcodec.conf ? | 16:52 |
| abranson | no, in your log. looks like it's being configured as standard NV12, not the usual QCOM_FORMATYUV420PackedSemiPlanar32m | 16:53 |
| abranson | which is odd | 16:54 |
| abranson | but if that's being incorrectly labelled on the way out, then it would cause those colour problems | 16:54 |
| vknecht | abranson, if there's anything I can do to help fixing it (short of crushing the device with a foot)... :-) | 17:01 |
| vknecht | here's a logcat from Webcat playing YT video, which doesn't have color bug : https://pastebin.com/6BtdCb1W | 17:01 |
| vknecht | but it's not the same graphic pipeline or something like that ? | 17:03 |
| piggz | mal: abranson: | 17:38 |
| piggz | Jun 25 18:37:56 Sailfish tracker-extract[5446]: [mov,mp4,m4a,3gp,3g2,mj2 @ 0xe6e00dc0] invalid STSD entries 0 | 17:38 |
| piggz | Jun 25 18:37:56 Sailfish tracker-extract[5446]: [mov,mp4,m4a,3gp,3g2,mj2 @ 0xe6e00dc0] error reading header: -1052488119 | 17:39 |
| piggz | 8 lines like that, and 8 'no thumbnails' | 17:39 |
| piggz | videos dont play in gallery, but do in lls | 17:39 |
| mal | piggz: which device, did videos work before? | 17:44 |
| piggz | mal: yes ... this is since switching from OMX to droidmedia | 17:45 |
| piggz | (ie, the preferred method) | 17:45 |
| mal | piggz: I assume that is mido? | 17:47 |
| mal | piggz: pastebin ls -l /etc/ and ls -l /system/etc/ | 17:48 |
| mal | also ls -l /dev/video* | 17:48 |
| piggz | mal: https://bpaste.net/show/75b56855c90e | 17:49 |
| piggz | i thinks it must be related to recording, as older recorded videos work | 17:50 |
| piggz | so, perhaps different how droidmedia and omx outputted | 17:50 |
| vknecht | piggz, maybe related to gstdroidcodec.conf ? | 17:53 |
| mal | piggz: have you tried videos you haven't recorded yourself, like big bug bunny? | 17:54 |
| vknecht | *buck | 17:55 |
| mal | piggz: just in case add mixer_paths_mtp.xml symlink | 17:55 |
| mal | yep, typo | 17:55 |
| vknecht | :) | 17:56 |
| mal | piggz: if it's recording that fails show logcat when you try to record | 18:00 |
| piggz | mal: recording seems t work, and the recorded video works in lls | 18:09 |
| piggz | mal: somewhere in https://bpaste.net/show/4c34e887177e i recorded for a few seconds | 18:11 |
| potter | hey ther in doc section 4.1 should i need to put those commands on terminal? | 18:52 |
| abranson | piggz: did you have ffmpeg installed? | 19:00 |
| abranson | because if you had gst-libav with ffmpeg, then you might have things like tracker built against that instead of libav. they both provide the same components. | 19:01 |
| abranson | then anything that uses libav would crash, including tracker | 19:01 |
| mal | potter: yes, the ones after HOST $ should be run at terminal, commands after PLATFORM_SDK $ are run in sdk (which is started by sfosdk) | 19:03 |
| potter | ok | 19:03 |
| potter | and the sdk means andriod sdk? | 19:04 |
| piggz | abranson: i uninstalled ffmpeg earliler | 19:05 |
| piggz | but no rebuilds | 19:05 |
| piggz | i have nemo-qml-plugin-thumbnailer-qt5-libav | 19:06 |
| piggz | but not gstreamer1.0-libav instlled | 19:06 |
| mal | potter: no, it's the platform sdk | 19:07 |
| potter | ok | 19:08 |
| abranson | piggz: is tracker your own build or from the standard repos? | 19:09 |
| piggz | abranson: standard afaik | 19:09 |
| piggz | my obs has prefer ffmpe-devel | 19:09 |
| piggz | but that was for the omx / libav stuff earlier | 19:09 |
| abranson | yeah maybe drop that | 19:09 |
| abranson | piggz: where's your obs? | 19:10 |
| piggz | https://build.merproject.org/project/show/nemo:devel:hw:xiaomi:mido | 19:10 |
| piggz | should i delete gst-libav and ffmpeg packages in there | 19:11 |
| abranson | very yes | 19:12 |
| piggz | mal: safe to also delete ngfd-plugin-droid-vibrator | 19:13 |
| piggz | ? | 19:13 |
| mal | piggz: from devel yes if you use native already | 19:15 |
| mal | piggz: not 100% sure how OBS handles if you delete a package from a project, does it delete the rpms, this is only relevant for testing project so it won't break only targets | 19:16 |
| mal | *old targets | 19:16 |
| potter | how can i get ubu-chroot,? | 19:20 |
| piggz | mal: it seems to delete them | 19:22 |
| abranson | potter: install android-tools-hadk | 19:23 |
| potter | ok | 19:26 |
| mal | piggz: ok, so then better just disable the package in testing is you have some old targets that still need it | 19:32 |
| T4 | <abhishek_0> conclusion ... to fix bad thumbnails of the recorded videos and colour issue .. gst-libav and ffmpeg should be removed ? | 19:37 |
| mal | try it | 19:38 |
| piggz | @abhishek_0: we shall find out soon.... | 19:39 |
| piggz | abranson: i cant see it making any difference, zypper dup does not propose any inteesting changes | 19:41 |
| piggz | mal: abranson: ah wait, its fixed | 19:44 |
| piggz | newly recorded videos are working | 19:44 |
| mal | is logcat nicer now? | 19:45 |
| mal | there were some errors there previously | 19:45 |
| piggz | mal: na, it still looks like trash! | 19:46 |
| mal | but at least it works | 19:46 |
| piggz | https://bpaste.net/show/b359fba46811 | 19:47 |
| piggz | yes | 19:47 |
| piggz | very | 19:47 |
| piggz | mal: that just leaves the compass and colorspace issue | 19:49 |
| piggz | but, i think some BLE programmig now | 19:49 |
| abranson | piggz: what changed? | 19:54 |
| piggz | abranson: i think likely the removal of ffmpeg, thats all i can think of ... unless, i havnt tried recording in a few days since making other changes in that area | 19:55 |
| piggz | abranson: if you want to debug the problemati files i can send you one | 19:57 |
| abranson | piggz: maybe it was removing ffmpeg that fixed it, but reinstalling libav. they conflict. | 20:02 |
| abranson | might be that didn't get reinstalled straight away | 20:02 |
| piggz | abranson: i have libav installed, but not the gstreamer plugin for it | 20:03 |
| Mr_Doge | removed gstreamer1.0-libav and ffmpeg, recorded videos now work, videos in 9gag still doesnt play and camera is giving me fc | 20:03 |
| piggz | ah, so that could be it then | 20:03 |
| r0kk3rz[m] | Interesting, dodgy libav codecs | 20:06 |
| abranson | piggz: even though the gstreamer plugin is called gst-libav, it only works with ffmpeg! | 20:10 |
| piggz | sheesh | 20:11 |
| abranson | because libav is a fork of ffmpeg, and ffmpeg used that name before. quite rude really | 20:11 |
| abranson | but really we should move to ffmpeg, because libav has lost popularity lately and a lot of distros have switched back | 20:12 |
| piggz | i knew they were forks, they should make their plugin names match tho! | 20:12 |
| r0kk3rz | gst-libav doesnt have ffmpeg as a dep? | 20:12 |
| abranson | no, it depends on libavcodec etc that both provide, but it will only build against ffmpeg. it checks for special version numbers that only ffmpeg uses. | 20:12 |
| abranson | they're supposed to be compatible, but y'know | 20:13 |
| piggz | well, atm, I have -droid, -plugins-* and that is it. .... is that correct? | 20:13 |
| r0kk3rz | yeah | 20:14 |
| abranson | yep, though r0kk3rz has an interesting build of gst-libav with ffmpeg statically built. that works. | 20:14 |
| r0kk3rz | thats for gst 14.1 though | 20:15 |
| piggz | well, so long as i have basic media playing, i dont really care....it s aphone :D | 20:15 |
| piggz | r0kk3rz: what magic will you add? | 20:15 |
| r0kk3rz | but the codec list in the build output is impressive | 20:15 |
| r0kk3rz | most of them im sure ive never heard of | 20:15 |
| piggz | there was a guy wanting to play .TS files i see.... | 20:15 |
| piggz | and, actually, i already have a .ts file on my phone! | 20:16 |
| abranson | our libav has most stuff disabled because we have to watch out for licensed and patented stuff | 20:16 |
| abranson | but you guys can build whatever you like ;) | 20:16 |
| r0kk3rz | i think the normal gst stuff should do .ts files | 20:16 |
| r0kk3rz | you can always use the new gstreamer, you dont need to wait until the next release | 20:18 |
| abranson | piggz: so where are you up to now? is that colour problem only in the browser? do videos play properly elsewhere? | 20:20 |
| piggz | abranson: yes, jut browser | 20:21 |
| abranson | ok, I have a suspicion for vknecht's problem that a colour format is being selected that isn't completely truthful, and the colour converter is leaving it as COLOR_FormatYUV420SemiPlanar when it should be converting. | 20:23 |
| abranson | because it's actually COLOR_QCOM_FORMATYUV420PackedSemiPlanar32m | 20:23 |
| abranson | question is whether we can get it to select the packed PackedSemiPlanar32m as it should | 20:24 |
| piggz | anything I can help with quickly? | 20:24 |
| piggz | BTW,, do you guys @ jolla test stuff on the jolla tablet? | 20:25 |
| abranson | GST_DEBUG=5 sailfish-browser might tell us more about why it's selecting the wrong one | 20:25 |
| piggz | it also had horrible color space problems in the broser | 20:25 |
| abranson | I have one that was ok last time I tried. | 20:25 |
| abranson | we've had to change quite a lot in the colour conversion lately for different devices with different colour formats and conversion capabilities. | 20:26 |
| piggz | it always works fisrt time .... but then messes up after about the third video, and needs reatrted | 20:26 |
| r0kk3rz | weird | 20:26 |
| abranson | and it's highlighted some inconsistencies in older adaptations | 20:26 |
| abranson | oh yeah that's weird | 20:26 |
| piggz | abranson: woop, i have a new problem | 20:30 |
| piggz | browser crashes when video starts | 20:30 |
| abranson | progress! | 20:30 |
| abranson | did you get any debug before it crashed? | 20:31 |
| piggz | ome moment... | 20:31 |
| piggz | abranson: https://paste.ubuntu.com/p/pFhRRCRf5j/ | 20:43 |
| abranson | hmm, looks like the transformation is being done inside gstreamer because the internal colour converter doesn't recognise the format maybe? doesn't really say why it chose the wrong one though. can you do a log cat at the same time? | 20:52 |
| abranson | i think it's crashing because the colour format is lies | 20:52 |
| abranson | internally gst-droid uses an incorrect format. usually it goes straight to the eglsink so it doesn't matter, and it outputs I420 for things that care | 20:53 |
| piggz | abranson: hmm, omx appears... | 20:54 |
| piggz | abranson: https://bpaste.net/show/f186a3186e2c | 20:55 |
| abranson | omx is the internal android codec interface. gst-omx uses it directly, but it's used by the droid media service too. | 20:55 |
| piggz | k | 20:59 |
| abranson | were there a few video starts in that log? it seems to go through codec setup a lot. doesn't seem to go well though | 20:59 |
| piggz | no, just the one | 20:59 |
| abranson | ah it's kept it all. i guess the 20:54 is the one you did | 21:01 |
| abranson | logcat just dumps the log. it can spool through a lot of previous entries when you launch it | 21:01 |
| piggz | anything helpful? | 21:03 |
| abranson | hmm no nothing jumps out at me :/ | 21:10 |
| piggz | abranson: so, you want me to add debugging to acodec? | 21:16 |
| abranson | that might help, especially if you see the color-format=21 on the output buffer of the codec that vknecht saw | 21:23 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!