Thaodan | voidanix elros34: I reported the issue upstream, my workaround is so far that I build mic without 1e5a09cc51a6ae034b7cedbbf4c1def98dcf5ae6. | 02:58 |
---|---|---|
T42 | <AntonlX> Hello, Does anyone know what line 171 does in the ramdisk initialization script? | 06:53 |
T42 | <AntonlX> https://github.com/mer-hybris/hybris-boot/blob/f0298c45c093a194d0a4301ae3405121501aa3ae/init-script#L171 | 06:54 |
T42 | <AntonlX> I have error on this line. | 06:54 |
T42 | <elros34> nothing itself, just helper function. I guess to make usb initialization looks like android's ini*usb.rc | 07:09 |
mal | @AntonlX that is just wrapper function to simplify writing to files | 10:48 |
mal | hey, there was talk about that suplpatcher some time ago, does that do anything else then just change the url and port of supl server? if so why is that not done via the normal config file for geoclue-providers-hybris | 15:23 |
mal | @elros34 ^ | 15:24 |
mal | maybe it does something else also | 15:25 |
T42 | <Spidey24Z> What about hybris-17.1 ? Could this work also (re @SailfishFreenodeIRCBridgeBot: <mal>https://github....) | 16:40 |
mal | please don't use reply, I can't see what the reply is referring to, Í only see first few letters, does that mean the codec bitness fix, that is in 17.1 already | 16:49 |
mal | that looked this on irc "What about hybris-17.1 ? Could this work also (re @SailfishFreenodeIRCBridgeBot: <mal>https://github....)" | 16:50 |
T42 | <Spidey24Z> @mal this one https://github.com/mer-hybris/hybris-patches/pull/26 | 16:51 |
mal | those patches are in 17.1 branch already | 16:52 |
T42 | <Verevka86> how best to debug video playback through sfos gallery? | 18:23 |
T42 | <Verevka86> in 1 out of 10 attempts the video plays, in the remaining 9 the gallery crashes | 18:24 |
voidanix[m] | did you try `GST_DEBUG=droid*:5 jolla-gallery`? | 18:25 |
voidanix[m] | had the same issue some time ago and downgrading droidmedia fixed it | 18:26 |
T42 | <Verevka86> i will try, thanks | 18:27 |
T42 | <Spidey24Z> i fix this issue by bulding libExynosOMX_Core and libExynosOMX_Resourcemanager from source (re @Verevka86: in 1 out of 10 attem...) | 18:33 |
T42 | <Verevka86> I have a qualcomm (re @Spidey24Z: i fix this issue by ...) | 18:35 |
mal | @Verevka86 which android base? | 18:49 |
T42 | <Verevka86> Lineage18.1 | 18:52 |
T42 | <adampigg> That matches my experience too, with a qcom halium port | 19:00 |
T42 | <elros34> mal: no, AFAIK url/port change is additional manual step but suplpatcher talks to modem | 19:03 |
mal | @elros34 I though url/port is one option and patching something else is another option | 19:20 |
mal | would be nice to know what that suplpatcher does internally | 19:20 |
T42 | <elros34> yes: https://forum.sailfishos.org/t/gps-stopped-working/1181/426 | 19:28 |
mal | maybe I should send a private message on forums asking for sources | 19:40 |
T42 | <adampigg> @Verevka86 what does logcat look like for you when it fails playing back video? | 19:46 |
T42 | <Verevka86> https://paste.ubuntu.com/p/QbVwXHtXhw/ (re @adampigg: @Verevka86 what does...) | 20:10 |
mal | I might have found something related to 17.1 base, let me prepare a patch | 20:22 |
mal | so the patches from yesterday fixed video recording but playback still has issues? | 20:23 |
T42 | <Verevka86> Yes | 20:24 |
mal | hmm, odd | 20:24 |
T42 | <b100dian> Wait, which patches fixed @Verevka86's video rwcording? | 20:30 |
T42 | <Verevka86> ^ (re @SailfishFreenodeIRCBridgeBot: <mal>https://github....) | 20:31 |
T42 | <Verevka86> @b100dian hybris-18.1 only | 20:31 |
T42 | <adampigg> This looks suspiciously like what happens here too (re @Verevka86: https://paste.ubuntu...) | 20:31 |
T42 | <b100dian> @Verevka86 still useful, maybe I should re-base my port:) | 20:34 |
mal | I should merge those | 20:36 |
mal | still would be nice to find out why playback still fails, usually that is the easier part | 20:36 |
poetaster | damn, those pastbins are being cut off from whatever bridde that is. | 20:48 |
poetaster | s/bridde/bridge/ | 20:49 |
T42 | <adampigg> Mal, build on my laptop has caused it to lockup until its done 😂 | 20:50 |
Thaodan | The cut-offs only happen on replies to not dump the whole copy of the quoted message into irc. | 20:52 |
poetaster | Thaodan, well, it's impossible to follow the links. harumph | 20:55 |
poetaster | not that I'd be that usefull, but I can try. | 20:55 |
Thaodan | It's just for context usualy but admit it doesn't help if the original message isn't annymore in your log. | 21:00 |
poetaster | Thaodan, I just didn't scroll up high enough. at it now. thanks! | 21:01 |
poetaster | ohh. eerie: W OMXUtils: do not know color format 0x7f.... what is it about me and color formats. | 21:03 |
mal | @Verevka86 do you have any quirks enabled in gst-droid configs? | 21:05 |
mal | voidanix[m]: to which version of droidmedia? | 21:05 |
poetaster | piggz, volla, logcat, record bust: https://paste.opensuse.org/60661377 | 21:08 |
voidanix[m] | @mal i think https://github.com/sailfishos/droidmedia/commit/41039f332db73d03e5d3d101df9180db68960ff2 | 21:08 |
voidanix[m] | the commit later started the issue AFAIR | 21:09 |
piggz | mal: ive been trying use-codec-supplied-height and width | 21:10 |
mal | piggz: ok, and not difference? | 21:11 |
piggz | crshes with them ... let me try without .. i probably havnt tried that | 21:11 |
mal | piggz: crashes the same way or some different way? | 21:12 |
piggz | the crash is the the bT i sent you the other day .. where i added printf's but it didnt make sense | 21:12 |
mal | ok | 21:13 |
T42 | <Verevka86> mal: this - https://github.com/mer-hybris/droid-hal-configs/blob/master/sparse-11/etc/gst-droid/gstdroidcodec.conf | 21:14 |
mal | ok | 21:14 |
mal | @Verevka86 have you tried without those decoder-quirks? | 21:15 |
T42 | <Verevka86> I'll check now | 21:16 |
piggz | mal: i removed the quirks .. and nothing played, tried 5 or so times ... i added them back, and first run, a recorded video crashed, second it worked, and after that the built-in video also worked | 21:16 |
mal | piggz: but did it crash without those quirks? | 21:17 |
piggz | mal: without, it crashed every time | 21:17 |
voidanix[m] | @piggz did you also try checking out droidmedia at https://github.com/sailfishos/droidmedia/commit/41039f332db73d03e5d3d101df9180db68960ff2? | 21:17 |
piggz | id probably have to try 20 or so times to be sure it _never_ works | 21:17 |
T42 | <Verevka86> mal: I deleted it, the video played once, then it still crashes, but sometimes it plays without deleting, it's definitely not right | 21:24 |
piggz | voidanix[m]: no, i havnt tried that ... which issue is yours, the 17.1 record, or the 18.1 playback? | 21:24 |
mal | ok, then it's not the quirks | 21:24 |
voidanix[m] | @piggz playback on A11 | 21:36 |
mal | voidanix[m]: I have a theory, there could be a version mismatch, if the droidmedia-devel used to build gst-droid doesn't match the actual droidmedia then there could be problems, like the change after the last working commit you mentioned has a api breakage which changes a size of a struct and the memset done in gst-droid then would be using wrong size | 21:39 |
mal | usually when building a release one should be checking that droidmedia used has correct version, comparing it to changelog in release notes is how I do it | 21:42 |
voidanix[m] | interesting, so i assume when 4.5.0 hits this should stop being a problem? | 21:44 |
mal | yes, it will have latest droidmedia-devel package | 21:44 |
mal | @Verevka86 so 18.1 playback issue is likely to be mismatch of droidmedia and droidmedia-devel used to build gst-droid, can you check which droidmedia you have on your device | 21:47 |
T42 | <Verevka86> mal: Latest, gst-droid I also compiled the latest available | 21:57 |
mal | hmm, are you sure you had the correct droidmedia-devel? | 21:58 |
T42 | <Verevka86> not sure which version of droidmedia i should check? | 21:59 |
mal | well you need to figure out which droidmedia-devel was use to build gst-droid, if you built your own gst-droid | 22:01 |
T42 | <elros34> If you do not know then you used wrong version: zypper se -s droidmedia in target should give you the idea | 22:02 |
mal | just to note that when you build droidmedia local build it does not anymore build droidmedia-devel package, that comes from jolla repos, unless you of course build your own droidmedia-devel | 22:02 |
mal | yeah, building custom droidmedia-devel is not automatic, nor is it documented anywhere | 22:03 |
mal | @Verevka86 if you want to make a quick test, build droidmedia-devel from the normal droidmedia repo (using droidmedia-devel.spec) and then install that to your target and build gst-droid again, then install gst-droid on device and see if it works | 22:09 |
budrz | Adapting Google Pixel 3a XL. After running rpm/dhd/helpers/build_packages.sh --mic, couldn't I just use fastboot flash boot /path/to/hybris-boot.img instead of sideloading ? | 22:12 |
mal | hybris-boot.img is just the kernel, rest of the os is in the zip | 22:14 |
T42 | <Verevka86> mal: Thanks, I'll check tomorrow morning, now it's very late. | 22:17 |
T42 | <Verevka86> droidmedia- 0.20220716 | 22:19 |
T42 | <Verevka86> droidmedia-devel- 0.20211101 | 22:19 |
mal | yeah, that difference explains the issue probably | 22:29 |
mal | so two options, downgrading droidmedia to that version mentioned in devel package, or building new droidmedia-devel and rebuilding gst-droid against it | 22:29 |
budrz | mal: I tried sideloading sailfish os with TWRP but I get "signal 11" | 22:36 |
budrz | I intalled 1) stock Android 11 rom, then 2) LOS 18.1 | 22:36 |
mal | ok | 22:39 |
mal | well you can extract the rootfs (the tar.bz2 inside the zip) manually via telnet if you store it in sdcard for example | 22:41 |
mal | then mount it in "telnet 192.168.2.15 23" and extract using correct parameters like this: | 22:41 |
mal | mkdir -p /data/.stowaways/sailfishos | 22:41 |
mal | tar --numeric-owner -xvjf some_name.tar.bz2 -C /data/.stowaways/sailfishos | 22:41 |
budrz | ok, I will try that | 22:54 |
budrz | I'm guessing you can get telnet by flashing hybris-recovery.img somehow? | 23:23 |
budrz | I've never used telnet before | 23:23 |
deathmist1 | budrz: shouldn't even need hybris-recovery, hybris-boot fallbacks to telnet on port 23 if stuff doesn't work out assuming kernel stays booted | 23:39 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!