| *** n2klb_ is now known as n2klb | 01:55 | |
| sharks_ | It was briefly mentioned a few weeks ago, but is there any way to handle parse-android-dynparts completely agnostic of which slot you're flashed onto? | 03:36 |
|---|---|---|
| sharks_ | https://irclogs.sailfishos.org/logs/%2523sailfishos-porters/2026/%2523sailfishos-porters.2026-06-29.log.html#t2026-06-29T11:02:32 | 03:36 |
| sharks_ | https://irclogs.sailfishos.org/logs/%23sailfishos-porters/2026/%23sailfishos-porters.2026-06-29.log.html#t2026-06-29T11:02:32 | 03:40 |
| sharks_ | Sorry, for some reason sailfish browser turns "%23" into "%2523" | 03:41 |
| Mister_Magister | sailbot: it encodes %, %25 is % encoded | 07:26 |
| Mister_Magister | sorry i meant sharks xd | 08:25 |
| Mister_Magister | poor bot catching strays xd | 08:26 |
| sharks | Lol, yeah no I know it's encoding % as %25 but it seems counter-intuitive given that the % is only there to encode # as %23 | 09:38 |
| Mister_Magister | don't shoot the messenger :P | 09:38 |
| sharks | I did not intend to! | 09:44 |
| Mister_Magister | I went to sleep and now phone is booting on exact same kernel, magic innit | 11:09 |
| Mister_Magister | but seals did not help, heck, browser gets permission denied when trying to use it lol | 11:09 |
| Mister_Magister | okay so in shmem_setattr there is return -EPERM hm | 11:24 |
| Mister_Magister | mal: would you be so kind as to strace sfos browser on fp2? | 11:33 |
| mal | what are you looking for from strace? | 11:44 |
| Mister_Magister | just to compare | 11:45 |
| Mister_Magister | syscalls mostly | 11:45 |
| Mister_Magister | its nice to have working strace to compare against | 11:45 |
| Mister_Magister | xulrunner-qt5-debuginfo-91.13.1+git8-1.25.1.jolla.armv7hl needs 2243MB more space on the / filesystem | 11:47 |
| Mister_Magister | excuse me? HOW MUCH?! | 11:47 |
| Mister_Magister | i don't have that kind of storage | 11:48 |
| Mister_Magister | holdup i only got 200MB free? | 11:48 |
| mal | gecko debug symbols are huge in the old build | 11:52 |
| Mister_Magister | it loops and dies in xulrunner | 11:54 |
| Mister_Magister | but when i strace i don't see the SEAL syscall | 11:54 |
| Mister_Magister | but i see it complaining when i run it without strace | 11:55 |
| Mister_Magister | so after installing debuginfo i tried gdb and it's just hanging there xd | 12:06 |
| Mister_Magister | can it even load such insanely big debuginfo? | 12:06 |
| mal | it can take minutes even on faster devices | 12:10 |
| Mister_Magister | by the looks of it it will take hours | 12:16 |
| Mister_Magister | yeah i don't think I have ram for this | 12:17 |
| mal | that is probably also an issue | 12:18 |
| Mister_Magister | mal: I FIXED IT! | 12:21 |
| Mister_Magister | browser works! | 12:21 |
| Mister_Magister | it was two lines that your fp2 had and I didn't | 12:21 |
| Mister_Magister | https://github.com/mlehtima/android_kernel_fairphone_fp2/blob/hybris-21.08.1-rel/mm/shmem.c#L2441 | 12:22 |
| Mister_Magister | these 2 | 12:22 |
| Mister_Magister | I still can't ssh dunno why (but I can on fresh flash) | 12:22 |
| Mister_Magister | and DNS is still very much not working | 12:23 |
| Mister_Magister | socket(AF_INET, SOCK_DGRAM|SOCK_CLOEXEC|SOCK_NONBLOCK, IPPROTO_IP) = -1 EACCES (Permission denied) is probably the cause | 12:23 |
| Mister_Magister | warning: BFD: error: /usr/lib/debug/usr/lib/xulrunner-qt5-91.13.1/libxul.so-91.13.1+git8-1.25.1.jolla.arm.debug(.debug_loc) is too large (0x178bd5cb bytes) | 13:50 |
| Mister_Magister | you can't even load it | 13:50 |
| Mister_Magister | wait i'm running gdb when i already fixed browser, i'm seriously dumb | 13:51 |
| Mister_Magister | hm, while the patches have fixed browser on x2, on g2 opening browser causes segfualt | 14:02 |
| Mister_Magister | great now on x2 gallery doesn't open anymore even though that was fixed already | 14:05 |
| Mister_Magister | and another segfault in netlink | 14:06 |
| Mister_Magister | lol if i run browser through strace, browser segfaults, if i run it without strace, kernel segfaults | 15:40 |
| Mister_Magister | and the backtrace makes no sense | 16:04 |
| Mister_Magister | apparently one of the lines is if (err <= 0) which makes absolutely 0 sense, theres no funciton call | 16:05 |
| T42 | <elros34> sharks: I didn't test but maybe it's enough to just change this line https://github.com/tchebb/parse-android-dynparts/blob/master/main.cpp#L22 to ReadMetadata(argv[1], 1); for 'b' slot | 16:27 |
| Mister_Magister | cheeky revert of this fixed the issue https://github.com/VerdandiTeam/android_kernel_motorola_msm8226/commit/01526b95dec5f67048afe7b4a5257fe8362d7ed6 | 16:32 |
| Mister_Magister | mal: so I got browser working on 3.4, it's impossibly slow but no matter, but what's the issue is that there's no audio in youtube videos, is that… normal? | 16:36 |
| Mister_Magister | it looks like there's no hw video decoding at all | 16:38 |
| Mister_Magister | like it struggles with 240p youtube when it was doing 1080p just fine | 16:38 |
| Mister_Magister | i see the sink-input in pactl list but it's silent | 16:59 |
| Mister_Magister | but in other apps audio works | 17:01 |
| Mister_Magister | and like there's no audio from browser in both 3.4 devices | 17:01 |
| Mister_Magister | also after I scroll a bit in browser i lose touch | 18:09 |
| Mister_Magister | that normal? | 18:09 |
| Mister_Magister | hmm x2 doesn't experience neither of those issues | 18:12 |
| mal | hw codecs should work as before, anything in logcat? | 18:39 |
| Mister_Magister | nope | 18:40 |
| Mister_Magister | any clue why audio doesn't work on either of the devices? does it work on fp2? | 18:41 |
| Mister_Magister | like audio from browser, like video audio | 18:41 |
| Mister_Magister | on g2 and x2 if i play youtube video there's 0 audio | 18:42 |
| mal | need to test | 18:42 |
| Mister_Magister | mhm | 18:42 |
| Mister_Magister | i still need to fix the socket stuff and memfd stuff and figure out why browser on g2 is impossibly slow | 18:42 |
| Mister_Magister | and why it loses touch | 18:42 |
| Mister_Magister | man 5.1 is really impossible to work with | 18:43 |
| mal | seems to have issues with audio in browser | 18:50 |
| Mister_Magister | Hurray! i mean i'm sorry to hear (pun not intended) that but at least i'm not alone | 18:51 |
| rinigus | @b100dian: the manifest snapshot at github - does it correspond to 5.1 release? then we can tag it as well. for 1iv it does. | 18:53 |
| T42 | <b100dian> yes. that's what I shipped. | 18:55 |
| mal | @b100dian it seems this breaks hw codecs on very old android bases https://github.com/sailfishos/droidmedia/commit/43dfe53a9a398b9e1da651d30034f426d00a48f2 I got errors like "E OMX-VDEC-1080P: ERROR: Buffers requested exceeds max limit 53" | 19:05 |
| mal | I'll probably add some conditional for old bases or build time env var override for the value | 19:05 |
| mal | @b100dian the fallback didn't seem to trigger | 19:06 |
| Mister_Magister | mal: but i did not update droidmedia | 19:06 |
| T42 | <b100dian> Ah. So whats max acquired buffers in thise bases? | 19:06 |
| mal | NUM_BUFFER_SLOTS = 64 | 19:08 |
| mal | but it seems hw codec doesn't like very large values | 19:08 |
| mal | Mister_Magister: do hw codecs work in gallery app on your device | 19:08 |
| Mister_Magister | they seem so yeah | 19:09 |
| Mister_Magister | at least in my app they work | 19:09 |
| Mister_Magister | seems I can't open gallery anymore… again | 19:11 |
| Mister_Magister | ye gallery works, but as i said i didn't update droidmedia | 19:12 |
| Mister_Magister | i have droidmedia from like 2024 | 19:13 |
| T42 | <b100dian> mal: I am surprised the message is not about 32 then.. | 19:13 |
| mal | yeah | 19:13 |
| mal | @b100dian the value in that is the actual buffer count, not the limit | 19:15 |
| mal | but still why does get so many buffers | 19:15 |
| T42 | <b100dian> wait, the split into 2 means that the rest of the buffers can be held by e.g. qtmultimedia? | 19:15 |
| T42 | <b100dian> Does it not start playing back at all or it plays a couple of seconds? | 19:17 |
| mal | failed pretty much right away | 19:22 |
| mal | I think didn't even start | 19:22 |
| mal | after changing it back to 4 it worked fine | 19:22 |
| mal | pactl shows reasonable info when browser is supposed to play audio | 19:38 |
| Mister_Magister | i told you that already | 19:43 |
| Mister_Magister | its almost as if you don't believe me xd | 19:43 |
| mal | I haven't read all the irc logs | 19:44 |
| Mister_Magister | :P | 19:44 |
| mal | I do wonder about these "E: [pulseaudio] client-ext.c: can't obtain command line" | 19:45 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!