| 18 Apr 2024 |
 Managor | I tried using that. Changing the values had no effect | 23:11:09 |
| Managor | I'll do a quick test by running kodi with startx. On my main user. Maybe it's some xorg setting that's different between users | 23:17:22 |
| Managor | Didn't help | 23:29:40 |
 lrusak | Use xrandr to disable dpms | 23:30:21 |
| Managor | xranr can't detect display. I'll try to edit the x11 configs. | 23:33:08 |
| 19 Apr 2024 |
| Managor | Darn, that didn't help either | 00:08:29 |
| Managor | I made the unfortunate mistake of deleting my previous configs | 00:09:25 |
| Managor | I'll try once without the pulse8 adapter | 00:11:55 |
| Managor | Pure TTY doesn't cause video feed to cut off, so that's a relief | 00:13:14 |
| Managor | So it might have been the pulse8 adapter | 00:20:59 |
| Managor | Oh never mind. | 00:22:19 |
 bgt lover | kodi still segfaults at the end of a playlist for example, it doesn't always start up from the first pressing of the desktop icon, I suspect it's pipewire. What should I do to debug the segfault? | 10:58:53 |
| bgt lover | PIPEWIRE_DEBUG=1 doesn't show anything | 11:05:29 |
| lrusak | bgt lover: you need to post the backtrace | 14:06:24 |
| bgt lover | that's the problem, there's no backtrace I can find, it even says so in the log. Anything I can do to enable one? | 14:07:23 |
| lrusak | Does a crash log exist next to the logfile? | 14:07:55 |
| bgt lover | yes, there is one, but the information is sparse | 14:33:46 |
| bgt lover | I'll upload one | 14:33:47 |
| bgt lover | sorry, apparently it doesn't happen now, since this is nondeterministic. I know of a reliable way to eventually make it happen, and it will do so, but it'll take some time. In any case, the thing doesn't show any backtrace, only the terminal where I run it shows that I got a coredump. | 14:41:59 |
| lrusak | bgt lover: you'll need gdb installed in order to get a useful stack trace, also debug symbol would help | 14:51:58 |
| bgt lover | I have gdb installed, but kodi is running as a flatpak | 15:02:58 |
| bgt lover | should I attach to the process kodi made and do the stack trace that way? | 15:09:58 |
| bgt lover | it happened, again not enough info...should I upload this log file as it is? also, how do I get pipewire debug symbols, or are they included in the arch package? | 15:10:50 |
| bgt lover | ok, attaching to the kodi process while having pipewire access blocks the whole damn sound server. Like, in one instance, I got speech, another...boom | 15:18:33 |
| lrusak | If it's flatpak I don't know | 15:27:21 |
| lrusak | I imagine flatpak has a way to capture crashes | 15:27:44 |
| bgt lover | darn it, more issues, already not looking good...
warning: "target:/app/lib/kodi/kodi.bin": could not open as an executable file:
Operation not permitted.
warning: `target:/app/lib/kodi/kodi.bin': can't open to read symbols: Operation
not permitted.
warning: Could not load vsyscall page because no executable was specified
warning: Target and debugger are in different PID namespaces; thread lists and o
ther data are likely unreliable. Connect to gdbserver inside the container.
0x000061939d1b208a in ?? ()
(gdb) cont
Continuing.
| 15:29:06 |
| bgt lover | and that almost crashed my pipewire too | 15:29:16 |
| bgt lover | ok, got it!
Thread 3 "kodi.bin" received signal SIGSEGV, Segmentation fault.
[Switching to LWP 17726]
0x0000713183e71182 in ?? ()
(gdb)
| 15:31:47 |
| bgt lover | and of course, now it doesn't recognise anything in the flatpak as symbols...wth have I gotten myself into | 15:33:14 |
