Can't get output.srt to work?

[Radio-biscuit]

Hello,

I've been trying to get my windows PC to send a stream to my Linux pc using the output.srt function. The application keeps exiting with nothing suspicious in the log, it just shuts down. I feel that I must be overlooking something simple here, but I'm not event getting packets out of the interface (according to wireshark). I've tried all the combinations I can think of. The stream is pcm from a portaudio device and I buffer it enough to stop glitches. I've tested the stream back into another portaudio device directly and it works fine. I can't even get it to work with the host set to localhost. Here's the commands and logs if anyone can help?

i = buffer(buffer = 5., fallible=true, input.portaudio(device_id = 44, buflen=4096 ))
o = output.srt(fallible=true, host="192.168.2.7", %wav(header = false), i)

2023/04/11 15:56:38 >>> LOG START
2023/04/11 15:56:37 [main:3] Liquidsoap 2.1.4
2023/04/11 15:56:37 [main:3] Using: bytes=[distributed with OCaml 4.02 or above] pcre=7.5.0 sedlex=3.0 menhirLib=20210929 curl=0.9.2 uri=4.2.0 memtrace=v0.2.2 mem_usage=0.0.3 dtools=0.4.4 duppy=0.9.2 cry=0.6.6 mm=0.8.1 xmlplaylist=0.1.5 lastfm=0.3.3 ogg=0.7.2 ogg.decoder=0.7.2 opus=0.2.2 opus.decoder=0.2.2 dynlink=[distributed with Ocaml] ffmpeg=1.1.8 ao=0.2.4 samplerate=0.1.6 taglib=0.3.10 ssl=0.5.9 camomile=1.0.2 portaudio=0.2.3 srt.constants=0.3.0 srt.types=0.3.0 srt.stubs=0.3.0 srt.stubs.locked=0.3.0 srt=0.3.0 winsvc=1.0.1
2023/04/11 15:56:37 [clock:3] Using builtin (low-precision) implementation for latency control
2023/04/11 15:56:37 [frame:3] Using 44100Hz audio, 25Hz video, 44100Hz main.
2023/04/11 15:56:37 [frame:3] Video frame size set to: 1280x720
2023/04/11 15:56:37 [frame:3] Frame size must be a multiple of 1764 ticks = 1764 audio samples = 1 video samples.
2023/04/11 15:56:37 [frame:3] Targeting 'frame.duration': 0.04s = 1764 audio samples = 1764 ticks.
2023/04/11 15:56:37 [frame:3] Frames last 0.04s = 1764 audio samples = 1 video samples = 1764 ticks.
2023/04/11 15:56:37 [sandbox:3] Could not find binary bwrap, disabling sandboxing.
2023/04/11 15:56:37 [video.converter:3] Using preferred video converter: ffmpeg.
2023/04/11 15:56:37 [audio.converter:3] Using samplerate converter: libsamplerate.
2023/04/11 15:56:38 [video.text:3] Using native implementation
2023/04/11 15:57:08 [clock:4] Currently 2 clock(s) allocated.
2023/04/11 15:57:08 [clock.main:4] Starting 1 source(s)...
2023/04/11 15:57:08 [source:4] Source srt_0 gets up with content kind: {audio=pcm(stereo),video=none,midi=none}.
2023/04/11 15:57:08 [output.srt:5] Clock is main[].
2023/04/11 15:57:08 [output.srt:5] Content kind: {audio=pcm(stereo),video=none,midi=none}, content type: {audio=pcm(stereo),video=none,midi=none}
2023/04/11 15:57:08 [output.srt:4] Content type is {audio=pcm(stereo),video=none,midi=none}.
2023/04/11 15:57:08 [source:4] Source buffer.producer_0 gets up with content kind: {audio=pcm(stereo),video=none,midi=none}.
2023/04/11 15:57:08 [buffer.producer_0:5] Clock is main[].
2023/04/11 15:57:08 [buffer.producer_0:5] Content kind: {audio=pcm(stereo),video=none,midi=none}, content type: {audio=pcm(stereo),video=none,midi=none}
2023/04/11 15:57:08 [buffer.producer_0:4] Content type is {audio=pcm(stereo),video=none,midi=none}.
2023/04/11 15:57:08 [buffer.producer_0:5] Activations changed: static=[output.srt:output.srt], dynamic=[].
2023/04/11 15:57:08 [output.srt:5] Activations changed: static=[output.srt], dynamic=[].
2023/04/11 15:57:08 [output.srt:5] Enabling caching mode: active source.
2023/04/11 15:57:08 [threads:4] Created thread "clock_main" (1 total).
2023/04/11 15:57:08 [clock.portaudio:4] Starting 2 source(s)...
2023/04/11 15:57:08 [source:4] Source portaudio_in_0 gets up with content kind: {audio=pcm(stereo),video=none,midi=none}.
2023/04/11 15:57:08 [portaudio_in_0:5] Content kind: {audio=pcm(stereo),video=none,midi=none}, content type: {audio=pcm(stereo),video=none,midi=none}
2023/04/11 15:57:08 [clock.main:3] Streaming loop starts in auto-sync mode
2023/04/11 15:57:08 [clock.main:3] Delegating synchronisation to CPU clock
2023/04/11 15:57:08 [portaudio_in_0:5] Activations changed: static=[portaudio_in_0], dynamic=[].
2023/04/11 15:57:08 [portaudio_in_0:5] Enabling caching mode: active source.
2023/04/11 15:57:08 [source:4] Source buffer.consumer_0 gets up with content kind: {audio=pcm(stereo),video=none,midi=none}.
2023/04/11 15:57:08 [buffer.consumer_0:5] Clock is portaudio[].
2023/04/11 15:57:08 [buffer.consumer_0:5] Content kind: {audio=pcm(stereo),video=none,midi=none}, content type: {audio=pcm(stereo),video=none,midi=none}
2023/04/11 15:57:08 [buffer.consumer_0:4] Content type is {audio=pcm(stereo),video=none,midi=none}.
2023/04/11 15:57:08 [portaudio_in_0:5] Activations changed: static=[buffer.consumer_0:buffer.consumer_0, portaudio_in_0], dynamic=[].
2023/04/11 15:57:08 [buffer.consumer_0:5] Activations changed: static=[buffer.consumer_0], dynamic=[].
2023/04/11 15:57:08 [buffer.consumer_0:5] Enabling caching mode: active source.
2023/04/11 15:57:08 [threads:4] Created thread "clock_portaudio" (2 total).
2023/04/11 15:57:08 [clock.portaudio:3] Streaming loop starts in auto-sync mode
2023/04/11 15:57:08 [clock.portaudio:3] Delegating synchronisation to active sources
2023/04/11 15:57:09 [output.srt:3] Connecting to srt://192.168.2.7:8000..

It drops back to the command prompt here. Log level was set to 7.

Thanks for any help!

[toots]

This looks like a crash. We've had report of issues with the windows build of libsrt. The library is actively developed so it's hard to track all the bugs that are fixed on each version.

I have just updated our built images to build against the latest libsrt (version 1.5.1 as of now). You should be able to download the new windows build here: https://github.com/savonet/liquidsoap/releases/tag/rolling-release-v2.2.x

Would you mind giving it a try? This is 2.2.x so your script might need some (minor) adjustments..

[Radio-biscuit]

Yes, many years ago. I'll take a look and see if I can catch what's going on.

Thanks!

[Radio-biscuit]

Okay, got that working (I haven't used MinGW in over a decade I think!). I'm getting this though: liquidsoap-2a3bf62-2.2.0-win64/liquidsoap.exe": not in executable format: File format not recognized

Er, what do I need to do to get a debug version? It doesn't like if I add --debug to the end of the command line?

[Radio-biscuit]

Oh yes, here's the build log:
build.log

[toots]

Ha damn sorry to hear. Did you build the binary yourself? I was hoping we could use the binary from our .zip distribution.

I'll have to see if I can try this inside a virtual machine..

[Radio-biscuit]

Gosh, I can't remember if I did or not. I've been working on both Windows and Linux and I can't remember where I got each version from. Sorry. I'll keep working on my project and come back to this. I'm using VB-Audio VBAN to send between machines at the moment. I'd rather have Liquidsoap though!

[Radio-biscuit]

Sorry for the delay, work has to take precedence unfortunately.

So I managed to run the new 2.2.x build with gdb and I get this:

Starting program: c:\progra~1\liquidsoap-2a3bf62-2.2.0-win64\liquidsoap.exe --interactive SRT_out.liq
[New Thread 9416.0x45d0]
[New Thread 9416.0x51e4]
[New Thread 9416.0x3bfc]
[New Thread 9416.0x2618]
[New Thread 9416.0xdd0]
[New Thread 9416.0x3b8]
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(1) tid(4978) 800401F0 CoInitialize has not been called.
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(2) tid(4978) 800401F0 CoInitialize has not been called.
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(3) tid(4978) 800401F0 CoInitialize has not been called.
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(4) tid(4978) 800401F0 CoInitialize has not been called.
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(5) tid(4978) 800401F0 CoInitialize has not been called.
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(6) tid(4978) 800401F0 CoInitialize has not been called.
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(7) tid(4978) 800401F0 CoInitialize has not been called.
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(8) tid(4978) 800401F0 CoInitialize has not been called.
[New Thread 9416.0x41c8]
[New Thread 9416.0xe0c]
[New Thread 9416.0x2f60]
[Thread 9416.0xe0c exited with code 0]
[New Thread 9416.0x24b8]
[Thread 9416.0x24b8 exited with code 0]
[New Thread 9416.0x4acc]
[Thread 9416.0x4acc exited with code 0]
[New Thread 9416.0x938]
[Thread 9416.0x938 exited with code 0]
[New Thread 9416.0x203c]
[Thread 9416.0x203c exited with code 0]
[New Thread 9416.0x74c]
[Thread 9416.0x74c exited with code 0]
[New Thread 9416.0xcb4]
[Thread 9416.0xcb4 exited with code 0]
[New Thread 9416.0x2e50]
[Thread 9416.0x2e50 exited with code 0]
[New Thread 9416.0x5398]
[Thread 9416.0x5398 exited with code 0]
[New Thread 9416.0x430c]

Welcome to the liquidsoap interactive loop.

You may enter any sequence of expressions, terminated by ";;".
Each input will be fully processed: parsing, type-checking,
evaluation (forces default types), output startup (forces default clock).

Logs can be found in "C:\\Users\\Andy\\AppData\\Local\\Temp\\liquidsoap-596-52202d.log".

[New Thread 9416.0x5158]
[New Thread 9416.0x5104]
# warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(9) tid(5104) 800401F0 CoInitialize has not been called.
[New Thread 9416.0x317c]
[New Thread 9416.0x1e8c]
[New Thread 9416.0x1dbc]
[New Thread 9416.0x2804]
[New Thread 9416.0x3624]
[New Thread 9416.0x1b5c]
warning: onecore\com\combase\objact\objact.cxx(827)\combase.dll!00007FFBB390C189: (caller: 00007FFBB390B4AA) ReturnHr(10) tid(317c) 800401F0 CoInitialize has not been called.
[New Thread 9416.0x2a24]

Thread 23 received signal SIGSEGV, Segmentation fault.
[Switching to Thread 9416.0x1e8c]
0x00007ff65a5430c7 in ctypes_read_pointer (src_=3146264139528) at /home/opam/.opam/4.12.0/.opam-switch/build/ctypes-windows.0.17.1/src/ctypes/type_info_stubs.c:191
191     /home/opam/.opam/4.12.0/.opam-switch/build/ctypes-windows.0.17.1/src/ctypes/type_info_stubs.c: No such file or directory.

Of course that means nothing to me, but I'm sure it means something to you!

This was the script:

log.level.set(7)
log.stdout.set(true)

i=buffer.adaptative(input.portaudio(device_id=44,buflen=4096))

x=output.portaudio(fallible=true,device_id=38,i)

o=output.srt(fallible=true,host="localhost",%wav(stereo=true,channels=2,samplesize=16,header=false),i)

Let me know if there's more information you need?

A.

[toots]

That's great thanks! Any chance you could also get the full backtrace? I think you can do: thread apply all bt after the segfault.

[Radio-biscuit]

Yes, there is a chance .. :) Attached!
backtrace.txt

[toots]

Yes, unfortunately, this does not tell us much more. Still, the segfault comes from ocaml-ctypes and not the underlying libsrt, which is a surprised. I'll have a look at our binding to see if there is anything fishy in there.

[Radio-biscuit]

Ugh. I hate looking for needles in haystacks! Thanks again.