AIS MOB SART Crashes OpenCPN on PI

[John2000]

I have two copies of v5.8.4 running and both receive the same TCP data feed from a network receiver.

Someone appears accidentally to have triggered an AIS MOB SART locally.
On the Windows 10 PC, OpenCPN shows this as expected.

But on the Raspberry PI, OpenCPN crashes every time the SART message is received. I've been watching both running copies, and when the 'message age' on the PC goes to zero, OpenCPN on the PI crashes.

The MoB SART has been running for hours, showing a position on land and this behaviour looks consistent.

OpenCPN on PI doesn't crash if I turn off the comms config for the AIS data feed.

Is this a known issue?

PS - There are no obviously relevant entries in the log file

[John2000]

Update:
As a work-around I added the AIS SART's MMSI to the 'ignore' list and OpenCPN is now not crashing on the Raspberry PI.

[stevead]

Have you configured OpenCPN on the Pi to play a sound when it receives an AIS MOB SART alert?

[John2000]

Yes, it was configured to play a sound, that being the default.
I was not able to uncheck the box to turn off sound in the UI, which seems odd.

Now disabled sounds with
bAIS_GCPA_AlertAudio=0
bAIS_SART_AlertAudio=0
bAIS_DSC_AlertAudio=0

I'll try to test this with a locally generated AIS SART message.

[John2000]

Using python, I sent synthesised AIS SART messages to both copies of OpenCPN.

On the Windows PC this was displayed as expected.
On the Raspberry PI, it crashed immediately. That's with sounds disabled as above.

[stevead]

Issue has already been raised on github.

Whether it gets fixed is another question.

[John2000]

Many thanks. I'm not sure that this is the same bug though.

On Raspberry PI, OpenCPN crashes and deoesn't get as far as playing the wrong sound.

There is no crash using the 'test' button where sounds are setup, but with no speaker on the PI, I can't say further.

[bdbcat]

John2000...
Sounds like you have a new issue.
Issues on github are addressed for new releases, especially if the fault involves a crash.
Please add an issue on github, including information on how to simulate an AIS-SART message for testing. Those things are (thankfully) a little rare IRL.


Thanks
Dave

[stevead]

including information on how to simulate an AIS-SART message for testing

Please don't tell me there are features that are checked in without a unit test? NMEA 183 or NMEA 2000?

[bdbcat]

Steve...
"Please don't tell me there are features that are checked in without a unit test?"
Of course there are. Our basic platform(CPU//OS/Version) grid has about 20 dimensions. Add "n" more dimensions for various I/O devices. And plugins...
There will be corner cases missed in testing.



" NMEA 183 or NMEA 2000?"
Either will do, the issue is not related to the comm channel, I suspect.


Thanks
Dave

[stevead]

Position your vessel in Port Philip Bay, Victoria, Australia. (about 38 16.0 S, 144 40.0E)

Save the following to a file and replay using canplayer.

(1690319715.581237) can0 11F80EC8#801C019221113A88
(1690319715.581283) can0 11F80EC8#81453B56AC7A31E9
(1690319715.581283) can0 11F80EC8#8250B77A66000000
(1690319715.581283) can0 11F80EC8#8300B77A0000CE00
(1690319715.581283) can0 11F80EC8#840AFFFFFFFFFFFF
(1690319715.590039) can0 11FB0AC8#80140E9221113AE0
(1690319715.590338) can0 11FB0AC8#810E014550495242
(1690319715.590338) can0 11FB0AC8#8220414354495645

The last 3 frames correspond to an AIS Broadcast Safety Message, which I don't believe is handled by Timo's library (PGN 129802)