OpenCPN crashes when receiving certain NMEA messages

[rhandersen]

Hello,


I'm experiencing a weird problem, that might be difficult to recreate.


I have made a program written in Python that takes some sensor data (GPS, Radar, AIS, etc.) either directly from the sensors or from recorded text files. The program runs some processing and then sends the processed information to OpenCPN via UDP for visualization. This has worked very well for a while, but recently OpenCPN has started crashing when receiving certain NMEA messages.


If I send the following 3 NMEA messages to OpenCPN, it freezes and I have to restart it:
$HEHDT,90.0,T*16
$HEGGA,000000.00,0.0,N,0.0,E,2,11,0.8,-0.5,M,44.1,M,,*64


So the program just sends this out when no info is available from the sensors. As seen it just sets the own ship to coordinates (0,0) with a heading of 90 degrees.


Here comes the even weirder part:
If I only sends one of those messages there are no problems.

If I send both messages, but change the heading in the HDT message to 89.9 degrees or 90.1 degrees, there are no problems.
It is only when I send both messages and keep the heading at 90 degrees that OpenCPN crashes.


Has anyone experienced something similar or maybe know what could cause this? I'm using OpenCPN 5.2.4.

[Hakan]

@rhandersen
Yes on Win10 I can reproduce that O is hanging when Lat/Lon is exactly 0.0 and HDT is exactly 90.0.
I can't debug it deep enough to find the cause but the same situation would likely not happen in real life. If this seems as a real problem a simple hack would be to change if hdt=90.000 then hdt = hdt + 0.0001. For your tests I'd recommend to use hdt = 90.0001 instead?

[rhandersen]

Thank you. I will change the program to use different default values.

It's interesting that it doesn't work with these exact values, but you're right it's a very rare real life case.

[Hakan]

For you info has Dave now corrected this in the master code for next O release.