I recompiled tactics_pi which I forked from your master branch on my Linux Ubuntu 16.04 64-bit Xenial Xerus and with gdb found that crashes were typically caused by division by zero. There was no crashes if no NMEA data (in home) but only when connected to my Miniplex (in boat). I deduced that since Miniplex is sending soooo much data over WiFi, not all TWS etc. variables were available when the new instrument objects already start their calculations. I made a few checks against that on methods which I found caused crashes. Then I recompiled everything also w/ VisualStudio against O 4.6.1 and wxWidgets 3.0.3. This weekend I went sailing for 36 hours with these modifications on board and I did not have a single crash of O on my Windows 10 Surface which I use for navigation (battery consumption rocketed high, though). I also included your fix in tactics_pi.cpp four days ago - now laylines are much better, thank you (Miniplex provides HDT-sentence).
Oh, I got my NKE mast display showing alright the values from your plugin - the trick is to limit with Miniplex the sentences going to NKE TL25 to a strict minimum when one is teaching the NMEA sentences - there is only 20 second window available for that and the display obviously missed the rarely sent sentences from tactics_pi. I had to try a few times to make it work. Once succeeded and during my 36 hour stint the new display functions also worked reliably.
Hope this will help you a bit and I do hope that you keep up the excellent work; I am really impressed with what you have done w/ O and I do appreciate that I do not need to purchase Adrena software - I am a non-professional sailor and that money is better spent elsewhere on my boat!
Hi, Thomas!
Thanks for the excellent plugin - it has helped me in my sailing and in improving my polars. In http://www.cruisersforum.com/forums/f134/tactics-plugin-166909-19.html#post2383419 I was complaining the plugin sometimes making O to get stuck.
I recompiled tactics_pi which I forked from your master branch on my Linux Ubuntu 16.04 64-bit Xenial Xerus and with gdb found that crashes were typically caused by division by zero. There was no crashes if no NMEA data (in home) but only when connected to my Miniplex (in boat). I deduced that since Miniplex is sending soooo much data over WiFi, not all TWS etc. variables were available when the new instrument objects already start their calculations. I made a few checks against that on methods which I found caused crashes. Then I recompiled everything also w/ VisualStudio against O 4.6.1 and wxWidgets 3.0.3. This weekend I went sailing for 36 hours with these modifications on board and I did not have a single crash of O on my Windows 10 Surface which I use for navigation (battery consumption rocketed high, though). I also included your fix in tactics_pi.cpp four days ago - now laylines are much better, thank you (Miniplex provides HDT-sentence).
Oh, I got my NKE mast display showing alright the values from your plugin - the trick is to limit with Miniplex the sentences going to NKE TL25 to a strict minimum when one is teaching the NMEA sentences - there is only 20 second window available for that and the display obviously missed the rarely sent sentences from tactics_pi. I had to try a few times to make it work. Once succeeded and during my 36 hour stint the new display functions also worked reliably.
Hope this will help you a bit and I do hope that you keep up the excellent work; I am really impressed with what you have done w/ O and I do appreciate that I do not need to purchase Adrena software - I am a non-professional sailor and that money is better spent elsewhere on my boat!
Best regards,
Petri