fill 0x0000000040232502 0x0 fill 0x0000000040232502 0x0 fill 0x0000000040232502 0x0 fill 0x0000000040232502 0x0 fill 0x0000000040232502 0x0 fill 0x0000000040232502 0x0 fill 0x0000000040232502 0x2 .text._ZN12MethodCallerIM21SerialReadingDelegateFvR6StreamctEE6invokeES2_ct 0x0000000040232504 0x2a out/build/app_app.a(application.o) 0x0000000040232504 _ZN12MethodCallerIM21SerialReadingDelegateFvR6StreamctEE6invokeES2_ct
@hreintke Yes, I do have a serial callback on the Serial. I have been tearing things apart looking for the problem. I was able to get my callback code working fine in basic_serial and http_configNetwork. Now I removed my rboot OTA stuff and things seem more stable...gonna slowly put that back in and see where it breaks.
probably something stupid I'm doing with variables or memory access or something
@kireevco I will try the binary build...yesterday I was trying from the build script
@piperpilot : If you want and can test without the serial callback , you just can remove line 246 from HardwareSerial.cpp. That "just removes the calling of the delegate" but leaves all other in place.
when it jumps into my connectSuccessCallback, if I call any kind of function, it randomly will fail with an IlegalInstructionCause. Examples of functions that will trigger this:
whats really wierd is that the failure doesn't happen inside that callback...it happens in the Serial OnData delegate...but if I remove this callback or take out the function calls from it...everything works fine
Interestingly enough, when I roll back to SDK 1.3...which means rolling develop back to commit 77d3dfdfb68ea35a542ed3628cba75eb175afda1 everything seems to be happy. So either there's an SDK issue that is screwing with the delegates or one of the recent changes to memory management or something else is screwing with things. I'll try to get a simple example together to reproduce the issue and open a bug so that someone that knows more about the guts can take a look at this...been banging my head against the wall for 2 days on this...