RPC on the M4 core needs -fexceptions in cflags.txt and cxxflags.txt files

[hpssjellis]

As far as I know for the M4 core to run RPC commands it needs cflags.tx and cxxflags.txt to have -fexceptions inserted at the bottom of the file.

Is this something that should be inserted by default for everyone? Would you like a PR for it?

[sebromero]

@hpssjellis How do you compile your sketches? I didn't have to change any flags.

[hpssjellis]

Things may have changed. I did it a while ago. Nothing worked on the M4 core, made the changes and things worked. I will remove them and test it. It was mentioned on the arduino forum.

Here is the link https://forum.arduino.cc/index.php?topic=693851.msg4665427#msg4665427

perhaps it was for an earlier version and I just continued to use it with the new version.

[hpssjellis]

@sbhklr I first made sure this program compiled to the M4 core, I then deleted -fexceptions from the 2 files (cflags.txt and cxxflags.txt) and re compiled without even rebooting the Arduino IDE and got this error. Replaced -fexceptions and no error. I hope that is helpful.

arduino-beta\hardware\mbed\1.2.2\libraries\rpclib\src/rpc/msgpack/v1/detail/cpp11_zone.hpp:75:38: error: exception handling disabled, use -fexceptions to enable

                 throw std::bad_alloc();

[sebromero]

@hpssjellis it is, thank you! We'll look into it.

[sebromero]

@hpssjellis It will be solved with the the core update.

[hpssjellis]

@sbhklr Thank you. Looking forward to the next update as I have a lot of examples to re-test. Any idea if we are talking days or month until the next release?

[sebromero]

@hpssjellis We're talking days :-)

[hpssjellis]

@sbhklr Things are working great with the new mbed Version 1.2.3

One minor thing. RPC1.begin() would be more fault tolerant if it could have a serial speed that it then ignores. Presently I am converting Serial.begin(115200); using

#define Serial RPC1 // So the M4 regular serial prints to RPC

but then the 115200 halts the compile.

Probably not needed for this version of mbed, but something to think about for the future.

[sebromero]

@hpssjellis Thanks for the input! We'll consider this for the "virtual serial" implementation for M4.
/cc @facchinm