How many failures? Read bellow:
1. I burned a PIC16F877A (first time trying to program bare-metal) connecting the power pins in a wrong way on a breadboard - I started to hate the breadboards;
2. I bricked an ATmega32 at first bare-metal programing trial. Until then, I enjoyed the demo it came with;
3. I erased the demo from my Nucleo board and failed to flash the new testing program using qstlink2 program, and because of this I entered in panic mode. Every attempt to write the test program using qstlink2 application ended with an orange LED, signaling a communication error.
After many hours of frustration, I found a new program for linux, stlink on github, and I was able to flash testing programs created with STM32CubeMX, OpenSTM32 and ARMmbed.org. What didn't worked at all? The "drag and drop" programming. But as I said, all "mbed" applications created on-line can be programed with the stlink program.
Over the years, I avoided using 3.3V microcontrollers (actually I refused to take them into consideration) as I knew that 5V microcontrollers are much more immune to noise in the industrial environment. But seeing the (European) Nucleo boards, I decided that is time to start this new adventure of setting a development tool chain, acquiring 3.3V peripherals and learning a new API SDK.
But I assure you, setting a free development environment under Linux for this board is much harder than just installing Arduino IDE and only a young engineer can be motivated enough to endure all the hassle. I needed almost half a day to solve all the issues, others may be lucky, but this is just the beginning of the journey. It is worth it? This board and the microcontroller equipped attracts me more than a ChipKit Max32 board based on PIC32 microcontroller.
So I will try to create a folder on my github account where I will store successful OpenSTM32 (AC6) projects for this board.