Sometimes when running Resilio Sync on Windows, one of the machines may get stuck for a long time at a certain percentage, with no clear reason why it fails to complete.
For some reason, the Virtual Com Port (VCP) drivers for STM32 chips were acting up and not properly enumerating a usable COM port on Windows when I was testing an STM32L476 chip.
I’m not sure if this is a chip-specific issue or if this problem shows up on more chips using the STM32Cube Hardware Abstraction Layer.
(Even after I installed Atollic TrueSTUDIO and the older STSW-LINK004 Windows utility, which I like better than the new STM32CubeProgrammer because it doesn’t require Java.)
Only the AVR series chips from the Atmel acquisition have this support, when using the open source GCC suite. And even then, their official compiler build is pretty old.
A quick note: I finally got around to updating mbed-memory-status to support outputting data using the Serial Wire Output port on a Nordic nRF52 Development Kit.
The trick to getting it work was, as usual with Nordic stuff, a compiler #define (-DENABLE_SWO to be exact). Took a bit of digging to figure this one out and get it working.
This actually lead to some further investigation and the interesting discovery that SWO could possibly offer an extra output-only serial port on most Cortex-M3-and-up chips. But, there are still some blockers on getting that thought working and I haven’t had the time to clear them.
An oldie but a goodie, an old hackathon demo built using Intel’s now-discontinued Edison IoT kit and Seeed’s Grove sensor kit.
It actually did win a small prize, and was actually pretty fun to build, and eventually led to the embedded systems job I’m about to leave.