Calling Serial.flush() insta-kills Pico

[maxgerhardt]

Consider the sketch

void setup() {
  Serial.begin(115200);
  while(!Serial) {}
  delay(200);

  Serial.println("This is a test!");
  Serial.flush();
  Serial.println("We survived!!");
}

void loop() { Serial.println("alive.."); delay(1000); }

The output of which is

This is a test!

followed by a hangup of the serial monitor window (cannot be closed). The on-board LED starts blinking on the board in the pattern dictated by mbed_die(). The hangup persists until the Pico is physically removed from the USB port.

Now consider the same sketch, but comment out the Serial.flush();. The output is now

This is a test!
We survived!!
alive..
alive..
alive..

Conclusion: Your Serial.flush() implementation kills the Pico, instantly.

(You will have to put the Pico into bootloader mode when plugging it in and selecting a fake upload port, otherwise you cannot reprogramm the Pico).

This is using the latest-released 2.5.2 core on the latest Arduino IDE 1.8.16, selected board is the Raspberry Pi Pico.

[facchinm]

Should be fixed by #341 (in master but not yet released).
@maxgerhardt would you mind applying the patch manually and reporting if it fixes the issue?

[maxgerhardt]

Yes, editing the file locally in C:\Users\<user>\AppData\Local\Arduino15\packages\arduino\hardware\mbed_rp2040\2.4.1\cores\arduino\Serial.cpp with the fix and reuploading fixes the problem.

This is a test!
We survived!!
alive..
alive..

Serial.flush() crashing the board in the common Serial USB configuration seems pretty crucial to me, I'm sure some sketches and libraries rely on it. The RP2040_RTC library linked above e.g. does to print its debug output. Don't you want to do a re-release with the bugfix?

[facchinm]

The plan was re-releasing this week but we are waiting a bit more to merge some more goodies, so next week the change will likely be live