HardwareSerial::flush() returns immediately if called as Stream*->flush() interface

[mlaga97]

Please, before reporting any issue

• Make sure you are using the latest Arduino_Core_STM32 version.
  https://github.com/stm32duino/Arduino_Core_STM32/releases/latest
• Make sure the issue is not already reported/fixed on GitHub or discussed on the stm32duino forum
• Submit a GitHub issue only for reporting a problem related to the Arduino_Core_STM32.
• Avoid to submit a GitHub issue for project troubleshooting.

Any questions/feedback/suggestions should be discussed on the stm32duino forum:

• questions on the STM32 Core
• bugs/enhancements on the STM core: Bugs and enhancements

⚠ When reporting any issue, please try to provide all relevant information to help on its resolution.

Describe the bug
HardwareSerial::flush() returns early if called via Stream* interface

To Reproduce

const int PIN_EN = PA1;
const int PIN_TX = PA2;
const int PIN_RX = PA3;

HardwareSerial Serial1(PIN_RX, PIN_TX);
Stream* streamTest = &Serial1;

void setup() {
  Serial1.begin(115200);
  pinMode(PIN_EN, OUTPUT);
}

void loop() {
  // Calling directly
  Serial1.print("aaaaaa");

  digitalWrite(PIN_EN, HIGH);
  Serial1.flush();
  digitalWrite(PIN_EN, LOW); // Goes low only after data has been fully transmitted

  delay(1);

  // Calling via generic stream interface
  streamTest->print("bbbbbb");

  digitalWrite(PIN_EN, HIGH);
  streamTest->flush();
  digitalWrite(PIN_EN, LOW); // Goes low immediately

  delay(1);
}

Expected behavior
In the above code, streamTest->flush() should return only once the data has been fully flushed.

Screenshots

Desktop (please complete the following information):

• OS: Ubuntu 22.04
• Arduino IDE version: [e.g. 1.8.8]
• STM32 core version: 2.7.1, issue appears to be a result of feat(serial): add timeout support to flush() #2124
• Tools menu settings if not the default: default
• Upload method: ST-LINKv2

Board (please complete the following information):

• Name: Generic STM32G030
• Hardware Revision: N/A
• Extra hardware used if any: N/A

Additional context
This appears to be a side effect of #2124, as reverting that commit results in the issue going away.

It seems as though the optional parameter is preventing HardwareSerial::flush(uint32_t timeout = 0) from properly overloading Stream::flush(void).

Therefore, another method of resolving is by modifying HardwareSerial.h from

virtual void flush(uint32_t timeout = 0);

to

virtual void flush()
{
  flush(0);
}
virtual void flush(uint32_t timeout);

[mlaga97]

Need to search better. #2254

[fpistm]

Hi @mlaga97
No worry, fix will be available within next release in the coming weeks.