MicroPython Forum

The micropython test revealed the offset (same as Roberthh's finding): >>> while True: ... gc.collect() ... t=time.ticks_ms() ... if time.ticks_us()<20: ... print("us overflow") ... time.sleep_us(20) ... delta = time.ticks_diff(time.ticks_ms(), t) ... if delta < 0 or delta >5: ... print("Fail time i...

jomas wrote: ↑

Tue May 21, 2019 9:04 am

MICROPY_PY_UTIME_TICKS_PERIOD = 2^30.
So MICROPY_PY_UTIME_TICKS_PERIOD-1 is used as a mask of 30bits (0b111111111111111111111111111111)

Ah thanks. This however gives exactly the same result.

I tried to run some tests on my arduino: uint32_t system_time_high_word=0; uint32_t current_time = 4294967295; //just before overflow uint32_t MICROPY_PY_UTIME_TICKS_PERIOD =0; uint32_t system_get_time(void) { return current_time; } uint32_t mp_hal_ticks_ms(void) { return ((uint64_t)system_time_high...

I had an event with @pythoncoder's simple test code this night. I used the settings for COUNT, MAX, MIN = 10000, 38, 34. There was a fail with a time gap of 4295005 ms, which is 1h 11m 35s 5 ms, or about 2^32 µs. A single pass runs ~35ms, which is the deviation to 2^32. Thank you @Roberthh for test...

No, your assumption is right. The whole uasyncio is blocking for 1h 11 minutes, not only one coroutine. I tested this when I first discovered the problem by running multiple testing codes (publishing sensor values and logging if there's a too big delay between executions in another coroutine) and al...

I thought I understand the implications but had to edit my post because I had thought it wrong. In uasyncio core.py in method run(): [...] if not self.runq: delay = -1 if self.waitq: tnow = self.time() t = self.waitq.peektime() delay = time.ticks_diff(t, tnow) if delay < 0: delay = 0 self.wait(delay...

That is an interesting find! If that's true, it could explain why only the esp8266 is affected as the esp32 seems to use "esp_timer_get_time()" instead in mphalport. I checked the functions you mention and to me (I don't know the inner workings of micropython) it looks like it is indeed possible to ...

Creates this output: 2019-05-20 14:56:03 (YYYY-MM-DD HH:mm:SS)

This is a very interesting project! I'm currently building a small Greenhouse on my balcony... Got almost all libraries ready for it. Why not send the current task by mqtt instead of all daily tasks? I see no need for a separate socket server (but could point you to a library pythoncoder and I creat...

I would also assume that it is something esp8266 specific. Either in micropython firmware or the esp-idf. I'm only running an ESP32 occasionally so I don't have a good comparison but didn't encounter the problem there. Only devices anyone reported this problem yet is the esp8266. If the problem affe...