Peter Hinch Pushbutton class stopped working

1

Hi, hoping somebody can help - even if just by reproducing tthe issue I am seeing.

I have used Peter Hinch’s excellent asyncio repo extensively see here… it has many fantastic modules, including a Pushbutton class, which I’ve been happily using for over a year. So… got a PCB fabbed for my project, so I can migrate off the dodgy breadboard setup my prototype is currently running on. And that’s where the problems surfaced…

I have a 5-way (up/down/left/right/OK) navigation switch… that would not work on my PCB. I suspected ground loop issues. Nope. Same result when I ran the Pico on a second bare-bones breadboard. Changed EVERY component of my setup .. same. Tried a different Pico. Tried re-flashing. Tried different versions of MicroPython. Nup.

The core issue (sorry for the pun) is that the Pushbutton driver uses a delay_ms.py and launch.py…to achieve both debounce, double-click and long-click processing. Nice. Except… this now FAILS. Serious fail… as in:

Task exception wasn’t retrieved
future: coro= <generator object ‘_timer’ at 20013f90>
Traceback (most recent call last):
File “asyncio/core.py”, line 1, in run_until_complete
File “delay_ms.py”, line 49, in _timer
TypeError: ‘module’ object isn’t callable

Is there anyone else using Peter Hinch’s async repo that I can compare notes with? Having changed everything… I am kinda desperate for some fresh eyes,

Thanks,
T

2

Hi!

From the error you have shared it looks like the delay_ms might be imported incorrectly. _timer expects a fuction oe class, so if the module itself is called, it will fail.

You just try few things:

  1. Check that you’re importing exactly what pushbutton expects from delay_ms.py
  2. Test a minimal setup with only one pushbutton on your PCB.
  3. Double check wiring and switch orientation against your breadboard version.
  4. Make sure you are using the latest version of Peter Hinch’s repo

Try these things and tell me is it working or not.

3

Thanks @Aakriti314205 for your input - appreciated.

Yes I check Peter’s repo… there has been no change since I grabbed his code, but note that I have both a) Peter’s original delay_ms (causing crashes)… and b) a cutdown minimal version I have been using for over a year, which is similar, but not identical (it does NOT call the _timer method…) but I am going down rabbit holes. My problem surfaced (initially) using the cut-down version… worked on my prototype… but failed on PCB, and also on my bare-bones B/B. This gets a bit convoluted… stick with me, I am determined to crack this tonight!

I have cut down to a single pushbutton on my breadboard… NOTHING else. So there is not much wiring to check!! I assure you, this weirdness is no simple wiring issue. I am however, seeing other unexpected behaviour, and am now documenting everything I do, every change I make. There is clearly something sus going on… at the MicroPython level, as I just had a 100% perfect operation, using my original Pico, but on a stripped down breadboard… and then in the VS terminal… I got no response… eventually followed by this: (the numbers are just my quick way of testing REPL response… type a number, it should be echoed back. It isn’t…)

>>> Error executing commandException occured

>>> 
>>> 1

>>> 
>>> 
>>> 2

>>> Error executing commandException occured

>>> 
>>> 7

>>> Error executing commandException occured

>>> 
>>> 5

>>> Error executing commandException occured

So the REPL was NOT responding normally.

I will continue methodical testing and documenting… I now suspect it may have something to do with my import statements… we’ll see (hopefully, before the night is over!)

If you have a Pico 2 and a pushbutton switch… and Peter’s Pushbutton class etc… I’d REALLY appreciate it if you could do a test… but let me poke around a bit more first.

Cheers, T

4

Problem solved… I discovered my “new” Pico 2 had a slightly different version of launch.py than the old one. No idea how/when that happened. And that is why nothing worked properly.

Peter has recommended I install his full repo on my Picos… had I done that way back, this issue would perhaps never have arisen. Dang…

Cheers,
T.

2 Likes
5

Ahh makes sense now. Those small version changes can cause big issues.

I’m glad you got it working.

1 Like
6

Diagnosing this was made more difficult as there were actually two contributing factors (mabe three… if you include my stupidity)

  1. I messed up the import . Don’t know why I changed it…
  2. I was using an erzatz launch method - as I could not find Peter’s offical one (now solved)

When I started this project, a) I had never used Python, b) I had never used microcontrollers, and c) it began on a Pico 1… with limited resources. Which contributed to my NOT installing Peter’s repo in full (space consideratons). WHich led to problems downstream.

I’ve learned a lot since then :wink: Good news is that last night I got my made codebase working on my PCB platform… albeit I still have some strange voltages messing up some functions, but that is a whiole other story…

Thanks for you comments. In particular, the one about importing and modules… there was a hint in the error message, I was too focussed on the “launch” failure to heed it.

Cheers,

2 Likes