@robalstona pin 0 is the clock, not ws, according to the sticker on the back. Don't know if that helps.

Thinking about how to test it: the M5StickC has no speaker.
OK, I have the Core Gray: It has a speaker! But no mic.
Maybe I could hook both of them together...
This is getting way too complicated for the idea I had in mind, which was some responsive blinky lights for Halloween. (Yes, it's next week. Your point?) :) I think I'll use the accelerometer instead.

Thanks! Going off that, I found this post, which points to the M5GO blob, which has

from machine import I2S

Going into REPL:

 from machine import I2S
>>> dir(I2S)
['__class__', '__name__', 'read', 'start', 'stop', 'write', '__bases__', '__dict__', 'CHANNEL_ALL_LEFT', 'CHANNEL_ALL_RIGHT', 'CHANNEL_ONLY_LEFT', 'CHANNEL_ONLY_RIGHT', 'CHANNEL_RIGHT_LEFT', 'DAC_BOTH_EN', 'DAC_DISABLE', 'DAC_LEFT_EN', 'DAC_RIGHT_EN', 'FORMAT_I2S', 'FORMAT_I2S_LSB', 'FORMAT_I2S_MSB', 'FORMAT_PCM', 'FORMAT_PCM_LONG', 'FORMAT_PCM_SHORT', 'I2S_NUM_0', 'I2S_NUM_1', 'MODE_ADC_BUILT_IN', 'MODE_DAC_BUILT_IN', 'MODE_MASTER', 'MODE_PDM', 'MODE_RX', 'MODE_SLAVE', 'MODE_TX', 'adc_enable', 'bits', 'deinit', 'init', 'nchannels', 'sample_rate', 'set_adc_pin', 'set_dac_mode', 'set_pin', 'volume']
>>> import wave
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: no module named 'wave'

So at least part of it is there! I found a listing of the class, but no documentation.

Then I found some examples, but they say Micropython after June 2021, so may not apply??

That talks about a "WS" pin, which I found described here.

The schematic shows only 2 pins however: SCL & SDA. The Unit unit :) appears to have WS tied to ground. Mouser indicates that the pin is tied low internally, and that is the left channel, but the arduino code sets it to I2S_CHANNEL_FMT_ALL_RIGHT.
Here's an I2S tutorial.
So, time to play I guess!

The posts are very confusing! Half say that it is i2s, and the Arduino code seems to support that. But the other half say that it is just an adc on port 34, and the micropython examples seem to support that!
I'm interested in using micropython.

Rshell works pretty well, although the newer M5 releases are a lot harder to get into the correct mode, particularly on the atom.

Uiflow is micropython: I have often written something in uiflow and then switched to the micropython side to see what the python is for a given uiflow thing. And then I can cut and paste into repl.

For long-term debugging (fails once a day or week), I leave the device connected to putty.

I have considered adding code that would catch Exception () in the main loop, do a stack trace, and save it into a file so that I don't need it to always be connected, but have never implemented it.

@medy mostly:

print(dir(wifiCfg))
print(wifiCfg.__dict__())

I'm sure that someone has written a function that formats all this, recurses, etc, but I usually just do it manually, because there are usually only a few things that make sense to try.

Also googling for micropython wifiCfg (or whatever), because micropython is not just for M5, and a lot of the code is on the internet. It's not always exactly the same, but it often is, or is close enough to provide hints for more things to look at.

I'm assuming that you know how to use rshell or some other tool to get to repl so that you can just type in statements as they occur to you, rather than writing a script for every new thing you want to try.

@keesjan With micropython, port 22 works fine, and as expected:

CDS_PIN = 33
TX = 19
RX = 22

As an aside, the idea that the labeling of the pins is "oriental logic" shows an unawareness of the history of rs232 and is a bit insulting. The question of what to label the pins depends on whether the manufacturer thinks of the device as DTE or DCE: StackExchange has a good explanation of the history.

In particular:

'some manufacturers try to "help" by effectively labelling their DTE equipment as if it was a piece of DCE. They mark their data input pin as Tx so that user just connects "Tx" from the external device (which, if it's DTE, will be the data output from there) to the pin marked "Tx" on their equipment (which they know is an input). Thereby allowing them to say "just connect Tx on your device to Tx on our equipment". They think they are trying to help, but such labelling often just adds to the confusion.'

Hmm, OK, get_wifi() works on 1.5.4, so maybe I just guessed wrong back then.

Some time ago, I asked where the wifi config was held.
At that time, the answer was:

 import wifiCfg
      self.ssid, self.password = wifiCfg.deviceCfg.wifi_read_from_flash()

Fast forward to 1.7.0:
I have set the WiFi credentials with M5burner 2.2.2 via "Configuration". But the above returns an empty list, rather than actual credentials.

Apparently the new hotness is:

wifiCfg.deviceCfg.get_wifi()

It's a bit frustrating when this keeps changing, and there's no user-visible notice, just broken code. Can we get some kind of announcement of functionality-breaking changes?

Re: Running code with rshell?

I had rshell working for a while, but now it is failing. I filed https://github.com/dhylands/rshell/issues/138, but I wonder if it is because of the fancy UISHELL ascii logo?

This is on the Atom Mini, so it's a bit tough to figure out!

I'm running v1.6.2-lite on it, and REPL says:

MicroPython 400ff3bbf-dirty on 2020-08-06; M5Stack with ESP32