rp2040-zero not correctly handled

[LeonHagenaarsKeus]

What were you trying to do?

Edit code on a WaveShare rp2040-zero running CircuitPython

What steps did you take to trigger the issue?

• Open MU editor
• Switch to rp2040 mode
• Restart MU editor
  • MU suggests that it detects a CircuitPython device, asks to switch to CircuitPython mode
• Press OK
  • MU complains:
    

What did you expect to happen?

A working editor in CircuitPython mode on the waveshare rp2040-zero

What actually happened?

An editor working with a local folder (not the circuitpython device)

Operating System Version

Windows 11 21H2 build 22000.708

Mu Version

1.1.1

Other Info

Using version below as requested in mu-editor/crash-reports#148:
https://github.com/mu-editor/mu/actions/runs/2444128277

Editor Log

mu.log

[carlosperate]

Hi @LeonHagenaarsKeus,

Thanks for opening the separate GitHub issue.

How does your CircuitPython device mount itself in the operating system? Is the device name CIRCUITPY? What drive letter does it have?
What build of CircuitPython is it running?

[dhalbert]

It looks like it was detected and then was lost five seconds later:

2022-06-07 12:51:39,497 - mu.interface.main:1072(show_confirmation) DEBUG: Detected new CircuitPython board device.
2022-06-07 12:51:39,497 - mu.interface.main:1073(show_confirmation) DEBUG: Would you like to change Mu to the CircuitPython mode?
2022-06-07 12:51:44,649 - mu.modes.base:61(get_default_workspace) INFO: Using workspace C:\Users\Leon Hagenaars-Keus\mu_code from settings file
2022-06-07 12:51:44,651 - mu.interface.main:1044(show_message) DEBUG: Could not find an attached CircuitPython device.

[carlosperate]

Without having a deep look I think the detection that asks the user to change the mode (via adafruit_board_toolkit using pyserial and USB data) is probably different than the detection that tries to configure Mu to save code into the CircuitPython MSD drive, which checks the volume name via Windows C apis.

[LeonHagenaarsKeus]

Hi @dhalbert

• Device name: PIDOXL
  (I'm creating my own 'ergodox' based keyboard with a rp2040, the L makes KMK recognize it's the left part of a split keyboard)
• Mount point: D:
• boot_out.txt (Which includes the CircuitPython version and board ID, in case it helps):

Adafruit CircuitPython 7.3.0 on 2022-05-23; Waveshare RP2040-Zero with rp2040
Board ID:waveshare_rp2040_zero

[carlosperate]

• Device name: PIDOXL

Right, that'd be the reason. Mu expects the CircuitPython devices to have CIRCUITPY as the volume name.

I'm not too familiar with the general guidelines for making your own CircuitPython devices, is the current recommendation to change the drive volume name? Is it up to be board designer? or maybe there isn't any guidelines in this area?

[LeonHagenaarsKeus]

No, CircuitPython doesn't suggest this, KMK does suggest it in order to enable autodetection of split keyboards. Having different names also helps me detect which code is going to which side.

https://github.com/KMKfw/kmk_firmware/blob/master/docs/split_keyboards.md#ee-hands

The name out of the box was CIRCUITPY. Mu does detect the serial port, so I can use it for serial communication out of the box.
Maybe being able to choose the device by hand (or putting a .mu file on the device) might be an option for those who want/need to change the name?

Perhaps detecting a code.py in the root of a drive might also be an option? (Installing CircuitPython will create a code.py file with 'hello-world'-like content)

[LeonHagenaarsKeus]

boot_out.txt is generated at each connect, this file does contain the string 'CircuitPython', so maybe it can be used for the same check?