I am not using Micropython anymore, so this repo is UNMAINTAINED

ESP32 Micropython build and tool container

This docker container will build, deploy, and monitor your esp32 micropython firmware. Using this is pretty straightforward and doesn't change a lot compared to using a simple build environment on your host or on a VM. Also, it allows you to build micropython without having to juggle with python venv, ESP-IDF versions, and environment variable. The image comes with ESP-IDF v4.0.1, python3, and rshell. The container image size is 1.88GB, but could surely be smaller. Suggestions welcome.

Getting the container image:

Recommended: Pull from docker:

You can simply pull this container image from docker hub by issuing:

docker pull tionebrr/micropython-tools-esp32

OR Build the container image from github:

You can also clone this repository in the folder of your choice, then cd into it and build the container image yourself.

docker build -t micropython-tools-esp32 .

The build will take about 5 to 10 minutes. Enjoy a coffee or a good green tea. =)

Prepare the micropython source:

The micropython source is not present in this image. Your micropython folder will be used "in place" for the build. If you don't already have the micropython sources present locally, you'll need to clone the micropython github repo in a folder of your choice.

git clone https://github.com/micropython/micropython

You'll also need to add a makefile in the ports/esp32 folder of micropython. To do this:

cd your/path/to/micropython/ports/esp32
nano makefile

This makefile works with my board, but you may need to change those values according to your hardware. Please refer to the micropython documentation for more details.

BOARD ?= GENERIC
FLASH_MODE ?= dio
FLASH_SIZE ?= 4MB

include Makefile

If you have read the micropython esp32 readme, you may be wondering where are the CROSS_COMPILE, ESPIDF and PORT environment variables. In our case, they are automatically set in the container (by the Dockerfile) during the image building process. This makefile now only contains informations about the target board.

Run the container:

This command will start the container in interactive mode. You'll end up in the container bash shell, ready to build. Don't forget to replace the paths with your micropython folder and your ESP32 serial device.

docker run -it --rm -v path/to/your/micropython/:/root/micropython --device=/dev/ttyUSBx:/dev/ttyESP tionebrr/micropython-tools-esp32 

If you want to use rshell from this container to upload custom scripts to your device, you will need to specify a volume in the docker run command. Simply add -v /path/to/your/µpy/scripts:/root/scripts before the container image name (tionebrr/micro...) like this:

docker run -it --rm -v path/to/your/micropython/:/root/micropython -v path/to/your/micropython-scripts:/root/scripts --device=/dev/ttyUSBx:/dev/ttyESP tionebrr/micropython-tools-esp32 

Build micropython, deploy to ESP32:

Once in the container shell, you can build and deploy using the same commands you would use with a normal micropython build environment.

If it was not already done in your micropython folder, you need to build the mpy cross compiler first:

cd micropython/mpy-cross
CROSS_COMPILE='' make

I may remove the capability of building mpy-cross from the container in the future; most developers already have gcc and make installed for their system and could build mpy-cross from the host.

You can now build and deploy micropython to your ESP32.

cd ../ports/esp32
make submodules
make
make erase
make deploy

You can find the output files in micropython/ports/esp32/build-GENERIC and these files will stay available even after you exited the container.

Start rshell:

After building and deploying, you can interact with the micropython REPL using rshell.

rshell -p $PORT -b 115200

Have fun.