/docker-nodemcu-build

Docker image to build NodeMCU firmware for the ESP8266 on your machine

Primary LanguageShellMIT LicenseMIT

Docker NodeMCU build and LFS images

Docker Pulls Docker Stars License

Clone and edit the NodeMCU firmware locally on your platform. This image will take it from there and turn your code into a binary which you then can flash to the ESP8266 or the ESP32. It can also create LFS images from your Lua sources.

中文文档请参阅 README-CN.md

Target audience

There seem to be three types of NodeMCU developers:

  • NodeMCU "application developers"

    They just need a ready-made firmware. I created a cloud build service with a nice UI and configuration options for them. However, if they use LFS they might want to build their LFS images as an alternative to Terry Ellison's online service. Then this image is right for them!

  • Occasional NodeMCU firmware hackers

    They don't need full control over the complete tool chain and don't want to setup a Linux VM with the build environment. This image is exactly for them!

  • NodeMCU firmware developers

    They commit or contribute to the project on GitHub and need their own full fledged build environment with the complete tool chain. They still might find this Docker image useful.

‼️ Regular updates

If you have previously pulled this Docker image then you should update the image from time to time to pull in the latest bug fixes:

docker pull marcelstoer/nodemcu-build

Usage

Install Docker

Follow the instructions at https://docs.docker.com/get-started/.

Clone the NodeMCU firmware repository

Docker runs on a VirtualBox VM which by default only shares the user directory from the underlying guest OS. On Windows that is c:/Users/<user> and on Mac it's /Users/<user>. Hence, you need to clone the NodeMCU firmware repository to your user directory. If you want to place it outside the user directory you need to adjust the VirtualBox VM sharing settings accordingly.

git clone --recurse-submodules https://github.com/nodemcu/nodemcu-firmware.git

For ESP32 you would then switch to the dev-esp32 branch and update the submodules:

git checkout dev-esp32
git submodule update --recursive

Build for ESP8266

Configure modules and features

Note The build script adds information about the options you set below to the NodeMCU boot message (dumped to console on application start).

To configure the modules to be built into the firmware edit app/include/user_modules.h. Also consider turning on SSL or LFS in app/include/user_config.h. #define LUA_NUMBER_INTEGRAL in the same file gives you control over whether to build a firmware with floating point support or without. See the NodeMCU documentation on build options for other options and details.

The version information and build date are correctly set automatically unless you modify the parameters in app/include/user_version.h.

Build the firmware

Start Docker and change to the NodeMCU firmware directory (in the Docker console). To build the firmware run:

docker run --rm -ti -v `pwd`:/opt/nodemcu-firmware marcelstoer/nodemcu-build build

Depending on the performance of your system it takes 1-3min until the compilation finishes. The first time you run this it takes longer because Docker needs to download the image and create a container.

Output

All outputs will be created in the bin subfolder of your NodeMCU repository's root directory. They will be:

  • nodemcu_${BUILD_TYPE}_${IMAGE_NAME}.bin is the combined firmware image you can flash. BUILD_TYPE is integer or float. For IMAGE_NAME, see the Options chapter below.
    • Almost same but with .map ending, a mapfile will be saved that contains the relative offsets of functions.
  • 0x00000.bin will contain just the firmware.
  • 0x10000.bin will contain the SPIFFS.

Flash the firmware

There are several tools to flash the firmware to the ESP8266.

Create an LFS image for ESP8266

Start Docker and change to the NodeMCU firmware directory (in the Docker console). To create the LFS image run:

docker run --rm -ti -v `pwd`:/opt/nodemcu-firmware -v {PathToLuaSourceFolder}:/opt/lua marcelstoer/nodemcu-build lfs-image

This will compile and store all Lua files in the given folder including subfolders.

To only add specific files you can prepare a file containing the files to add and give them as paramater.

docker run --rm -ti -v `pwd`:/opt/nodemcu-firmware -v {PathToLuaSourceFolder}:/opt/lua marcelstoer/nodemcu-build lfs-image final/files.lst

Assume the following content of files.lst:

lib/*.lua    main.lua
../baseProject/*.lua

NOTE: use linux path separator '/' instead of Windows type ''. Basically this is just an ls expression as long as it contains no spaces and other shell escapable characters.

Assume the following files inside {PathToLuaSourceFolder} which is mounted as /opt/lua

baseProject/base.lua
baseProject/lib/baseLib.lua
final/files.lst
final/lib/lib1.lua
final/main.lua
main.lua

this would add the following files

baseProject/base.lua
final/lib/lib1.lua
final/main.lua

Output

Depending on what type(s) of firmware you built this will create one or two LFS images in the root of your Lua folder.

Build for ESP32

NodeMCU for ESP32 is built on the ESP-IDF (Espressif IoT Development Framework). It uses a menu-driven user interface Kconfig to configure all firmware features and options. Hence, building NodeMCU for ESP32 is a two step process and you will launch the Docker container twice. First to start Kconfig, select all options and write the configuration file. Then to actually build the firmware.

Note make sure you have got the Git submodules loaded as described above.

Configure modules and features

docker run --rm -ti -v `pwd`:/opt/nodemcu-firmware marcelstoer/nodemcu-build configure-esp32

This internally will run make menuconfig in the firmware directory. It will generate a sdkconfig file in the same.

Build the firmware

docker run --rm -ti -v `pwd`:/opt/nodemcu-firmware marcelstoer/nodemcu-build build

That is the exact same command as for building for the ESP8266. It analyses the available files to figure out whether you checked out NodeMCU for ESP32 or ESP8266. The build command is thus a shortcut to using build-esp32.

The process will fail early with a meaningful error message if it does not find a sdkconfig file in the firmware directory.

Output

All outputs will be created in the bin subfolder of your NodeMCU repository's root directory. They will be:

  • nodemcu_${IMAGE_NAME}.bin will be the firmware image. For IMAGE_NAME, see the Options chapter below.

Options

You can pass the following optional parameters to the Docker build like so docker run -e "<parameter>=value" -e ....

  • BUILD_MAKE_TARGETS A space-separated list of custom make targets to build, instead of the default ones.
  • IMAGE_NAME can be set to save the output files (see your platform's "Output" section for above) with fixed names. If it is not set or empty, the branch name and a timestamp will be used.
  • TZ By default the Docker container will run in UTC timezone. Hence, the time in the timestamp of the default image name (see IMAGE_NAME option above) will not be same as your host system time - unless that is UTC as well of course. To fix this you can set the TZ parameter to any valid timezone name e.g. -e TZ=Asia/Kolkata.

INTEGER_ONLY and FLOAT_ONLY are not supported anymore. Please configure LUA_NUMBER_INTEGRAL in app/include/user_config.h as described above.

Notes for Windows users

(Docker on) Windows handles paths slightly differently. You need to specify the full path to the NodeMCU firmware directory in the command:

docker run --rm -it -v c:/Users/<user>/<nodemcu-firmware>:/opt/nodemcu-firmware marcelstoer/nodemcu-build build

If the Windows path contains spaces it would have to be wrapped in quotes as usual on Windows.

docker run --rm -it -v "c:/Users/monster tune/<nodemcu-firmware>":/opt/nodemcu-firmware marcelstoer/nodemcu-build build

If this Docker container hangs on sharing the drive (or starting) check whether the Windows service 'LanmanServer' is running. See DockerBug #2196 for details.

Notes for macOS users

Docker for Mac is slow. Period. However, much of the I/O-related latency can be significantly reduced with tuned volume mounts. Docker for Mac 17.04 introduced a "delegated" flag to avoid keeping host file system and container file system in perfect sync all the time. "delegated" postpones writing back changes in the container to the host in order to achieve higher filesystem throughput.

So, instead of -v `pwd`:/opt/nodemcu-firmware you would say -v `pwd`:/opt/nodemcu-firmware:delegated (note the flag at the end).

Updating NodeMCU

The NodeMCU team hopes that you will want to regularly pull their latest updates into your cloned repository and build a new firmware. There is more than one way to skin a cat and thus this chapter has unfortunately-but-intentionally to be brief.

Starting over

The simplest process is to discard your local changes, update the firmware, and then manually reapply them. That may be appropriate if all you changed are a handful of settings in the .h files.

git reset --hard origin/<name-of-the-branch-you-work-with>
git submodule update --recursive

Afterwards you would manually re-edit the files and run Docker again.

Attempt to preserve your changes

Git is extremely flexible and powerful. What process you follow is very often just a matter of taste. In any case, unless you are familiar with Git-fu the least you want to be dealing with is conflict resolution on the command line ( both you and NodeMCU updated the same file => potential conflict).

One way to attempt to preserve your changes is using git stash. I say "attempt" because you still might end up with conflicts.

git stash
git pull
git submodule update --recursive
git stash pop

Support

Ask a question on StackOverflow and assign the nodemcu and docker tags.

For bugs and improvement suggestions create an issue at https://github.com/marcelstoer/docker-nodemcu-build/issues.

Credits

Thanks to Paul Sokolovsky who created and maintains esp-open-sdk.

A big "Thank You!" goes to Gregor Hartmann who implemented LFS-support and removed the ill-designed INTEGER_ONLY / FLOAT_ONLY parameters for this image.

Author

https://frightanic.com