Please read the documentation on the website!
- Install PlatformIO Core or PlatformIO IDE (which installs PlatformIO core automatically), as well as the packages specified in
scripts/requirements.txt
(e.g.pip install -r scripts/requirements.txt
). - For LUA-scripting support (see env:LIGHT_EDITION_DEV_LUA): Install SWIG (also available in most package managers, e.g.
brew install swig
orapt install swig
) - Then clone this repository:
As this repository contains binary data (e.g. schematics or images), make sure to have
$ git clone --recurse-submodules https://github.com/Open-Smartwatch/open-smartwatch-os.git
git-lfs
installed! Cloning this repository withoutgit-lfs
will result in missing or broken files. - To update the sources later on, run:
$ git pull $ git submodule update --init --recursive
The master
branch is a stable version and the develop
branch is our beta version. Recommended is, that you upload the master
branch - except if you want to help us by testing the next release or need (very) new features 😉.
Open the cloned repo in VSCode:
$ code open-smartwatch-os
You may rename the file include/config.h.example
to include/config.h
and adapt the values according to your requirements. That config is only applied once after you wiped the flash of the watch or changed the config-version numer in osw_config.h
(...).
Alternatively, you can go to the repo folder with your terminal and run...
$ pio run -t upload
If you want to compile for a specific model, you can use the -e
flag with an env
name from the platformio.ini
file.
To get started, take a look into the examples in the src/apps/examples
folder - or just into any other app. If you want to compile the examples or other (by default) excluded applications, take a look into the main.cpp
file and add the respective flags to the platformio.ini
file.
If you want to print out the log for debugging (also including decoded exception traces), use the following command:
$ pio device monitor
- Wi-Fi needs to be able to connect for this to work.
- You will need bash and ImageMagick for the helper scripts to work
- You'll need to enable the respective feature flag to enable it (see below)
The raw screenserver runs in the background and should report via serial:
Started RAW ScreenServer under http://.../api/screenserver
- Add
-D RAW_SCREEN_SERVER
to your build flags inplatformio.ini
- Build + Flash + Reboot the watch
- Connect + wait for the server to be started (see msg above)
- Run
bash fetchScreen.sh <IP_OF_WATCH> screenshot.png
- Run
bash composeScreen.sh screenshot.png screenshot_composed.png
If there is no curl
, it must be installed.
$ apt install curl -y
The fetchScreen.sh
downloads the raw image buffer from the running screen server, and converts the image to png. The composeScreen.sh
creates the image with a surrounding smartwatch "overlay" (light edition).
Run the following inside the open-smartwatch-os
directory:
$ cd scripts/screen_capture/
$ ./createScreenshot.sh <IP_OF_WATCH> <SCREENSHOT>
- The captured file can be found in the
screenshot/
folder inside theopen-smartwatch-os
directory.
For more information on troubleshooting, see Wiki.
You did not clone the repository with the --recursive-submodules
flag.
You did not hold down BTN1(FLASH) and then tap the RESET button on the watch whilst platform.io was trying to connect.
The OS itself can be executed as a regular program on your machine. This saves you time compiling for the watch and flashing the OS, every time you make a minor change - e.g. while developing the UI or a game, which not explicitly depend on the hardware of the watch.
This also implies some limitations what you can do with the emulator, as we had to hack and reimplement some of the Arduino-specific libraries and their (conflicting) simplifications. This also means, that it maybe necessary to extend those extensions down the road as we (likely) missed that one specific function you try to use... 😉
The emulator can be build using the CMakeLists.txt
file - you may need to install additional libraries to be able to use it.
Here is a small example running on "Ubuntu 22.04 LTS":
$ sudo apt install libsdl2-dev libsdl2-image-dev g++ gcc cmake make build-essential
$ mkdir build && cd build
$ cmake ..
$ make -j $(nproc)
$ ./emulator.run
You also may extend the cmake
-command with -DCMAKE_BUILD_TYPE=Release
to get an even faster and smaller binary.
If a library is unavailable, you can still use the emulator using docker (e.g. on Ubuntu 20.04 SDL2 is too old). Proceed with a typical docker installation. Showing an application running in docker requires some additional steps:
Tested on Ubuntu 20.04.
$ xhost +
$ xauth list # Copy the result of the command.
$ docker run --net=host -e DISPLAY -v /tmp/.X11-unix -d --name OSW -p 22:22 -it --privileged ubuntu:22.04
$ xauth add <'xauth list' command result>
Everything in this repository is under a GPL-3.0 license, see here for more details.