commit | 513164be0df5277392b9a1ad5f4144bbde9b3883 | [log] [tgz] |
---|---|---|
author | Zach Deibert <zachdeibert@gmail.com> | Thu Feb 15 22:11:29 2024 -0600 |
committer | Petteri Aimonen <jpa@github.mail.kapsi.fi> | Tue Feb 20 11:29:13 2024 +0200 |
tree | b9fe91c0b57f851cea97e9464472ab48d27c3c8e | |
parent | 00227bd69df8467c5ed46f1051b0e5f6f8016771 [diff] |
Update FindNanopb to follow more modern CMake best practices
Nanopb is a small code-size Protocol Buffers implementation in ansi C. It is especially suitable for use in microcontrollers, but fits any memory restricted system.
To use the nanopb library, you need to do two things:
protoc
.The easiest way to get started is to study the project in “examples/simple”. It contains a Makefile, which should work directly under most Linux systems. However, for any other kind of build system, see the manual steps in README.txt in that folder.
Protocol Buffers messages are defined in a .proto
file, which follows a standard format that is compatible with all Protocol Buffers libraries. To use it with nanopb, you need to generate .pb.c
and .pb.h
files from it:
python generator/nanopb_generator.py myprotocol.proto # For source checkout generator-bin/nanopb_generator myprotocol.proto # For binary package
(Note: For instructions for nanopb-0.3.9.x and older, see the documentation of that particular version here)
The binary packages for Windows, Linux and Mac OS X should contain all necessary dependencies, including Python, python-protobuf library and protoc. If you are using a git checkout or a plain source distribution, you will need to install Python separately. Once you have Python, you can install the other dependencies with pip install --upgrade protobuf grpcio-tools
.
You can further customize the header generation by creating an .options
file. See documentation for details.
If you want to perform further development of the nanopb core, or to verify its functionality using your compiler and platform, you'll want to run the test suite. The build rules for the test suite are implemented using Scons, so you need to have that installed (ex: sudo apt install scons
or pip install scons
). To run the tests:
cd tests scons
This will show the progress of various test cases. If the output does not end in an error, the test cases were successful.
Note: Mac OS X by default aliases ‘clang’ as ‘gcc’, while not actually supporting the same command line options as gcc does. To run tests on Mac OS X, use: scons CC=clang CXX=clang++
. Same way can be used to run tests with different compilers on any platform.
For embedded platforms, there is currently support for running the tests on STM32 discovery board and simavr AVR simulator. Use scons PLATFORM=STM32
and scons PLATFORM=AVR
to run these tests.
Nanopb C code itself is designed to be portable and easy to build on any platform. Often the bigger hurdle is running the generator which takes in the .proto
files and outputs .pb.c
definitions.
There exist build rules for several systems:
extra/nanopb.mk
, see examples/simple
extra/FindNanopb.cmake
, see examples/cmake
tests/site_scons
(generator only)BUILD.bazel
in source rootconanfile.py
in source rootAnd also integration to platform interfaces: