Compilation Database Generator

Tool for generating Clang's JSON Compilation Database file for GNU make-based build systems.

It's aimed mainly at non-cmake (cmake already generates compilation database) large codebases. Inspired by projects like YCM-Generator and Bear, but faster (mainly with large projects), since in most cases it doesn't need a clean build (as the mentioned tools do) to generate the compilation database file, to achieve this it uses the make options such as -n/--dry-run and -k/--keep-going to extract the compile commands. Also, it's more cross-compiling friendly than YCM-generator's fake-toolchanin approach.

Installation

# pip install compiledb

• Supports Python 2.x and 3.x (for now, tested only with 2.7 and 3.6 versions)
• For bash completion support, add the content of sh-completion/compiledb.bash file to your .bashrc file, for example.
• ZSH completion coming soon :)

Usage

compiledb provides a make python wrapper script which, besides to execute the make build command, updates the JSON compilation database file corresponding to that build, resulting in a command-line interface similar to Bear.

To generate compile_commands.json file using compiledb-generator's "make wrapper" script, executing Makefile target all:

$ compiledb make

compiledb forwards all the options/arguments passed after make subcommand to GNU Make, so one can, for example, generate compile_commands.json using core/main.mk as main makefile (-f flag), starting the build from build directory (-C flag):

$ compiledb make -f core/main.mk -C build

By default, compiledb make generates the compilation database and runs the actual build command requested (acting as a make wrapper), the build step can be skipped using the -n or --no-build options.

$ compiledb -n make

compiledb base command has been designed so that it can be used to parse compile commands from arbitrary text files (or stdin), assuming it has a build log (ideally generated using make -Bnwk command), and generates the corresponding JSON Compilation database.

For example, to generate the compilation database from build-log.txt file, use the following command.

$ compiledb --parse build-log.txt

or its equivalent:

$ compiledb < build-log.txt

Or even, to pipe make's output and print the compilation database to the standard output:

$ make -Bnwk | compiledb -o-

Testing / Contributing

I've implemented this tool because I needed to index some AOSP's modules for navigating and studying purposes (after having no satisfatory results with current tools available by the time such as YCM-Generator and Bear). So I've reworked YCM-Generator, which resulted in the initial version of compiledb/parser.py and used successfully to generate compile_commands.json for some AOSP modules in ~1min running in a Docker container and then could use it with some great tools, such as:

• Vim + YouCompleteMe + rtags + chromatica.nvim
• Neovim + LanguageClient-neovim + cquery + deoplete

Notice:

• Windows: tested on Windows 10 with cmd, wsl(Ubuntu), mingw32
• Linux: tested only on Arch Linux so far

Patches are always welcome :)

License

GNU GPLv3