Implement Ninja persistent mode.

This is the final CL that enables persistent mode for Ninja.
The feature is enabled by setting `NINJA_PERSISTENT_MODE=1`
in your environment, and will ensure that Ninja launches a
background persistent process for the current build directory,
if one does not exists.

This server process shuts down gracefully after 5 minutes
if no client connects to it, in order to release system
resources automatically. This delay can be changed by
setting `NINJA_PERSISTENT_TIMEOUT_SECONDS` in your
environment.

It it possible to use the new `-t server` tool, which
supports three commands:

  ninja [-C <dir>] -t server status
    Reports whether a server is running for <dir>

  ninja [-C <dir>] -t server stop
    Stops any running server serving <dir>.

  ninja [-C <dir>] -t server pid
    Return the PID of the running server, or -1 if
    there is none.

For debugging, it is possible to set `NINJA_PERSISTENT_MODE=server`
in your environment to start the server in the foreground process.
In this case, the client must be started in another terminal,
with `NINJA_PERSISTENT_MODE=1` to connect to it. Both must use
the same build directory (specified with `-C <dir>` or the working
directory if this option is not used).

Note that currently, any requests on a large build plan like
the Fuchsia one takes 1 to 2 seconds, because each query will
still need to stat() all paths in the build graph. This may be
cut down by a future patch implementing a proper timestamp cache
invalidated by inotify-based filesystem watches.

There are a number of BUGS / CAVEATS:

- Does not work on Windows at the moment due to a Win32
  technical limitation (console handles cannot be duplicated
  in other processes). A work-around will be provided in
  a future patch.

- Ninja tools that run after the build manifest is loaded
  (i.e. most of them) are always run in the client process.

Fuchsia-Topic: persistent-mode
Original-Change-Id: Ice3f5dd04410d6e3d9bf137d1ad97383260c6653
Original-Change-Id: I81e4efac0f8c56979e8abdcd7459dada62aca929
Original-Change-Id: I19e9d7d1f2c410742c745d473d430b018c38c289
Original-Change-Id: I2a1d34f848144601a320c0b8eb6bbb3071df6ca5
Original-Change-Id: Ica91c9d138ab9f355b81ea20ad89330badc5aebd
Original-Change-Id: Ifeca100d77e97193a9d347ccbdbee73661a550d8
Change-Id: Ib26c95af19050784ffdc215fb8f44882ec7f786d
Reviewed-on: https://fuchsia-review.googlesource.com/c/third_party/github.com/ninja-build/ninja/+/1071429
Reviewed-by: David Fang <fangism@google.com>
6 files changed
tree: d47e2ca2fb16d9b18bbf35d35c426975d2c6f76a
  1. .github/
  2. doc/
  3. misc/
  4. src/
  5. windows/
  6. .clang-format
  7. .clang-tidy
  8. .editorconfig
  9. .gitignore
  10. appveyor.yml
  11. CMakeLists.txt
  12. configure.py
  13. CONTRIBUTING.md
  14. COPYING
  15. README.fuchsia
  16. README.md
  17. RELEASING.md
README.md

Ninja

Ninja is a small build system with a focus on speed. https://ninja-build.org/

See the manual or doc/manual.asciidoc included in the distribution for background and more details.

Binaries for Linux, Mac and Windows are available on GitHub. Run ./ninja -h for Ninja help.

Installation is not necessary because the only required file is the resulting ninja binary. However, to enable features like Bash completion and Emacs and Vim editing modes, some files in misc/ must be copied to appropriate locations.

If you're interested in making changes to Ninja, read CONTRIBUTING.md first.

Building Ninja itself

You can either build Ninja via the custom generator script written in Python or via CMake. For more details see the wiki.

Python

./configure.py --bootstrap

This will generate the ninja binary and a build.ninja file you can now use to build Ninja with itself.

If you have a GoogleTest source directory, you can build the tests by passing its path with --gtest-source-dir=PATH option, or the GTEST_SOURCE_DIR environment variable, e.g.:

./configure.py --bootstrap --gtest-source-dir=/path/to/googletest
./ninja all     # build ninja_test and other auxiliary binaries
./ninja_test`   # run the unit-test suite.

Use the CMake build below if you want to use a preinstalled binary version of the library.

CMake

cmake -Bbuild-cmake
cmake --build build-cmake

The ninja binary will now be inside the build-cmake directory (you can choose any other name you like).

To run the unit tests:

./build-cmake/ninja_test

Generating documentation

Ninja Manual

You must have asciidoc and xsltproc in your PATH, then do:

./configure.py
ninja manual doc/manual.pdf

Which will generate doc/manual.html.

To generate the PDF version of the manual, you must have dblatext in your PATH then do:

./configure.py    # only if you didn't do it previously.
ninja doc/manual.pdf

Which will generate doc/manual.pdf.

Doxygen documentation

If you have doxygen installed, you can build documentation extracted from C++ declarations and comments to help you navigate the code. Note that Ninja is a standalone executable, not a library, so there is no public API, all details exposed here are internal.

./configure.py   # if needed
ninja doxygen

Then open doc/doxygen/html/index.html in a browser to look at it.