Coroutine-based network library
- Fast event loop based on libev or libuv.
- Lightweight execution units based on greenlets.
- API that re-uses concepts from the Python standard library (for examples there are events and queues).
- Cooperative sockets with SSL support
- Cooperative DNS queries performed through a threadpool, dnspython, or c-ares.
- Monkey patching utility to get 3rd party modules to become cooperative
- TCP/UDP/HTTP servers
- Subprocess support (through gevent.subprocess)
- Thread pools
gevent is inspired by eventlet but features a more consistent API, simpler implementation and better performance. Read why others use gevent and check out the list of the open source projects based on gevent.
gevent was written by Denis Bilenko.
See what’s new in the latest major release.
Check out the detailed changelog for this version.
Read the documentation online at http://www.gevent.org.
Installation and Requirements
gevent 1.3 runs on Python 2.7 and Python 3. Releases 3.4, 3.5 and 3.6 of Python 3 are supported. (Users of older versions of Python 2 need to install gevent 1.0.x (2.5), 1.1.x (2.6) or 1.2.x (<=2.7.8); gevent 1.2 can be installed on Python 3.3.) gevent requires the greenlet library and will install the cffi library by default on Windows.
gevent 1.3 also runs on PyPy 5.5 and above, although 5.9 or above is strongly recommended. On PyPy, there are no external dependencies.
gevent is tested on Windows, OS X, and Linux, and should run on most other Unix-like operating systems (e.g., FreeBSD, Solaris, etc.)
On Windows using the libev backend, gevent is limited to a maximum of 1024 open sockets due to limitations in libev. This limitation should not exist with the default libuv backend.
This section is about installing released versions of gevent as distributed on the Python Package Index
On Windows, OS X, and Linux, both gevent and greenlet are distributed as binary wheels.
You need Pip 8.0 or later, or buildout 2.10.0 to install the binary wheels.
On Linux, you’ll need to install gevent from source if you wish to use the libuv loop implementation. This is because the manylinux1 specification for the distributed wheels does not support libuv. The cffi library must be installed at build time.
Installing From Source
If you are unable to use the binary wheels (for platforms where no pre-built wheels are available or if wheel installation is disabled, e.g., for libuv support on Linux), here are some things you need to know.
- You can install gevent from source with pip install --no-binary gevent gevent.
- You’ll need a working C compiler that can build Python extensions. On some platforms, you may need to install Python development packages.
- Installing from source requires setuptools. This is installed automatically in virtual environments and by buildout. However, gevent uses PEP 496 environment markers in setup.py. Consequently, you’ll need a version of setuptools newer than 25 (mid 2016) to install gevent from source. Older versions of pipenv may also have issues installing gevent for this reason.
- To build the libuv backend (which is required on Windows and optional elsewhere), or the CFFI-based libev backend, you must install cffi before attempting to install gevent on CPython (on PyPy this step is not necessary).
Common Installation Issues
The following are some common installation problems and solutions for those compiling gevent from source.
- Some Linux distributions are now mounting their temporary directories with the noexec option. This can cause a standard pip install gevent to fail with an error like cannot run C compiled programs. One fix is to mount the temporary directory without that option. Another may be to use the --build option to pip install to specify another directory. See issue #570 and issue #612 for examples.
- Also check for conflicts with environment variables like CFLAGS. For example, see Library Updates.
- Users of a recent SmartOS release may need to customize the CPPFLAGS (the environment variable containing the default options for the C preprocessor) if they are using the libev shipped with gevent. See Operating Systems for more information.
gevent has no runtime dependencies outside the standard library, greenlet and (on some platforms) cffi. However, there are a number of additional libraries that extend gevent’s functionality and will be used if they are available.
The psutil library is needed to monitor memory usage.
zope.event is highly recommended for configurable event support; it can be installed with the events extra, e.g., pip install gevent[events].
To install the latest development version:
pip install setuptools cffi 'cython>=0.28' git+git://github.com/gevent/gevent.git#egg=gevent
You will not be able to run gevent’s test suite using that method.
To hack on gevent (using a virtualenv):
$ git clone https://github.com/gevent/gevent.git $ cd gevent $ virtualenv env $ source env/bin/activate (env) $ pip install -r dev-requirements.txt
The notes above about installing from source apply here as well. The dev-requirements.txt file takes care of the library prerequisites (CFFI, Cython), but having a working C compiler that can create Python extensions is up to you.
There are a few different ways to run the tests. To simply run the tests on one version of Python during development, begin with the above instructions to install gevent in a virtual environment and then run:
(env) $ cd src/greentest (env) $ python ./testrunner.py
Before submitting a pull request, it’s a good idea to run the tests across all supported versions of Python, and to check the code quality using prospector. This is what is done on Travis CI. Locally it can be done using tox:
pip install tox tox
The testrunner accepts a --coverage argument to enable code coverage metrics through the coverage.py package. That would go something like this:
cd src/greentest python testrunner.py --coverage coverage combine coverage html -i <open htmlcov/index.html>
A test suite is run for every push and pull request submitted. Travis CI is used to test on Linux, and AppVeyor runs the builds on Windows.
Builds on Travis CI automatically submit updates to coveralls.io to monitor test coverage.
On Debian, you will probably need libpythonX.Y-testsuite installed to run all the tests.
Release history Release notifications
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.