Typing Stubs for gevent

Warnings

• breaking Type stubs are version-specific. Installing 'types-gevent' for 'gevent==X.Y' alongside a different version of 'gevent' (e.g., 'gevent==A.B' where A.B != X.Y) can lead to incorrect type checking errors or missed type issues. The 'types-gevent' package version `X.Y.Z.YYYYMMDD` indicates it targets `gevent==X.Y.*`.
  Fix: Ensure that the major and minor version numbers of `types-gevent` (e.g., '25.9' from '25.9.0.20260408') match or are compatible with your installed `gevent` runtime version. Pin both `gevent` and `types-gevent` to compatible versions in your project's dependencies.
• gotcha Stub packages like 'types-gevent' provide static type information only; they do not include any runtime code or functionality. Installing 'types-gevent' without the actual 'gevent' library will not make 'gevent' available for execution.
  Fix: Always install the 'gevent' runtime library (`pip install gevent`) in addition to its type stubs (`pip install types-gevent`) for your application to function correctly.
• gotcha When using `gevent.monkey.patch_all()`, the patching must occur as early as possible in your application's lifecycle, typically at the very beginning of your main script. Delaying patching, especially after other standard library modules (e.g., `concurrent.futures`, `threading`) have been imported, can result in those modules retaining their original blocking behavior or lead to conflicts.
  Fix: Place `gevent.monkey.patch_all()` at the top of your application's entry point file before any other imports that might use standard library I/O or concurrency primitives.
• gotcha Gevent is designed for I/O-bound concurrency, not CPU-bound parallelism. Running CPU-intensive tasks within a greenlet will block the entire event loop, preventing other greenlets from executing and negating the benefits of cooperative multitasking.
  Fix: Offload CPU-intensive operations to separate processes or threads (e.g., using `multiprocessing` or `concurrent.futures.ThreadPoolExecutor` where appropriate) to avoid blocking the gevent event loop.
• gotcha Combining `gevent` with Python's built-in `multiprocessing` module can introduce complex issues and is generally discouraged without careful management. Specifically, after forking on POSIX systems, `gevent`'s internal state in the child process can be ill-posed, leading to unexpected behavior.
  Fix: If multiprocessing is necessary, carefully review the `gevent` documentation on `gevent.subprocess` and `multiprocessing` interactions. Consider using process pools or alternative strategies for multi-core utilization that are compatible with `gevent`'s cooperative concurrency model.

Install

• pip install types-gevent Install stubs

Imports

• gevent

  import gevent

  types-gevent itself is not imported directly; it provides type information for imports from the 'gevent' library.
• gevent.monkey

  import gevent.monkey

  Used for applying monkey patches to standard library modules.

Quickstart

This example demonstrates how types-gevent provides type hints for the gevent library. After installing both 'gevent' and 'types-gevent', a type checker like MyPy can analyze the code. The type hints for `fetch_data` and `process_urls` will be checked, and `gevent.spawn` and `gevent.joinall` will have their expected types available for static analysis.

import gevent
import gevent.monkey
import typing

gevent.monkey.patch_all()

def fetch_data(url: str) -> str:
    """Simulates fetching data asynchronously."""
    gevent.sleep(0.1) # Simulate I/O delay
    return f"Data from {url}"

def process_urls(urls: typing.List[str]) -> typing.List[str]:
    greenlets = [gevent.spawn(fetch_data, url) for url in urls]
    gevent.joinall(greenlets)
    results = [g.value for g in greenlets]
    return results

if __name__ == "__main__":
    urls_to_process = ["http://example.com/1", "http://example.com/2"]
    processed_data = process_urls(urls_to_process)
    print(f"Processed: {processed_data}")

# To run type checking, save this as e.g., 'main.py' and run:
# mypy main.py