greenlet Namespace Reference

Namespaces
namespace	tests

Classes
class	AttributeError
class	BrokenGreenlet
class	ExceptionState
class	Greenlet
class	GreenletGlobals
class	LockInitError
class	MainGreenlet
class	PyErrOccurred
class	PyFatalError
struct	PythonAllocator
class	PythonState
class	PythonStateContext
class	StackState
class	SwitchingArgs
class	ThreadState
struct	ThreadState_DestroyNoGIL
class	ThreadStateCreator
class	TypeError
class	UserGreenlet
class	ValueError

Typedefs
typedef std::vector< ThreadState * >	cleanup_queue_t
typedef std::mutex	Mutex
typedef std::lock_guard< Mutex >	LockGuard
typedef void(*	ThreadStateDestructor) (ThreadState *const)

Functions
OwnedObject &	operator<<= (OwnedObject &lhs, greenlet::SwitchingArgs &rhs) noexcept
template<typename T >
void	operator<< (const PyThreadState *const tstate, T &exc)

Detailed Description

The root of the greenlet package.

Defines various utility functions to help greenlet integrate well with threads. This used to be needed when we supported Python 2.7 on Windows, which used a very old compiler. We wrote an alternative implementation using Python APIs and POSIX or Windows APIs, but that's no longer needed. So this file is a shadow of its former self — but may be needed in the future.

Implementation of greenlet::UserGreenlet.

Format with: clang-format -i –style=file src/greenlet/greenlet.c

Fix missing braces with: clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"

Function Documentation

operator<<=()

OwnedObject & greenlet::operator<<= ( OwnedObject &  lhs, greenlet::SwitchingArgs &  rhs  )

noexcept

CAUTION: May invoke arbitrary Python code.

Figure out what the result of greenlet.switch(arg, kwargs) should be and transfers ownership of it to the left-hand-side.

If switch() was just passed an arg tuple, then we'll just return that. If only keyword arguments were passed, then we'll pass the keyword argument dict. Otherwise, we'll create a tuple of (args, kwargs) and return both.

CAUTION: This may allocate a new tuple object, which may cause the Python garbage collector to run, which in turn may run arbitrary Python code that switches.