1 :mod:`thread` --- Multiple threads of control

     2 =============================================

     3 

     4 .. module:: thread

     5    :synopsis: Create multiple threads of control within one interpreter.

     6 

     7 .. note::

     8    The :mod:`thread` module has been renamed to :mod:`_thread` in Python 3.0.

     9    The :term:`2to3` tool will automatically adapt imports when converting your

    10    sources to 3.0; however, you should consider using the high-level

    11    :mod:`threading` module instead.

    12 

    13 

    14 .. index::

    15    single: light-weight processes

    16    single: processes, light-weight

    17    single: binary semaphores

    18    single: semaphores, binary

    19 

    20 This module provides low-level primitives for working with multiple threads

    21 (also called :dfn:`light-weight processes` or :dfn:`tasks`) --- multiple threads of

    22 control sharing their global data space.  For synchronization, simple locks

    23 (also called :dfn:`mutexes` or :dfn:`binary semaphores`) are provided.

    24 The :mod:`threading` module provides an easier to use and higher-level

    25 threading API built on top of this module.

    26 

    27 .. index::

    28    single: pthreads

    29    pair: threads; POSIX

    30 

    31 The module is optional.  It is supported on Windows, Linux, SGI IRIX, Solaris

    32 2.x, as well as on systems that have a POSIX thread (a.k.a. "pthread")

    33 implementation.  For systems lacking the :mod:`thread` module, the

    34 :mod:`dummy_thread` module is available. It duplicates this module's interface

    35 and can be used as a drop-in replacement.

    36 

    37 It defines the following constant and functions:

    38 

    39 

    40 .. exception:: error

    41 

    42    Raised on thread-specific errors.

    43 

    44 

    45 .. data:: LockType

    46 

    47    This is the type of lock objects.

    48 

    49 

    50 .. function:: start_new_thread(function, args[, kwargs])

    51 

    52    Start a new thread and return its identifier.  The thread executes the function

    53    *function* with the argument list *args* (which must be a tuple).  The optional

    54    *kwargs* argument specifies a dictionary of keyword arguments. When the function

    55    returns, the thread silently exits.  When the function terminates with an

    56    unhandled exception, a stack trace is printed and then the thread exits (but

    57    other threads continue to run).

    58 

    59 

    60 .. function:: interrupt_main()

    61 

    62    Raise a :exc:`KeyboardInterrupt` exception in the main thread.  A subthread can

    63    use this function to interrupt the main thread.

    64 

    65    .. versionadded:: 2.3

    66 

    67 

    68 .. function:: exit()

    69 

    70    Raise the :exc:`SystemExit` exception.  When not caught, this will cause the

    71    thread to exit silently.

    72 

    73 ..

    74    function:: exit_prog(status)

    75 

    76       Exit all threads and report the value of the integer argument

    77       *status* as the exit status of the entire program.

    78       **Caveat:** code in pending :keyword:`finally` clauses, in this thread

    79       or in other threads, is not executed.

    80 

    81 

    82 .. function:: allocate_lock()

    83 

    84    Return a new lock object.  Methods of locks are described below.  The lock is

    85    initially unlocked.

    86 

    87 

    88 .. function:: get_ident()

    89 

    90    Return the 'thread identifier' of the current thread.  This is a nonzero

    91    integer.  Its value has no direct meaning; it is intended as a magic cookie to

    92    be used e.g. to index a dictionary of thread-specific data.  Thread identifiers

    93    may be recycled when a thread exits and another thread is created.

    94 

    95 

    96 .. function:: stack_size([size])

    97 

    98    Return the thread stack size used when creating new threads.  The optional

    99    *size* argument specifies the stack size to be used for subsequently created

   100    threads, and must be 0 (use platform or configured default) or a positive

   101    integer value of at least 32,768 (32kB). If changing the thread stack size is

   102    unsupported, the :exc:`error` exception is raised.  If the specified stack size is

   103    invalid, a :exc:`ValueError` is raised and the stack size is unmodified.  32kB

   104    is currently the minimum supported stack size value to guarantee sufficient

   105    stack space for the interpreter itself.  Note that some platforms may have

   106    particular restrictions on values for the stack size, such as requiring a

   107    minimum stack size > 32kB or requiring allocation in multiples of the system

   108    memory page size - platform documentation should be referred to for more

   109    information (4kB pages are common; using multiples of 4096 for the stack size is

   110    the suggested approach in the absence of more specific information).

   111    Availability: Windows, systems with POSIX threads.

   112 

   113    .. versionadded:: 2.5

   114 

   115 Lock objects have the following methods:

   116 

   117 

   118 .. method:: lock.acquire([waitflag])

   119 

   120    Without the optional argument, this method acquires the lock unconditionally, if

   121    necessary waiting until it is released by another thread (only one thread at a

   122    time can acquire a lock --- that's their reason for existence).  If the integer

   123    *waitflag* argument is present, the action depends on its value: if it is zero,

   124    the lock is only acquired if it can be acquired immediately without waiting,

   125    while if it is nonzero, the lock is acquired unconditionally as before.  The

   126    return value is ``True`` if the lock is acquired successfully, ``False`` if not.

   127 

   128 

   129 .. method:: lock.release()

   130 

   131    Releases the lock.  The lock must have been acquired earlier, but not

   132    necessarily by the same thread.

   133 

   134 

   135 .. method:: lock.locked()

   136 

   137    Return the status of the lock: ``True`` if it has been acquired by some thread,

   138    ``False`` if not.

   139 

   140 In addition to these methods, lock objects can also be used via the

   141 :keyword:`with` statement, e.g.::

   142 

   143    import thread

   144 

   145    a_lock = thread.allocate_lock()

   146 

   147    with a_lock:

   148        print "a_lock is locked while this executes"

   149 

   150 **Caveats:**

   151 

   152   .. index:: module: signal

   153 

   154 * Threads interact strangely with interrupts: the :exc:`KeyboardInterrupt`

   155   exception will be received by an arbitrary thread.  (When the :mod:`signal`

   156   module is available, interrupts always go to the main thread.)

   157 

   158 * Calling :func:`sys.exit` or raising the :exc:`SystemExit` exception is

   159   equivalent to calling :func:`exit`.

   160 

   161 * Not all built-in functions that may block waiting for I/O allow other threads

   162   to run.  (The most popular ones (:func:`time.sleep`, :meth:`file.read`,

   163   :func:`select.select`) work as expected.)

   164 

   165 * It is not possible to interrupt the :meth:`acquire` method on a lock --- the

   166   :exc:`KeyboardInterrupt` exception will happen after the lock has been acquired.

   167 

   168   .. index:: pair: threads; IRIX

   169 

   170 * When the main thread exits, it is system defined whether the other threads

   171   survive.  On SGI IRIX using the native thread implementation, they survive.  On

   172   most other systems, they are killed without executing :keyword:`try` ...

   173   :keyword:`finally` clauses or executing object destructors.

   174 

   175 * When the main thread exits, it does not do any of its usual cleanup (except

   176   that :keyword:`try` ... :keyword:`finally` clauses are honored), and the

   177   standard I/O files are not flushed.

   178