thread.nim

Types

Thread* = pointer

The SDL thread structure, defined in SDL_thread.c

ThreadID* = culong

The SDL thread ID

TLSID* = cuint

Thread local storage ID, 0 is the invalid ID

ThreadPriority* {. size: sizeof(cint) .} = enum THREAD_PRIORITY_LOW, THREAD_PRIORITY_NORMAL, THREAD_PRIORITY_HIGH

The SDL thread priority.

Note On many systems you require special privileges to set high priority.

ThreadFunction* = proc (data: pointer): cint {. cdecl .}

The procedure passed to createThread(). It is passed a pointer user context parameter and returns an int.

Procs

proc createThread*(fn: ThreadFunction; name: cstring; data: pointer): Thread {. cdecl, importc: "SDL_CreateThread", dynlib: SDL2_LIB .}

Create a thread.

Thread naming is a little complicated: Most systems have very small limits for the string length (Haiku has 32 bytes, Linux currently has 16, Visual C++ 6.0 has nine!), and possibly other arbitrary rules. You'll have to see what happens with your system's debugger. The name should be UTF-8 (but using the naming limits of C identifiers is a better bet). There are no requirements for thread naming conventions, so long as the string is null-terminated UTF-8, but these guidelines are helpful in choosing a name:

http://stackoverflow.com/questions/149932/naming-conventions-for-threads

If a system imposes requirements, SDL will try to munge the string for it (truncate, etc), but the original string contents will be available from getThreadName().

proc getThreadName*(thread: Thread): cstring {. cdecl, importc: "SDL_GetThreadName", dynlib: SDL2_LIB .}

Get the thread name, as it was specified in createThread(). This procedure returns a pointer to a UTF-8 string that names the specified thread, or nil if it doesn't have a name. This is internal memory, not to be free()'d by the caller, and remains valid until the specified thread is cleaned up by waitThread().

proc threadID*(): ThreadID {. cdecl, importc: "SDL_ThreadID", dynlib: SDL2_LIB .}

Get the thread identifier for the current thread.

proc getThreadID*(thread: Thread): ThreadID {. cdecl, importc: "SDL_GetThreadID", dynlib: SDL2_LIB .}

Get the thread identifier for the specified thread.

Equivalent to threadID() if the specified thread is nil.

proc setThreadPriority*(priority: ThreadPriority): cint {. cdecl, importc: "SDL_SetThreadPriority", dynlib: SDL2_LIB .}

Set the priority for the current thread.

proc waitThread*(thread: Thread; status: ptr cint) {. cdecl, importc: "SDL_WaitThread", dynlib: SDL2_LIB .}

Wait for a thread to finish. Threads that haven't been detached will remain (as a "zombie") until this procedure cleans them up. Not doing so is a resource leak.

Once a thread has been cleaned up through this procedure, the Thread that references it becomes invalid and should not be referenced again. As such, only one thread may call WaitThread() on another.

The return code for the thread procedure is placed in the area pointed to by status, if status is not nil.

You may not wait on a thread that has been used in a call to detachThread(). Use either that procedure or this one, but not both, or behavior is undefined.

It is safe to pass nil to this procedure; it is a no-op.

proc detachThread*(thread: Thread) {. cdecl, importc: "SDL_DetachThread", dynlib: SDL2_LIB .}

A thread may be "detached" to signify that it should not remain until another thread has called waitThread() on it. Detaching a thread is useful for long-running threads that nothing needs to synchronize with or further manage. When a detached thread is done, it simply goes away.

There is no way to recover the return code of a detached thread. If you need this, don't detach the thread and instead use waitThread().

Once a thread is detached, you should usually assume the Thread isn't safe to reference again, as it will become invalid immediately upon the detached thread's exit, instead of remaining until someone has called waitThread() to finally clean it up. As such, don't detach the same thread more than once.

If a thread has already exited when passed to detachThread(), it will stop waiting for a call to waitThread() and clean up immediately. It is not safe to detach a thread that might be used with waitThread().

You may not call waitThread() on a thread that has been detached. Use either that procedure or this one, but not both, or behavior is undefined.

It is safe to pass nil to this procedure; it is a no-op.

proc tlsCreate*(): TLSID {. cdecl, importc: "SDL_TLSCreate", dynlib: SDL2_LIB .}

Create an identifier that is globally visible to all threads but refers to data that is thread-specific.

Return The newly created thread local storage identifier, or 0 on error.

var
  tls_lock: SpinLock
  thread_local_storage: SDL_TLSID

proc setMyThreadData(value: pointer) =
  if not thread_local_storage:
    atomicLock(addr(tls_lock))
      if not thread_local_storage:
        thread_local_storage = tlsCreate()
      atomicUnlock(addr(tls_lock))
  tlsSet(thread_local_storage, value, 0)

proc getMyThreadData(): pointer =
  return tlsGet(thread_local_storage)

Module thread

thread.nim

Types

Procs