[issue35813] shared memory construct to avoid need for serialization between processes

[Giampaolo Rodola']

Giampaolo Rodola' <g.rodola at gmail.com> added the comment:

> Because shared memory blocks are not "owned" by a single process, they are not destroyed/freed when a process exits. 

1) it seems SharedMemory.close() does not destroy the memory region (I'm able to re-attach to it via name). If I'm not mistaken only the manager can do that. As such it probably makes sense to have an unlink() or destroy() methods. Maybe SharedMemory should support the with statement. 


2) it seems SharedMemory is only supposed to handle `memoryview`s. I suggest to turn SharedMemory.buf in a read-onl property:

>>> sm = shared_memory.SharedMemory(None, size=23)
>>> sm.buf = [1,2,3]
>>> sm.close()
  File "/home/giampaolo/svn/cpython-shm/Lib/multiprocessing/shared_memory.py", line 166, in close
    self.buf.release()
AttributeError: 'list' object has no attribute 'release'


3) it seems "size" kwarg cannot be zero (current default). Suggestion: either choose default != 0 or make it a mandatory arg:

>>> shared_memory.SharedMemory(name=None)
ValueError: cannot mmap an empty file
>>> shared_memory.SharedMemory(name=None, size=0)
ValueError: cannot mmap an empty file
>>> shared_memory.SharedMemory(name=None, size=1)
PosixSharedMemory('psm_32161_20838', size=1)
>>> 


4) I wonder if we should have multiprocessing.active_memory_children() or something, similar to multiprocessing.active_children() so that one can do:

@atexit.register
def cleanup():
    for x in multiprocessing.active_memory_children():
        x.close()  # or unlink/destroy

I guess that would be useful for people not using a manager. Also, I'm not sure if active_memory_children() can return meaningful results with a brand new process (I suppose it can't).

----------

_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue35813>
_______________________________________