- Current – 0.9.4 (2 Sept 2012) –
Fixed a buglet. When creating shared memory under Linux and
specifying both a size and the read-only flag, creating the memory
would succeed but calling ftruncate() would fail.
The failure to change the size was correctly reported
but posix_ipc failed to clean up the shared memory segment
it had created. That's now fixed. Thanks to Kevin Miles for the bug
- 0.9.3 (2 Jan 2012) –
Added a bugfix/feature to raise an error (rather than segault)
when trying to use a closed semaphore.
Thanks to Russel for the suggestion and patch.
- 0.9.2 (6 Nov 2011) –
- Fixed a bug where timeouts in Semaphore.acquire(),
MessageQueue.send() and MessageQueue.receive()
were only accurate to about one second due to use of the C call
time(). Switching to gettimeofday() fixes
the problem. Thanks to Douglas Young for the bug report and
- Fixed a bug in prober.py that caused install to fail
under Ubuntu 11.10. prober.py specified link options
in the wrong order, and so linking one of the test
applications that's built during setup was failing. Thanks
to Kevin Miles for the bug report.
- Added a check in prober.py to see if
sysconf_names exists in the os module. It
doesn't exist under Cygwin, and this code caused an error
on that platform. Thanks to Rizwan Raza for the bug report.
- 0.9.1 (7 Apr 2011) –
- Fixed (?) a bug in message queue thread notification that caused
ceval: tstate mix-up and other fun messages. Thanks to
Lev Maximov for the bug report.
- Added the demo3 directory with demos of message queue.
This was supposed be included in version 0.9.0 but I accidentally
left it out. (Also reported by Lev.)
- 0.9.0 (31 Dec 2010) –
Added the demo3 directory with demos of message queue
notification techniques. Also, fixed two bugs related to message
queue notification. Big thanks to
Philip D. Bober for debugging and providing a patch to the
most difficult part of the code. The bugs were –
- First, the series of calls to set up the Python thread in
process_notification() were simply wrong. They worked
some (most?) of the time but would segfault eventually because
I was creating a Python thread state when I should not have.
- Second, the code in process_notification() failed
to consider that the user's callback might re-request
notification, thus overwriting pointers that I would later
decref. process_notification() is now thread-safe.
- 0.8.1 (15 Mar 2010) –
Fixed a sloppy declaration that caused a compile error under
Cygwin 1.7.1. Thanks to Jill McCutcheon for the bug report.
- 0.8.0 (2 Mar 2010) –
- Fixed message queue support detection in FreeBSD and
the platform-specific documentation about FreeBSD.
- Rearranged the documentation and split the history
(which you're reading now) into a separate file.
- I fixed two small bugs related to the confusing
message queue constants. The bugs and associated changes are
explained below. The explanation is really long not
because the changes were big (they weren't), but because
they and rationale behind them are subtle.
Fixing these bugs was made easier by this realization:
on all of the systems to which I have access that implement
message queues (FreeBSD, OpenSolaris, Linux, and Windows +
Cygwin), all except Linux implement them as
memory-mapped files or something similar. On these
non-Linux systems, the
maximum queue message count and size are pretty darn big
(LONG_MAX). Therefore, only on Linux is anyone likely to
encounter limits to message queue size and content.
The first bug I fixed was related to four message queue
constants mentioned in posix_ipc documentation:
QUEUE_MESSAGE_SIZE_MAX_DEFAULT. All four were defined
in the C
code, but the two XXX_DEFAULT constants weren't exposed on
the Python side.
The second bug was that under Linux, QUEUE_MESSAGES_MAX and
QUEUE_MESSAGE_SIZE_MAX were permanently fixed to their
values at posix_ipc's compile/install time even if the
relevant system values changed later. Thanks to Kyle Tippetts
for bringing this to my attention.
QUEUE_MESSAGES_MAX_DEFAULT was arbitrarily limited to
(at most) 1024. This wasn't a bug, just a bad choice.
I made a few changes in order to fix these problems –
- The constants QUEUE_MESSAGES_MAX and
have been deleted since they were only sure to
be accurate on systems where they were irrelevant. Furthermore,
Linux (the only place where they matter) exposes these values
through the file system (in
/proc/sys/fs/mqueue/msgsize_max respectively) so Python
apps that need them can read them without any help
- QUEUE_MESSAGES_MAX_DEFAULT and
QUEUE_MESSAGE_SIZE_MAX_DEFAULT are now exposed to
Python as they should have been all along.
QUEUE_MESSAGES_MAX_DEFAULT is now set to
LONG_MAX on all platforms except Linux, where
it's set at compile time from /proc/sys/fs/mqueue/msg_max.
- QUEUE_MESSAGE_SIZE_MAX_DEFAULT remains at the fairly
arbitrary value of 8k. It's not a good idea to make it too big
since a buffer of this size is allocated every time
MessageQueue.receive() is called. Under Linux, I
check the contents of /proc/sys/fs/mqueue/msgsize_max
and make QUEUE_MESSAGE_SIZE_MAX_DEFAULT smaller if
- 0.7.0 (21 Feb 2010) –
Added Python 3.1 support.
- 0.6.3 (15 Feb 2009) –
- 0.6.2 (30 Dec 2009) –
Fixed a bug where a MessageQueue's mode
attribute returned garbage. Grazie to Stefano Debenedetti for
the bug report.
- 0.6.1 (29 Nov 2009) –
There were no functional changes to the module in this version, but
I added the convenience function close_fd() and fixed
some docmentation and demo bugs/sloppiness.
- Added the convenience function SharedMemory.close_fd().
Thanks to Kyle Tippetts for pointing out the usefulness
- Added the module attributes __version__,
__copyright__, __author__ and
- Fixed the license info embedded in posix_ipc_module.c
which was still referring to GPL.
- Replaced file() in setup.py with
- Demo changes –
- Made the demo a bit faster, especially for large
shared memory chunks. Thanks to Andrew Trevorrow
for the suggestion and patch.
- Fixed a bug in premise.c; it wasn't closing the semaphore.
- Fixed a bug in premise.py; it wasn't closing the
shared memory's file descriptor.
- Fixed bugs in conclusion.py; it wasn't closing the
shared memory's file descriptor, the semaphore or
- 0.6 (5 Oct 2009) –
- Relicensed from the GPL to a BSD license to celebrate the
one year anniversary of this module.
- Updated Cygwin info.
- 0.5.5 (17 Sept 2009) –
- Set MQ_MAX_MESSAGES and MQ_MAX_MESSAGE_SIZE to
LONG_MAX under cygwin.
(Danke to René Liebscher.)
- Surrounded the #define PAGE_SIZE in probe_results.h with
#ifndef/#endif because it is already defined on some systems.
(Danke to René Liebscher, again.)
- Minor documentation changes.
- 0.5.4 (21 Jun 2009) –
- 0.5.3 (8 Mar 2009) –
- Added automatic generation of names.
- Changed status to beta.
- 0.5.2 (12 Feb 2009) –
- Fixed a memory leak in MessageQueue.receive().
- Fixed a bug where the name of the MessageQueue
current_messages attribute didn't match the name
given in the documentation.
- Added the VERSION attribute to the module.
- Fixed a documentation bug that said message queue
notifications were not yet supported.
- 0.5.1 (8 Feb 2009) –
- Fixed outdated info in setup.py that was showing up
in the Python package index. Updated README while I
was at it.
- 0.5 (8 Feb 2009) –
- Added the message queue notification feature.
- Added a mode attribute to each type.
- Added str() and repr() support to
- Added a demo for message queues.
- Fixed some minor documentation problems and added
some information (esp. about Windows + Cygwin).
- 0.4 (9 Jan 2009) –
- Added message queue support.
- Fixed the poor choices I'd made for names for classes and
errors by removing the leading "Posix" and "PosixIpc".
- Simplified the prober and expanded it (for message
- Cleaned up this documentation.
- 0.3.2 (4 Jan 2009) –
- Fixed an uninitialized value passed to PyMem_Free() when
invalid params were passed to either constructor.
- 0.3.1 (1 Jan 2009) –
- Fixed a big bug where the custom exceptions defined by this
module weren't visible.
- Fixed a compile complaint about the redefinition of
SEM_VALUE_MAX on Linux (Ubuntu) that I introduced
in the previous version.
- Fixed a bug in the demo program premise.c where I wasn't
closing the file descriptor associated with the shared
- Added the PAGE_SIZE attribute. This was already
available in the mmap module that you need to use shared
memory anyway, but adding it makes the interface more
consistent with the sysv_ipc module.
- 0.3 (19 Dec 2008) –
- Added informative custom errors instead of raising
OSError when something goes wrong.
- Made the code friendly to multi-threaded applications.
- Added the constants O_CREX and
- Added code to prohibit negative timeout values.
- 0.2 (4 Dec 2008) –
- Removed the un-Pythonic try_acquire() method. The
same functionality is now available by passing a timeout of
0 to the .acquire() method.
- Renamed the module constant ACQUIRE_TIMEOUT_SUPPORTED to
- Moved the demo code into its own directory and added C
versions of the Python scripts. The parameters are now in a
text file shared by the Python and C program, so you can
run the C version of Mrs. Premise and have it communicate with
the Python version of Mrs. Conclusion and vice versa.
- 0.1 (9 Oct 2008) – Original (alpha) version.