[shmem] New Shmem snapshot in Boost Vault

Hi to all, I've just uploaded a new Shmem snapshot to Boost Vault. This version has some interesting new features so that users can use Shmem to build more complex IPC mechanisms. The documentation is also available in the usual place: http://ice.prohosting.com/newfunk/boost/libs/shmem/index.html New features: -> Some bugfixes as always -> Named objects can be now destroyed via pointer without using the name -> More internal checks to avoid incorrect types, bad alignments or bad pointers. -> Added new STL-like allocator: cached_node_allocator. This allocator has a configurable private cache of nodes. The halfway between node_allocator and private_node_allocator. -> Atomic shared memory creation/destruction with optional user initialization/destruction functor. This allows the design of more advanced IPC mechanisms. -> Shared memory creation and destruction unification in Unix/Windows. When last connected process calls "close()", the shared memory is destroyed. -> Memory mapped file wrapper and Shmem adapted to work with memory mapped files. Now you can get automatic persistence of the classes you create in memory. -> New inter-process mechanism: A process-shared message queue. This message queue has priority-based message reception, and can work in blocking, non-blocking and timed way. This is a practical example of how Shmem classes can be used to implement higher order IPC mechanisms. That's all. All comments and suggestions are welcome! Ion

On 7/8/05, Ion Gaztañaga <ion_g_m@terra.es> wrote: Small typo in the docs I just noticed: ... Allocating dinamically ... Should be dynamically. -- Caleb Epstein caleb dot epstein at gmail dot com

Thanks, I forgot to correct your previous doc comments for this version, but they will be ready for the next. Thanks again! Ion ----- Original Message ----- From: "Caleb Epstein" <caleb.epstein@gmail.com> To: <boost@lists.boost.org> Sent: Saturday, July 09, 2005 4:23 AM Subject: Re: [boost] [shmem] New Shmem snapshot in Boost Vault
On 7/8/05, Ion Gaztañaga <ion_g_m@terra.es> wrote:
Small typo in the docs I just noticed:
... Allocating dinamically ...
Should be dynamically.
-- Caleb Epstein caleb dot epstein at gmail dot com
participants (2)
-
Caleb Epstein
-
Ion Gaztañaga