Afternoon all, I'm looking to merge the feature-visual-studio-updates branch into next, since it's been around a while and had what seems like a reasonable amount of community testing. Anyone any issu

Including test strategy: This is not OS dependant This is not middleware dependant If you build MAMA you will have mamalistencachedc and mamalistencachedcpp in the bin directory. Try running these up

From feb963ab29ec5423822307b18b866f0000fbefd6 Mon Sep 17 00:00:00 2001 From: Gary Molloy <gmolloy@...> Date: Wed, 2 Apr 2014 16:28:43 +0100 Subject: [PATCH] OpenSource MamaFieldCache for C and CPP Ope

Hi All, There have been a number of discussions (and code contributions) around making the bridge loading more flexible and dynamic, so I'd like to kick off a working group to hammer this out and get

Hi List, Attached is a patch fixing a memleak in mamaMsgImpl_setPayload. impl->payload is not checked for a owned previous payload before the assignment, leading to a memory leak, in particular was sh

From 9e5cc88d7d48adf57796f0dba35c4c5d34aa8eae Mon Sep 17 00:00:00 2001 From: Gary Molloy <gmolloy@...> Date: Mon, 27 Jan 2014 14:58:21 +0000 Subject: [PATCH] Title: Refresh acknowledgement processing

Fix in Darwin Semaphore. Count variable was not being checked before decrementing, which was causing issues in managing the queue size, when running high loads. Further change is to use Barrier versio

This change moves file parsing capabilities to an external module. This makes the code more readable and modular as well as allowing us to remove the file streams macro which was temporarily added to

The new internal fileparser module allows the caller to open a file in either mmap or file stream mode. The module will then allow the caller to read through the file and copy the output to a provided

This change adds mman.h and mman.c to the code base as licensed by the MIT license from https://code.google.com/p/mman-win32/. It also removes mmap.h and updates windows/wombat/port.h to use the new m

Hi I have been looking at adding a plugin loader functionality to MAMA. The idea being that it would a simple standard interface to add extra functionality alongside MAMA. The functionality is limited

- The roundtrip serialization unit test constructs a reasonably representative payload with most of the payload types - It then serialises and unserialises this into a second payload and then extracts

- Reversed all EXPECT_*, ASSERT_* where the actual and expected where in the format of EXPECT_*(actual, expected) instead of the suggested format of EXPECT_*(expected, actual). Doesn't impact function

- Fixed two types of leaks for reserved fields: (1) MamaCPP: Fields weren't being destroyed on the final (or any) Mama::close(). (2) MamaCPP: Field initialisation was called on every Mama::open(), cau

Fixed memory leak when calling dqPublisherManager_destroy() - The contents of the publisher topic map are now freed inline with the publisher map itself being destroyed. Signed-off-by: Lee Skillen <ls

Changed the mIsDispatching signalling variable into an atomic int since its visible, inspected and modified across thread boundaries. This is the cause of several errors in helgrind and has been obser

The queue in common code now uses an atomic (wInterlockedInt) variable for tracking whether or not the queue is being unblocked (which is often as a result of requiring to shut the queue down). A race

- When destroying the timers heap, the dispatching thread is now joined and the socket pair (pipe) is closed, freeing the resources for both. - When creating a timer, if the timer that has been passed

Libraries were delimited using semi-colons in 2005 which worked fine, but in import to MSVC 2010, it dislikes these and cowardly refuses to interpret them. Frank Quinn <fquinn@...> --- mama/c_cpp/src/

Hey This is a small tidy up for the SCons multithreading implementation. The native SCons –j option was essentially being overloaded, which isn’t really the clearest solution. This patch adds a “jobs”