|
[PATCH 20/50] [mama] transport write queue high and low watermarks
From: Mike Schonberg <mschonberg@...>
Allow applications to set high and low watermarks for internal write queues if
the underlying middleware queues writes. Socket based middlewares that
From: Mike Schonberg <mschonberg@...>
Allow applications to set high and low watermarks for internal write queues if
the underlying middleware queues writes. Socket based middlewares that
|
By
Michael Schonberg <mschonberg@...>
·
#130
·
|
|
[PATCH 19/50] [mama] Create an Internal Transport for Monitoring
From: Mike Schonberg <mschonberg@...>
OpenMAMA creates a dedicated internal transport for publishing statistics for
monitoring and other internal communication. This transport disables
From: Mike Schonberg <mschonberg@...>
OpenMAMA creates a dedicated internal transport for publishing statistics for
monitoring and other internal communication. This transport disables
|
By
Michael Schonberg <mschonberg@...>
·
#129
·
|
|
[PATCH 18/50] [mama] Subscription: don't put invalid actions on the throttle queue
From: Mike Schonberg <mschonberg@...>
Added code to avoid race conditions when enqueing shutdown logic on the
throttle.
Signed-off-by: Mike Schonberg <mschonberg@...>
---
From: Mike Schonberg <mschonberg@...>
Added code to avoid race conditions when enqueing shutdown logic on the
throttle.
Signed-off-by: Mike Schonberg <mschonberg@...>
---
|
By
Michael Schonberg <mschonberg@...>
·
#128
·
|
|
[PATCH 17/50] [mama] Configurable Schemes for data quality
From: Mike Schonberg <mschonberg@...>
The dqStratgyScheme deterines whether or not OpenMAMA ignores duplicate
messages. Valid values are:
DQ_SCHEME_DELIVER_ALL deliver all messages including
From: Mike Schonberg <mschonberg@...>
The dqStratgyScheme deterines whether or not OpenMAMA ignores duplicate
messages. Valid values are:
DQ_SCHEME_DELIVER_ALL deliver all messages including
|
By
Michael Schonberg <mschonberg@...>
·
#127
·
|
|
[PATCH 16/50] [mama] dqpublisher destroy
From: Mike Schonberg <mschonberg@...>
Added proper destroy implementation for dqpublisher to allow applications to
clean up and shutdown properly.
Signed-off-by: Mike Schonberg
From: Mike Schonberg <mschonberg@...>
Added proper destroy implementation for dqpublisher to allow applications to
clean up and shutdown properly.
Signed-off-by: Mike Schonberg
|
By
Michael Schonberg <mschonberg@...>
·
#126
·
|
|
[PATCH 15/50] [mama] dqpublisher manager bug fixes
From: Mike Schonberg <mschonberg@...>
Only invoke the application on create callback if it is not NULL. This avoids
core dumps if the application does not provide a callback.
The manager was not
From: Mike Schonberg <mschonberg@...>
Only invoke the application on create callback if it is not NULL. This avoids
core dumps if the application does not provide a callback.
The manager was not
|
By
Michael Schonberg <mschonberg@...>
·
#125
·
|
|
[PATCH 14/50] [mama] Add cache and closure to dqpublisher
From: Mike Schonberg <mschonberg@...>
These are both application supplied callbacks. The closure is intended for
application specific context for the publisher and the cache is for caching
data
From: Mike Schonberg <mschonberg@...>
These are both application supplied callbacks. The closure is intended for
application specific context for the publisher and the cache is for caching
data
|
By
Michael Schonberg <mschonberg@...>
·
#124
·
|
|
[PATCH 13/50] [mama] mamaDQPublisher_sendReplyWithHandle
From: Mike Schonberg <mschonberg@...>
This method allows applications to use a reply handle rather than detaching and
holding on to the original message. Since OpenMAMA attempts to reuse
From: Mike Schonberg <mschonberg@...>
This method allows applications to use a reply handle rather than detaching and
holding on to the original message. Since OpenMAMA attempts to reuse
|
By
Michael Schonberg <mschonberg@...>
·
#123
·
|
|
[PATCH 12/50] [mama][dqpublisher] Improved Message Handling
From: Mike Schonberg <mschonberg@...>
Since the caller may pass messages with the reserved files (ie. senderid,
seqno, etc.) populated, the dq publisher should try to be robust and update
From: Mike Schonberg <mschonberg@...>
Since the caller may pass messages with the reserved files (ie. senderid,
seqno, etc.) populated, the dq publisher should try to be robust and update
|
By
Michael Schonberg <mschonberg@...>
·
#122
·
|
|
[PATCH 11/50] [mama] minor datetime fixes
From: Mike Schonberg <mschonberg@...>
When formatting a date time with a time, separate the date and time with a
space.
Added parameter checking to makeTime to handle invalid values sanely: set
From: Mike Schonberg <mschonberg@...>
When formatting a date time with a time, separate the date and time with a
space.
Added parameter checking to makeTime to handle invalid values sanely: set
|
By
Michael Schonberg <mschonberg@...>
·
#121
·
|
|
[PATCH 10/50] [mama] Free CM Subscription in Destroy Callback
From: Mike Schonberg <mschonberg@...>
The CM subscription can not be freed until it destruction completes and invokes
the onDestroy() callback. This avoids potential core dumps on
From: Mike Schonberg <mschonberg@...>
The CM subscription can not be freed until it destruction completes and invokes
the onDestroy() callback. This avoids potential core dumps on
|
By
Michael Schonberg <mschonberg@...>
·
#120
·
|
|
[PATCH 09/50] [bridge] Added mamaTransport_forceClientDisconnect()
From: Mike Schonberg <mschonberg@...>
For TCP socket based middleware, allow transports to disconnect individual
clients. This is usefull for eliminating slow consumers or other applications
that
From: Mike Schonberg <mschonberg@...>
For TCP socket based middleware, allow transports to disconnect individual
clients. This is usefull for eliminating slow consumers or other applications
that
|
By
Michael Schonberg <mschonberg@...>
·
#119
·
|
|
[PATCH 08/50] [bridge] Timout for Stopping Internal Queue
From: Mike Schonberg <mschonberg@...>
When shutting down, don't wait forever for the internal queue to stop. If the
queue does not stop within the default queue timeout
(mama.defaultqueue.timeout
From: Mike Schonberg <mschonberg@...>
When shutting down, don't wait forever for the internal queue to stop. If the
queue does not stop within the default queue timeout
(mama.defaultqueue.timeout
|
By
Michael Schonberg <mschonberg@...>
·
#118
·
|
|
[PATCH 07/50] [windows] Macros for MSVC Version
From: Mike Schonberg <mschonberg@...>
These are necessary to accomodate different version of Visual Studio.
Signed-off-by: Mike Schonberg <mschonberg@...>
---
From: Mike Schonberg <mschonberg@...>
These are necessary to accomodate different version of Visual Studio.
Signed-off-by: Mike Schonberg <mschonberg@...>
---
|
By
Michael Schonberg <mschonberg@...>
·
#117
·
|
|
[PATCH 06/50] Adding new files for C++ and Windows Support
From: Mike Schonberg <mschonberg@...>
These files include support for uuid's required by avis for request reply. Also
adding wMessageStats for C++ and MAMDA.
Signed-off-by: Mike Schonberg
From: Mike Schonberg <mschonberg@...>
These files include support for uuid's required by avis for request reply. Also
adding wMessageStats for C++ and MAMDA.
Signed-off-by: Mike Schonberg
|
By
Michael Schonberg <mschonberg@...>
·
#160
·
|
|
[PATCH 05/50] Make includes paths consistent
From: Mike Schonberg <mschonberg@...>
wConfig.h and port.h should be included as wombat/xxx.h. This simplifies the
builds and makes the consistent across Windows and Linux.
Signed-off-by: Mike
From: Mike Schonberg <mschonberg@...>
wConfig.h and port.h should be included as wombat/xxx.h. This simplifies the
builds and makes the consistent across Windows and Linux.
Signed-off-by: Mike
|
By
Michael Schonberg <mschonberg@...>
·
#116
·
|
|
[PATCH 04/50] Added WCACHEExpDLL macro for building NYSE wirecache
From: Mike Schonberg <mschonberg@...>
This is all part of DLL hell for windows: when we build common we need to
indicate that the mehtods from common are export from the DLL, and then we need
to
From: Mike Schonberg <mschonberg@...>
This is all part of DLL hell for windows: when we build common we need to
indicate that the mehtods from common are export from the DLL, and then we need
to
|
By
Michael Schonberg <mschonberg@...>
·
#115
·
|
|
[PATCH 03/50] Fix timer implementation for Windows
From: Mike Schonberg <mschonberg@...>
Address subtle difference between Linux and Windows networking required by
timers. Since there are no pipes on Windows, the timer implementation use
From: Mike Schonberg <mschonberg@...>
Address subtle difference between Linux and Windows networking required by
timers. Since there are no pipes on Windows, the timer implementation use
|
By
Michael Schonberg <mschonberg@...>
·
#114
·
|
|
[PATCH 02/50] c linkage for wlock_xxx() methods
From: Mike Schonberg <mschonberg@...>
These must be declared with extern "C" to link correctly with C++ applications.
Signed-off-by: Mike Schonberg <mschonberg@...>
---
From: Mike Schonberg <mschonberg@...>
These must be declared with extern "C" to link correctly with C++ applications.
Signed-off-by: Mike Schonberg <mschonberg@...>
---
|
By
Michael Schonberg <mschonberg@...>
·
#113
·
|
|
[PATCH 01/50] Ownership of message buffers
From: Mike Schonberg <mschonberg@...>
The msg payload bridges now have a msgPayloadSetParent() call which is
used to set the parent mamaMsg object for a payload. This allows the
"ownership" of
From: Mike Schonberg <mschonberg@...>
The msg payload bridges now have a msgPayloadSetParent() call which is
used to set the parent mamaMsg object for a payload. This allows the
"ownership" of
|
By
Michael Schonberg <mschonberg@...>
·
#112
·
|