Open Hub
Open Hub

 Settings |  Report Duplicate

10
I Use This!
Inactive

Commits : Listings

Analyzed 21 days ago. based on code collected 21 days ago.
Jan 29, 2024 — Jan 29, 2025
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Update error msg to explain why the createSeqRefused fault is sent so as to aid diagnosis.
parsonsd
More... almost 15 years ago
Update error msg to explain why the createSeqRefused fault is sent so as to aid diagnosis.
David Parsons
More... almost 15 years ago
commit the patch for SANDESHA2-199
Amila Chinthaka Suriarachchi
More... almost 15 years ago
commit the patch for SANDESHA2-199
amilas
More... almost 15 years ago
Moving axis svn, part of TLP move INFRA-2441
gmcdonald
More... about 15 years ago
Moving axis svn, part of TLP move INFRA-2441
Gavin McDonald
More... about 15 years ago
Further update related to the last fix. A null check was added as Sync RM 1.0 msgCtx's may not have a to address.
David Parsons
More... about 15 years ago
Fix to Sandesha as problem highlighted after move to recent Axis2 release. The thread ABORT logic has changed slightly in Axis2 so was causing NPE's when an RM inorder response was being delivered to a client. The fix is to use a thread SUSPEND for inorder Sync responses.
David Parsons
More... about 15 years ago
Improvements to error logging of Sequence Reallocation
David Parsons
More... about 15 years ago
Fix added so that duplicate Seq ID's won't get added to either inmemory or permanent stores.
David Parsons
More... over 15 years ago
add transport dependencies
davanum
as Davanum Srinivas
More... over 15 years ago
better set of dependencies from Rick Slade
davanum
as Davanum Srinivas
More... over 15 years ago
add missing transports - fix for test failure
davanum
as Davanum Srinivas
More... over 15 years ago
fix the artifact id
davanum
as Davanum Srinivas
More... over 15 years ago
So that Sandesha is compliant with the ReliableSecureProfile spec. Piggybacked acks must have the mustUnderstand flag set to false.
David Parsons
More... over 15 years ago
Fix for failing Sync2Way RSP interop tests
David Parsons
More... over 15 years ago
remove the generated code and change the test senarios to suite for current axis2 generated code
Amila Chinthaka Suriarachchi
More... over 15 years ago
remove the tests which check whether Make connection is there for RM 1.1 in out scenarios
Amila Chinthaka Suriarachchi
More... over 15 years ago
General Performance Improvements
David Parsons
More... over 15 years ago
return the thread always if it is getting the response at the client side. there was a bug not returing 202 Accepted for response messages for dual channel case
Amila Chinthaka Suriarachchi
More... over 15 years ago
apply the patch for SANDESHA2-198
Amila Chinthaka Suriarachchi
More... over 15 years ago
setting the rms annonymous address to close sequence as well
Amila Chinthaka Suriarachchi
More... over 15 years ago
Fix committed for JIRA SANDESHA2-196. Thanks Rick
David Parsons
More... over 15 years ago
The matching of the primitive type reallocated field happened whether or not the field had been set in the matcher RMSBean. I've changed it so that it only matches upon request i.e. the field is set in the matcher bean.
David Parsons
More... over 15 years ago
Added reallocated field to match method in RMSBean. Thanks Francois Mireaux for pointing this out.
David Parsons
More... over 15 years ago
- When an inorder RM system is under load msgs aren't always being delivered to applications as the InvokerWorker currently working the seq has the lock on the seq but has decided that there are no more msgs to process as the SequenceProcessor thread hasn't quite got as far as inserting a new InvokerBean. This means neither thread does the work of delivering the msg to the application. The fix is to move the removing of the lock on the sequence to before we commit the tran used to find the InvokerBean in the InvokerWorker.
David Parsons
More... over 15 years ago
Fix committed for JIRA SANDESHA2-192. Thanks Rick.
David Parsons
More... over 15 years ago
- Msgs are being lost because RMDBeans are being updated to say msgs have been received before complete insertion of InvokerBeans. The fix is to make the updating and insertion complete under one transaction. - The sync1way reallocation code isn't triggered because the client doesn't receive any faults from the server. The fix is to set the faultTo so wsa has somewhere to send the fault.
David Parsons
More... over 15 years ago
applied the patch for SANDESHA2-191
Amila Chinthaka Suriarachchi
More... almost 16 years ago
- Fix newly added reallocation code as deadlock had been introduced. - Changed RMSBean to store EndpointReference for offered Endpoints so ref params don't get lost - Update so that all backchannel msgs on MC's are processed successfully - Commented out sending of AckRequested msgs as not needed as yet and impacts performance
David Parsons
More... almost 16 years ago