updateTask task.getTaskId has wrong value

[GoogleCodeExporter]

What steps will reproduce the problem?
1. We have our own call system with a lot of open tasks.
2. We have created a fragment plugin of the industrial core, with extension to 
the persistor.
3. I have a set of queries, with a total of like 300 (non-distinct (overlap)) 
tasks.

What is the expected output? What do you see instead?
From the moment we now want to update a task (eg. T01) and press the submit 
button, another unrelated task (eg. T99) gets updated. I have to note that that 
task is somewhere present in another query.
When I debug into updateTask and put a watch on I see the following:
- all task fields have the expected values of task T01
- but task.taskId has the wrong value and had value T99
It's like the connector is trying to do 2 requests at the same time. 
- updateTask for T01
- fetchTask for T99
and somehow those requests start to interfere with each other.

What version of the product are you using? On what operating system?
I'm using eclipse 3.6 on a Windows environment with the Industrial SQL 
Connector for Mylyn installed from the Marketplace.

Please provide any additional information below.
It might be interesting to know that I didn't have any problems until I created 
more queries...

Original issue reported on code.google.com by bert.str...@gmail.com on 10 Mar 2011 at 4:38

[GoogleCodeExporter]

Original comment by meijerf...@gmail.com on 31 Mar 2011 at 9:24

• Changed state: Accepted

[GoogleCodeExporter]

Did you use JPA or Ibatis based connector? 
When Ibatis I suspect it maybe because 
com.industrialtsi.mylyn.core.persistence.TasksSqlMapConfig.getSqlMapInstance(Tas
kRepository) is not synchronised.
See this article on singleton factory pattern:
http://java.sun.com/developer/technicalArticles/Programming/singletons/

Original comment by meijerf...@gmail.com on 1 Apr 2011 at 1:18

• Changed state: Started

[GoogleCodeExporter]

I didn't use both JPA or Ibatis. I started out with this webpage:
http://wiki.eclipse.org/Mylyn/Incubator/Generic_Industrial_Connector/Configuring
_Industrial_Connector_using_Nothing

Original comment by bert.str...@gmail.com on 5 Apr 2011 at 12:53

[GoogleCodeExporter]

I still suspect a synchronization error, but I cannot see the code you created.

Original comment by meijerf...@gmail.com on 5 Apr 2011 at 7:54

[GoogleCodeExporter]

I have the same or very similar problem using the "Nothing" technique with a 
web service on the back end. During a full sync of a query, which takes a long 
time (200+ tasks) if I open a task and synchronize it alone then its contents 
becomes that of a later task which appears to be somewhere close to where the 
full sync has gotten to. So if I resync task 001 and the query sync is up to 
task 100 then I get the contents of task 100 in task 001. Seen this happen 
multiple times. 

Thanks for making this connector available by the way.

Original comment by pdegrego...@pgmsystems.com on 16 Dec 2011 at 9:38