11:17 a.m.
Hi Thomas,
Here you have some answers (in blue) to your complains:
Issue 1: startProcess is semi-synchronous. It starts the process, runs
through as many nodes as it can until it gets to a non-synchronous task
(rules, long lived customer work item etc) at which point it returns – or it
returns when the entire process has been completed.
Challenge 1a: How to detect when the process has finished and use it in a
synchronous manner.
Solution: A ProcessEventListener which completes a countdown latch when teh
process has completed.
Challenge 1b: Subprocesses also trigger process events
Solution: Have the latch track the first process that started (the parent)
and only return when that process completes.
Another solution: after startProcess() returns, you can check the status of
the process (unfortunately, you need a cast):
WorkflowProcessInstance processInstance =
(WorkflowProcessInstance) ksession.startProcess("org.drools.state");
// should be in state A
assertEquals(ProcessInstance.STATE_ACTIVE,
processInstance.getState());
Issue 2: Drools Flow is littered with error checking code which just prints
to System.err and either carrys on or just stop. Eg if a work item handler
isn’t registered the process just stalls with a message.
Solution: Make sure code is perfectly written and doesn’t have any bugs...
hope that it doesn’t happen in production ... complain on mailing list and
add TODO to raise issue.
You are absolutely true about this. We are trying to remove all the
System.err and replace them with a proper error treatment (like abort the
process)
Issue 3: Fault nodes seem to be the correct way to say that something has
gone wrong and the process should terminate.
Challenge: How does the calling code actually work out what the fault was
and get the variable? The guide talks very very briefly about exception
handlers but from what I can understand this is a way for the process to
define what to do in this conditions. I want the process to be aborted and
the calling application to be informed of this and the details.
Solution: Add an event listener which detects when a fault node is being
entered and extracts the details – seems like a big hack surely there is a
better way.
Another Solution: you could throw an exception inside your exception
handler. This way you will abort the process and get the exception in your
code.
Issue 4: If a child process executes a fault node it is aborted. The parent
process is set to wait for completion and has the child node as not being
independent so I would expect that the parent process should also be marked
as aborted.
Challenge: The parent process isn’t marked as complete – it still has a
status of Executing. However the process is stalled. Looking at
subprocessNodeInstance internalTrigger (line 117) if the node returns a
state of completed then trigger completed is called, otherwise it just adds
a process listener – which will never be called as the process has already
aborted. The parent process just stalls left in the state of active forever.
Sollution: ?? Hack my fault listener to work round the issue and update the
parent process, make the processEventListener assert that if a child process
is aborted then we should return? Ask on the mailing list to get me out of a
blind alley and point me to the correct drools way of doing it...
Another solution: Throwing an exception inside the handler should solve this
issue too.
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Esteban Aliverti
- Developer @ http://www.plugtree.com
- Blog @ http://ilesteban.wordpress.com
2010/10/15 Swindells, Thomas <TSwindells(a)nds.com>
I’ve just started to use Drools Flow in anger and I’m really
struggling
to understand the correct way to use it and think I am missing something.
This is what I want to achieve:
The app code executes a StartProcess command and waits for the process to
return (1 main process per session, 1 session per thread).
The process either succeeds, or it fails with a result.
A process could be complicated so I want to split it up into a parent
process calling subprocesses.
I want to minimize boilerplate code (particularly in the ruleflow files)
and keep the workflows as simple and maintainable as possible.
It sounds simple enough and doesn’t seem to be particularly unusual.
Issue 1: startProcess is semi-synchronous. It starts the process, runs
through as many nodes as it can until it gets to a non-synchronous task
(rules, long lived customer work item etc) at which point it returns – or it
returns when the entire process has been completed.
Challenge 1a: How to detect when the process has finished and use it in a
synchronous manner.
Solution: A ProcessEventListener which completes a countdown latch when teh
process has completed.
Challenge 1b: Subprocesses also trigger process events
Solution: Have the latch track the first process that started (the parent)
and only return when that process completes.
Issue 2: Drools Flow is littered with error checking code which just prints
to System.err and either carrys on or just stop. Eg if a work item handler
isn’t registered the process just stalls with a message.
Solution: Make sure code is perfectly written and doesn’t have any bugs...
hope that it doesn’t happen in production ... complain on mailing list and
add TODO to raise issue.
Issue 3: Fault nodes seem to be the correct way to say that something has
gone wrong and the process should terminate.
Challenge: How does the calling code actually work out what the fault was
and get the variable? The guide talks very very briefly about exception
handlers but from what I can understand this is a way for the process to
define what to do in this conditions. I want the process to be aborted and
the calling application to be informed of this and the details.
Solution: Add an event listener which detects when a fault node is being
entered and extracts the details – seems like a big hack surely there is a
better way.
Issue 4: If a child process executes a fault node it is aborted. The parent
process is set to wait for completion and has the child node as not being
independent so I would expect that the parent process should also be marked
as aborted.
Challenge: The parent process isn’t marked as complete – it still has a
status of Executing. However the process is stalled. Looking at
subprocessNodeInstance internalTrigger (line 117) if the node returns a
state of completed then trigger completed is called, otherwise it just adds
a process listener – which will never be called as the process has already
aborted. The parent process just stalls left in the state of active forever.
Sollution: ?? Hack my fault listener to work round the issue and update the
parent process, make the processEventListener assert that if a child process
is aborted then we should return? Ask on the mailing list to get me out of a
blind alley and point me to the correct drools way of doing it...
Thomas
------------------------------
**************************************************************************************
This message is confidential and intended only for the addressee. If you
have received this message in error, please immediately notify the
postmaster(a)nds.com and delete it from your system as well as any copies.
The content of e-mails as well as traffic data may be monitored by NDS for
employment and security purposes. To protect the environment please do not
print this e-mail unless necessary.
NDS Limited. Registered Office: One London Road, Staines, Middlesex, TW18
4EX, United Kingdom. A company registered in England and Wales. Registered
no. 3080780. VAT no. GB 603 8808 40-00
**************************************************************************************
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users