We do it also like this:
a decision before each subpath of execution which moves directly to the join if
the condition is not met.
A custom fork... ah... ye... several times mentioned in the forum.
Maybe it is time to have a "standard" jBpm fork that does this trick :-), or to
have guar
Yeah, also had a hard time debugging why conditions did not work. Now it is on
our tips and tricks for jBpm wiki:
DON'T FORGET TO NAME THAT TRANSITION :-)
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3948265#3948265
Reply to the post :
http://www.jboss.co
strange, a join without a fork (or did I miss something)
Are you putting "less" in an Amount variable? (as a string)
do you do a "taskinstance.end()"?
Try having a tasknode with attribute signal="last-wait"
James
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=
All process (and task) variables are instantly available.
So you can do
| #{decisionVariable eq 'less'}
|
See source code of JbpmVariableResolver.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3942396#3942396
Reply to the post :
http://www.jboss.com/i
I also think you should do tok.end(false) and not tok.end() (which seems
dangerous). Will it be fixed? I'm regularly patching/extending the jBpm core.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3941477#3941477
Reply to the post :
http://www.jboss.com/inde
Relatively simple by having handlers creating/adding all sorts of vars like
this (package and imports not shown):
| public class VariableActionHandler implements ActionHandler {
|
| private String name;
|
| private String method;
|
| private String value;
|
Situation:
Fork with two legs, each having a tasknode with a task, each with its own
swimlane (so two different users could get a task in their worklist).
What happens if two users end their task simultaneously? Will the join execute
properly? After all, they fetch the current situation, do a tra
In jbpm 3.1 and our extensions, it works too :-)
We use the command pattern and strategy pattern heavily. Each node and task can
lookup the best strategy to perform assignments, execution, priority, swimlane,
concerned datablock, etc...
The lookups are by node-name, or a specified name on the ev
I don't get it exactly, but if you want to configure a process variable using
the definition, it is best to make a base class (I created an adapter handling
assignment, events and decision), having some general-purpose properties to use
e.g. a set of name-value pairs that will be created as vari
In Ghent, it was the infamous/notorious Rita De Caluwe
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3940501#3940501
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3940501
--
I feel neglected :-)
http://www.jboss.com/index.html?module=bb&op=viewtopic&t=80727
I think that, if you can map it in hibernate, it will be fairly easy. You have
to tell hibernate to always fetch the latest version (this may be difficult,
but you could have a "latest version" mapping, based on
any ideas on this matter?
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3938621#3938621
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3938621
---
Using Tomcat but ne
yeah, by implementing my own stack pointer (datastructures and algorithms 1,
following the 7-steps rule)
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3938620#3938620
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=393
We just made "undo" transitions were applicable, because our "undo's" are
business-related undo's.
If you need something like a real compensation, you could see if you could get
away with an exception handler, which parks the process in a custom node.
View the original post :
http://www.jboss.
If I deploy a process as a new version, it will not be used by processes using
this process as a subprocess. This means I have to redeploy all superprocesses
using this process. Is this desired behaviour?
E.g. the following business case:
we have a verification process that will be used by other
How can I specify a processdefinition that uses itself as a subprocess? Will
this deploy correctly?
James
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3935711#3935711
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=39
I hope the message was clear. We design by interfaces, we inject by interfaces.
So using a class instead of a defined interface totally breaks this concept.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3928949#3928949
Reply to the post :
http://www.jboss.co
PersistenceService:
why is the interface sometimes used, and sometimes the DBPersistenceService? If
you cast directly to DBPersistenceService, interfaces are of little use
(especially if you depend on them to make them configurable).
Now I have to make a subclass of DBPersistenceService to avoid
criess,
what will happen if, from within the jbpm code, a call to
JbpmConfiguration.getInstance() is made? Will you end up with the same
configuration as defined in spring?
James
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3927050#3927050
Reply to the p
The contract of the swimlane is clear afaik:
"A swimlane is a process role. It is a mechanism to specify that multiple tasks
in the process should be done by the same actor. So after the first task
instance is created for a given swimlane, the actor should be remembered in the
process for all su
see also http://www.jboss.com/index.html?module=bb&op=viewtopic&t=76546, but I
guess your version is more elaborated.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3926033#3926033
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=
getMethod(s) only give back public method(s).
If you specify in the contract you need "a" setter (with any visibility), then
you have to implement it, of course.
On the other hand, if you specify in your contract that the setter should be
public, you can get away with it :-).
View the original p
Somewhere in the BeanInstantiator, there is the following code:
| // create the setter method name from the property name
| String setterMethodName =
"set"+propertyName.substring(0,1).toUpperCase()+propertyName.substring(1);
|
| // find the setter method
| M
because it is not a full task-node, and I want tasks to be ready as soon as the
process starts, with the same task-node functionalities. My facade already
implements it now ;-)
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3924281#3924281
Reply to the post :
is it possible to do an automatic signal out of the start-state.
I tried to use an event on process-start, but I got:
"can't continue execution on locked tokens. signalling the same token that
executes an action is not allowed"
View the original post :
http://www.jboss.com/index.html?module=bb&
Some quick investigation shows that the events are handled bottom-up.
Good :-).
I've did a little quick-and-dirty (no unit tests) job of:
adjusting the xsd, to accept the "bubble-up" (true|false|yes|no)
Added a property to event (guess which one)
Adjusted the hibernate mapping and added a column to
Is it possible for event handlers to stop bubbling up?
e.g. I have a task-create event handler on process level, and a specific
task-create event handler on some task in the process. Is it possible to
specify that the event processing should stop there?
Of course, if events are processed top-down
Testclass: testTransientVariablesOnly will fail if you do the transient bit
inside the conditional. Will also prove that you have to do the reverse bit as
well.
| /*
| * JBoss, Home of Professional Open Source
| * Copyright 2005, JBoss Inc., and individual contributors as indicated
|
Without testing it:
Does this has to be inside the conditional? This means that you have to have at
least one read/read-write mapping.
If the contract would state that all current transient variables are
transferred, this could be sufficient (I don't think you would have
read/read-write issues w
Jira entry created on approval of one of the gurus :-):
http://jira.jboss.com/jira/browse/JBPM-539
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3921449#3921449
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3921449
-
nudging to know if it will be done in 3.1 or 3.2.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3921448#3921448
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3921448
---
Is it possible to have events indicating that a suspend, cancel, resume has
occurred?
The request is quick, the implementation will be rather elaborate, I guess
xsd will change, etc. etc...
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3921427#3921427
Look at the resource patterns (on www.workflowpatterns.com .
anonymous wrote : Initially a work item comes into existence in the created
state. This indicates that
| the preconditions required for its enablement have been satisfied and it is
capable of
| being executed. At this point however,
damn, I meant http://www.workflowpatterns.com
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3921389#3921389
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3921389
---
This is still implemented in 3.1b2. Should I report this as a bug?
James
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3921189#3921189
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3921189
--
Seems to me a recursive thingy. Best to use a unit test to discover this.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3921127#3921127
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3921127
---
Of course. There was some glitches in the configuration for the session
factory. It must depend on jbpmConfiguration, so:
| - depends-on="jbpmConfiguration" for the sessionFactory (believe me, it is
necessary :-))
| - The addition of my favorite
directories:-)/hibernate.queries.hbm.xml (the
I guess there must be already some Spring integration initiatives. As we are
heavily into Spring, we tried an integration as well. We have been using jBpm
3.1b2, because that looked like the easiest jBpm framework to "Springify" (the
author has made it clear that the whole intention of 3.1 is mo
Hello
where can I find the feature list of the current version of the graphical
designer? all the node types are in there now, but for some, functionality
seems rather limited:
- superstates: this is a node, but does not "surround" its substates
- process state: how to specify the subprocess
Al
Seems like a variant of this question:
http://www.jboss.com/index.html?module=bb&op=viewtopic&t=68905
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3916011#3916011
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3916011
Maybe the guard is on status of the taskinstance itself. It could be possible
that taskinstances may not end unless a special guard condition is fullfilled.
There is a little statemachine behind taskinstance itself, and maybe you want
to put guards on various transitions these states as well (e.g
I was following the other thread concerning transitions, guards and the like
(http://www.jboss.com/index.html?module=bb&op=viewtopic&t=73286)
Of course, you could also use guards on transitions, filter the ones that are
available, and choose the first one (hopefully, in this case, there will be
"some sort of" decision handler.
I mean that, when you do a signal without a transition name, it will do a
getDefaultLeavingTransition on the node (check the sourcecode for this). It
would be nice if this function also calls a handler, similar to a
DecisionHandler, which returns the name of the
I once raised the idea that every nodetype should be "conditional" instead of
having a special node for that purpose. If you implement a conditionhandler,
this could be used when signal() is called (as this tries to get a default
transition, which is now always the first transition).
See http:
3.2?
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3910201#3910201
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3910201
---
This SF.net email is sponsored by: Splun
45 matches
Mail list logo
