4:13 a.m.
I have been using the open source Drools 5.1 where I have been inserting
facts in the RHS of a rule and that has been causing subsequent rules to
fire just fine.
However, I am working at a customer that requires my to use a supported
product, so I have moved my rules over to JBoss Rules 5.0.2 and I find
that this does not work any more.
I have a declarative fact type as follows
declare QuestionValidation
question: String
state: String
mandatory: Boolean
type: String
end
I have an "init" rule with high salience that inserts a number of
QuestionValidations into the session.
In order to test what is going on, I created a rule:-
rule 'Test'
dialect 'mvel'
when
$qv: QuestionValidation()
then
System.out.println( "QuestionValidation found " + $qv.getQuestion()
+ "/" + $qv.getState() );
end
This rule never fires if I insert a QuestionValidation in the init rule!
I can see the facts being inserted in the log as follows
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
thread-pool-1; w: 7;|
OBJECT ASSERTED value:QuestionValidation( mandatory=true, state=Closed,
question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
But I don't see any associated "ACTIVATION CREATED" for this object.
However, if I insert a QuestionValidation fact created in Java by using
kbase.getFactType("SWWB2BOutgoing", "QuestionValidation") then the
Test
rule gets fired just fine.
Is there something I am missing? Do I need to to turn on rule chaining
or something like that? (Surely chaining is on by default).
Chris
5:50 a.m.
We aren't seeing the code where the insertions happen, but I suppose that's
identical to
repeats of
QuestionValidation qc = new QuestionValidation();
insert( qv );
I'm sure you've also ascertained that the inserted value:QuestionValidation
is the very same type (including package) as the one used in rule "Test".
I would try and use a Java equivalent for the declared class
QuestionValidation,
just to see whether this is at the root of the problem.
-W
On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk> wrote:
I have been using the open source Drools 5.1 where I have been
inserting
facts in the RHS of a rule and that has been causing subsequent rules to
fire just fine.
However, I am working at a customer that requires my to use a supported
product, so I have moved my rules over to JBoss Rules 5.0.2 and I find
that this does not work any more.
I have a declarative fact type as follows
declare QuestionValidation
question: String
state: String
mandatory: Boolean
type: String
end
I have an "init" rule with high salience that inserts a number of
QuestionValidations into the session.
In order to test what is going on, I created a rule:-
rule 'Test'
dialect 'mvel'
when
$qv: QuestionValidation()
then
System.out.println( "QuestionValidation found " + $qv.getQuestion()
+ "/" + $qv.getState() );
end
This rule never fires if I insert a QuestionValidation in the init rule!
I can see the facts being inserted in the log as follows
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
thread-pool-1; w: 7;|
OBJECT ASSERTED value:QuestionValidation( mandatory=true, state=Closed,
question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
But I don't see any associated "ACTIVATION CREATED" for this object.
However, if I insert a QuestionValidation fact created in Java by using
kbase.getFactType("SWWB2BOutgoing", "QuestionValidation") then the
Test
rule gets fired just fine.
Is there something I am missing? Do I need to to turn on rule chaining
or something like that? (Surely chaining is on by default).
Chris
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
6:05 a.m.
A closer look:
On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk> wrote:
I can see the facts being inserted in the log as follows
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
thread-pool-1; w: 7;|
OBJECT ASSERTED value:QuestionValidation( mandatory=true, state=Closed,
question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
Shouldn't the package name appear in the log as well?
However, if I insert a QuestionValidation fact created in Java by using
kbase.getFactType("SWWB2BOutgoing", "QuestionValidation") then the
Test
rule gets fired just fine.
That is, apparently "SWWB2BOutgoing".
-W
9:14 a.m.
Yes the package is "SWWB2BOutgoing" but it looks as though the package
name simple doesn't get reported in the log file.
I have been debugging through this thing and where I've got to so far is
that I am getting two (and maybe more) entries for the class
"SWWB2BOutgoing.QuestionValidation" in the
ObjectTypeConfigurationRepository.
So my current question is "How is this possible?" Is it maybe a
classloader issue?
Chris
On 01/12/2010 12:05, Wolfgang Laun wrote:
A closer look:
On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk
<mailto:chris@selwyn-family.me.uk>> wrote:
I can see the facts being inserted in the log as follows
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
thread-pool-1; w: 7;|
OBJECT ASSERTED value:QuestionValidation( mandatory=true,
state=Closed,
question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
Shouldn't the package name appear in the log as well?
However, if I insert a QuestionValidation fact created in Java by
using
kbase.getFactType("SWWB2BOutgoing", "QuestionValidation") then
the
Test
rule gets fired just fine.
That is, apparently "SWWB2BOutgoing".
-W
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
No virus found in this message.
Checked by AVG - www.avg.com <http://www.avg.com>
Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
3:29 p.m.
New subject: Rules not firing when facts asserted in a rule RHS in 5.0 - diagnosed with workaround (but not fixed)
Having spent all day on this one, I have finally got to the bottom of it.
The problem is reproducible if I have /both/ a declarative fact model
/and/ a function in my rules.
It doesn't matter whether the function is called or not... It's presence
is enough to cause the problem.
Any rule that is dependent on the existence of an instance of the
declarative fact model where the instance is created in the RHS of
another rule will not fire.
I suspect that this is a bug that has been fixed in 5.1 because I have
been using this technique in Drools 5.1... It's only since downgrading
to the (supported) JBoss Rules 5.0 that I have experienced the problem.
Does anyone know if there is a fix in the 5.0 branch?
If I create a "real" Java class and upload it in a jar file to the
Guvnor and use that in place of the declarative model then the problem
does not happen.
This is a workaround in the short term but I would much prefer not to
have to do this.
Chris Selwyn
On 01/12/2010 15:14, Chris Selwyn wrote:
Yes the package is "SWWB2BOutgoing" but it looks as though
the package
name simple doesn't get reported in the log file.
I have been debugging through this thing and where I've got to so far
is that I am getting two (and maybe more) entries for the class
"SWWB2BOutgoing.QuestionValidation" in the
ObjectTypeConfigurationRepository.
So my current question is "How is this possible?" Is it maybe a
classloader issue?
Chris
On 01/12/2010 12:05, Wolfgang Laun wrote:
> A closer look:
>
> On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk
> <mailto:chris@selwyn-family.me.uk>> wrote:
>
>
>
> I can see the facts being inserted in the log as follows
>
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
> thread-pool-1; w: 7;|
> OBJECT ASSERTED value:QuestionValidation( mandatory=true,
> state=Closed,
> question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
>
>
> Shouldn't the package name appear in the log as well?
>
>
> However, if I insert a QuestionValidation fact created in Java by
> using
> kbase.getFactType("SWWB2BOutgoing", "QuestionValidation")
then
> the Test
> rule gets fired just fine.
>
>
> That is, apparently "SWWB2BOutgoing".
>
> -W
>
>
> _______________________________________________
> rules-users mailing list
> rules-users(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
>
> No virus found in this message.
> Checked by AVG - www.avg.com <http://www.avg.com>
> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
No virus found in this message.
Checked by AVG - www.avg.com <http://www.avg.com>
Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
3:47 p.m.
New subject: Rules not firing when facts asserted in a rule RHS in 5.0 - diagnosed with workaround (but not fixed)
Hi Chris, could you attach the repository export xml from your repo that
I can use to reproduce the issue with? BRMS 5.1 (supported bits) is
right around the corner and I would be able to tell you if the issue
still persists with the currently tested bits for 5.1.
Thanks.
Tihomir
On 12/1/10 4:29 PM, Chris Selwyn wrote:
Having spent all day on this one, I have finally got to the bottom of
it.
The problem is reproducible if I have /both/ a declarative fact model
/and/ a function in my rules.
It doesn't matter whether the function is called or not... It's
presence is enough to cause the problem.
Any rule that is dependent on the existence of an instance of the
declarative fact model where the instance is created in the RHS of
another rule will not fire.
I suspect that this is a bug that has been fixed in 5.1 because I have
been using this technique in Drools 5.1... It's only since downgrading
to the (supported) JBoss Rules 5.0 that I have experienced the problem.
Does anyone know if there is a fix in the 5.0 branch?
If I create a "real" Java class and upload it in a jar file to the
Guvnor and use that in place of the declarative model then the problem
does not happen.
This is a workaround in the short term but I would much prefer not to
have to do this.
Chris Selwyn
On 01/12/2010 15:14, Chris Selwyn wrote:
> Yes the package is "SWWB2BOutgoing" but it looks as though the
> package name simple doesn't get reported in the log file.
>
> I have been debugging through this thing and where I've got to so far
> is that I am getting two (and maybe more) entries for the class
> "SWWB2BOutgoing.QuestionValidation" in the
> ObjectTypeConfigurationRepository.
> So my current question is "How is this possible?" Is it maybe a
> classloader issue?
>
> Chris
>
>
> On 01/12/2010 12:05, Wolfgang Laun wrote:
>> A closer look:
>>
>> On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk
>> <mailto:chris@selwyn-family.me.uk>> wrote:
>>
>>
>>
>> I can see the facts being inserted in the log as follows
>>
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
>> thread-pool-1; w: 7;|
>> OBJECT ASSERTED value:QuestionValidation( mandatory=true,
>> state=Closed,
>> question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
>>
>>
>> Shouldn't the package name appear in the log as well?
>>
>>
>> However, if I insert a QuestionValidation fact created in Java
>> by using
>> kbase.getFactType("SWWB2BOutgoing", "QuestionValidation")
then
>> the Test
>> rule gets fired just fine.
>>
>>
>> That is, apparently "SWWB2BOutgoing".
>>
>> -W
>>
>>
>> _______________________________________________
>> rules-users mailing list
>> rules-users(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/rules-users
>>
>>
>> No virus found in this message.
>> Checked by AVG - www.avg.com <http://www.avg.com>
>> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>>
>
> _______________________________________________
> rules-users mailing list
> rules-users(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
>
> No virus found in this message.
> Checked by AVG - www.avg.com <http://www.avg.com>
> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
4:09 p.m.
New subject: Rules not firing when facts asserted in a rule RHS in 5.0 - diagnosed with workaround (but not fixed)
Having just looked in the Drools JIRA repository, I pretty sure that it
is the same issue as was reported in JBRULES-2140
<https://jira.jboss.org/browse/JBRULES-2140>;.
That entry has a repro.
I am interested as to what you mean by "right around the corner"?
Here is a repro anyway...
<QUOTE>
package rulestest
global java.util.logging.Logger logger
declare QuestionValidation
question: String
mandatory: Boolean
state: String
type: String
end
function void testfunc() {
System.out.println("testfunc called");
}
rule 'init'
dialect "mvel"
when
eval(true)
then
QuestionValidation qv = new QuestionValidation();
qv.setState("ss");
qv.setQuestion("qq");
insert(qv);
logger.info("Hello world");
end
rule 'test'
dialect "mvel"
when
$qv: QuestionValidation()
then
logger.info("Found a qv " + $qv);
end
</QUOTE>
If the testfunc is present then the "test" rule never fires. If it is
not present then the "test" rule fires just fine.
Chris
On 01/12/2010 21:47, Tihomir Surdilovic wrote:
Hi Chris, could you attach the repository export xml from your repo
that I can use to reproduce the issue with? BRMS 5.1 (supported bits)
is right around the corner and I would be able to tell you if the
issue still persists with the currently tested bits for 5.1.
Thanks.
Tihomir
On 12/1/10 4:29 PM, Chris Selwyn wrote:
> Having spent all day on this one, I have finally got to the bottom of it.
>
> The problem is reproducible if I have /both/ a declarative fact model
> /and/ a function in my rules.
> It doesn't matter whether the function is called or not... It's
> presence is enough to cause the problem.
>
> Any rule that is dependent on the existence of an instance of the
> declarative fact model where the instance is created in the RHS of
> another rule will not fire.
>
> I suspect that this is a bug that has been fixed in 5.1 because I
> have been using this technique in Drools 5.1... It's only since
> downgrading to the (supported) JBoss Rules 5.0 that I have
> experienced the problem.
>
> Does anyone know if there is a fix in the 5.0 branch?
>
> If I create a "real" Java class and upload it in a jar file to the
> Guvnor and use that in place of the declarative model then the
> problem does not happen.
> This is a workaround in the short term but I would much prefer not to
> have to do this.
>
> Chris Selwyn
>
>
> On 01/12/2010 15:14, Chris Selwyn wrote:
>> Yes the package is "SWWB2BOutgoing" but it looks as though the
>> package name simple doesn't get reported in the log file.
>>
>> I have been debugging through this thing and where I've got to so
>> far is that I am getting two (and maybe more) entries for the class
>> "SWWB2BOutgoing.QuestionValidation" in the
>> ObjectTypeConfigurationRepository.
>> So my current question is "How is this possible?" Is it maybe a
>> classloader issue?
>>
>> Chris
>>
>>
>> On 01/12/2010 12:05, Wolfgang Laun wrote:
>>> A closer look:
>>>
>>> On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk
>>> <mailto:chris@selwyn-family.me.uk>> wrote:
>>>
>>>
>>>
>>> I can see the facts being inserted in the log as follows
>>>
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
>>> thread-pool-1; w: 7;|
>>> OBJECT ASSERTED value:QuestionValidation( mandatory=true,
>>> state=Closed,
>>> question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
>>>
>>>
>>> Shouldn't the package name appear in the log as well?
>>>
>>>
>>> However, if I insert a QuestionValidation fact created in Java
>>> by using
>>> kbase.getFactType("SWWB2BOutgoing",
"QuestionValidation") then
>>> the Test
>>> rule gets fired just fine.
>>>
>>>
>>> That is, apparently "SWWB2BOutgoing".
>>>
>>> -W
>>>
>>>
>>> _______________________________________________
>>> rules-users mailing list
>>> rules-users(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>
>>>
>>> No virus found in this message.
>>> Checked by AVG - www.avg.com <http://www.avg.com>
>>> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>>>
>>
>> _______________________________________________
>> rules-users mailing list
>> rules-users(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/rules-users
>>
>>
>> No virus found in this message.
>> Checked by AVG - www.avg.com <http://www.avg.com>
>> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>>
>
> _______________________________________________
> rules-users mailing list
> rules-users(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
No virus found in this message.
Checked by AVG - www.avg.com <http://www.avg.com>
Version: 10.0.1170 / Virus Database: 426/3291 - Release Date: 12/01/10
5:04 p.m.
New subject: Rules not firing when facts asserted in a rule RHS in 5.0 - diagnosed with workaround (but not fixed)
On 12/1/10 5:09 PM, Chris Selwyn wrote:
Having just looked in the Drools JIRA repository, I pretty sure that
it is the same issue as was reported in JBRULES-2140
<https://jira.jboss.org/browse/JBRULES-2140>;.
That entry has a repro.
thanks for finding
I am interested as to what you mean by "right around the corner"?
this month
Here is a repro anyway...
<QUOTE>
package rulestest
global java.util.logging.Logger logger
declare QuestionValidation
question: String
mandatory: Boolean
state: String
type: String
end
function void testfunc() {
System.out.println("testfunc called");
}
rule 'init'
dialect "mvel"
when
eval(true)
then
QuestionValidation qv = new QuestionValidation();
qv.setState("ss");
qv.setQuestion("qq");
insert(qv);
logger.info("Hello world");
end
rule 'test'
dialect "mvel"
when
$qv: QuestionValidation()
then
logger.info("Found a qv " + $qv);
end
</QUOTE>
If the testfunc is present then the "test" rule never fires. If it is
not present then the "test" rule fires just fine.
Chris
On 01/12/2010 21:47, Tihomir Surdilovic wrote:
> Hi Chris, could you attach the repository export xml from your repo
> that I can use to reproduce the issue with? BRMS 5.1 (supported bits)
> is right around the corner and I would be able to tell you if the
> issue still persists with the currently tested bits for 5.1.
>
> Thanks.
> Tihomir
>
> On 12/1/10 4:29 PM, Chris Selwyn wrote:
>> Having spent all day on this one, I have finally got to the bottom
>> of it.
>>
>> The problem is reproducible if I have /both/ a declarative fact
>> model /and/ a function in my rules.
>> It doesn't matter whether the function is called or not... It's
>> presence is enough to cause the problem.
>>
>> Any rule that is dependent on the existence of an instance of the
>> declarative fact model where the instance is created in the RHS of
>> another rule will not fire.
>>
>> I suspect that this is a bug that has been fixed in 5.1 because I
>> have been using this technique in Drools 5.1... It's only since
>> downgrading to the (supported) JBoss Rules 5.0 that I have
>> experienced the problem.
>>
>> Does anyone know if there is a fix in the 5.0 branch?
>>
>> If I create a "real" Java class and upload it in a jar file to the
>> Guvnor and use that in place of the declarative model then the
>> problem does not happen.
>> This is a workaround in the short term but I would much prefer not
>> to have to do this.
>>
>> Chris Selwyn
>>
>>
>> On 01/12/2010 15:14, Chris Selwyn wrote:
>>> Yes the package is "SWWB2BOutgoing" but it looks as though the
>>> package name simple doesn't get reported in the log file.
>>>
>>> I have been debugging through this thing and where I've got to so
>>> far is that I am getting two (and maybe more) entries for the class
>>> "SWWB2BOutgoing.QuestionValidation" in the
>>> ObjectTypeConfigurationRepository.
>>> So my current question is "How is this possible?" Is it maybe a
>>> classloader issue?
>>>
>>> Chris
>>>
>>>
>>> On 01/12/2010 12:05, Wolfgang Laun wrote:
>>>> A closer look:
>>>>
>>>> On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk
>>>> <mailto:chris@selwyn-family.me.uk>> wrote:
>>>>
>>>>
>>>>
>>>> I can see the facts being inserted in the log as follows
>>>>
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
>>>> thread-pool-1; w: 7;|
>>>> OBJECT ASSERTED value:QuestionValidation( mandatory=true,
>>>> state=Closed,
>>>> question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
>>>>
>>>>
>>>> Shouldn't the package name appear in the log as well?
>>>>
>>>>
>>>> However, if I insert a QuestionValidation fact created in Java
>>>> by using
>>>> kbase.getFactType("SWWB2BOutgoing",
"QuestionValidation") then
>>>> the Test
>>>> rule gets fired just fine.
>>>>
>>>>
>>>> That is, apparently "SWWB2BOutgoing".
>>>>
>>>> -W
>>>>
>>>>
>>>> _______________________________________________
>>>> rules-users mailing list
>>>> rules-users(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>>
>>>>
>>>> No virus found in this message.
>>>> Checked by AVG - www.avg.com <http://www.avg.com>
>>>> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>>>>
>>>
>>> _______________________________________________
>>> rules-users mailing list
>>> rules-users(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>
>>>
>>> No virus found in this message.
>>> Checked by AVG - www.avg.com <http://www.avg.com>
>>> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>>>
>>
>> _______________________________________________
>> rules-users mailing list
>> rules-users(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/rules-users
>
>
> _______________________________________________
> rules-users mailing list
> rules-users(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
>
> No virus found in this message.
> Checked by AVG - www.avg.com <http://www.avg.com>
> Version: 10.0.1170 / Virus Database: 426/3291 - Release Date: 12/01/10
>
5:25 p.m.
New subject: Rules not firing when facts asserted in a rule RHS in 5.0 - diagnosed with workaround (but not fixed)
On 12/1/10 6:04 PM, Tihomir Surdilovic wrote:
On 12/1/10 5:09 PM, Chris Selwyn wrote:
> Having just looked in the Drools JIRA repository, I pretty sure that
> it is the same issue as was reported in JBRULES-2140
> <https://jira.jboss.org/browse/JBRULES-2140>;.
Fix for JBRULES-2140
<https://jira.jboss.org/browse/JBRULES-2140> will
be included in the next release of the supported bits (5.1). Since you
already are using supported bits I assume you know where to check for
updates.
If you cannot upgrade, open a ticket with JBoss support and request a
backport fix.
> That entry has a repro.
thanks for finding
>
> I am interested as to what you mean by "right around the corner"?
>
this month
> Here is a repro anyway...
>
> <QUOTE>
> package rulestest
> global java.util.logging.Logger logger
>
> declare QuestionValidation
> question: String
> mandatory: Boolean
> state: String
> type: String
> end
>
> function void testfunc() {
> System.out.println("testfunc called");
>
> }
>
> rule 'init'
> dialect "mvel"
> when
> eval(true)
> then
> QuestionValidation qv = new QuestionValidation();
> qv.setState("ss");
> qv.setQuestion("qq");
> insert(qv);
> logger.info("Hello world");
>
> end
>
> rule 'test'
> dialect "mvel"
> when
> $qv: QuestionValidation()
> then
> logger.info("Found a qv " + $qv);
>
> end
> </QUOTE>
>
> If the testfunc is present then the "test" rule never fires. If it is
> not present then the "test" rule fires just fine.
>
> Chris
>
> On 01/12/2010 21:47, Tihomir Surdilovic wrote:
>> Hi Chris, could you attach the repository export xml from your repo
>> that I can use to reproduce the issue with? BRMS 5.1 (supported
>> bits) is right around the corner and I would be able to tell you if
>> the issue still persists with the currently tested bits for 5.1.
>>
>> Thanks.
>> Tihomir
>>
>> On 12/1/10 4:29 PM, Chris Selwyn wrote:
>>> Having spent all day on this one, I have finally got to the bottom
>>> of it.
>>>
>>> The problem is reproducible if I have /both/ a declarative fact
>>> model /and/ a function in my rules.
>>> It doesn't matter whether the function is called or not... It's
>>> presence is enough to cause the problem.
>>>
>>> Any rule that is dependent on the existence of an instance of the
>>> declarative fact model where the instance is created in the RHS of
>>> another rule will not fire.
>>>
>>> I suspect that this is a bug that has been fixed in 5.1 because I
>>> have been using this technique in Drools 5.1... It's only since
>>> downgrading to the (supported) JBoss Rules 5.0 that I have
>>> experienced the problem.
>>>
>>> Does anyone know if there is a fix in the 5.0 branch?
>>>
>>> If I create a "real" Java class and upload it in a jar file to the
>>> Guvnor and use that in place of the declarative model then the
>>> problem does not happen.
>>> This is a workaround in the short term but I would much prefer not
>>> to have to do this.
>>>
>>> Chris Selwyn
>>>
>>>
>>> On 01/12/2010 15:14, Chris Selwyn wrote:
>>>> Yes the package is "SWWB2BOutgoing" but it looks as though the
>>>> package name simple doesn't get reported in the log file.
>>>>
>>>> I have been debugging through this thing and where I've got to so
>>>> far is that I am getting two (and maybe more) entries for the
>>>> class "SWWB2BOutgoing.QuestionValidation" in the
>>>> ObjectTypeConfigurationRepository.
>>>> So my current question is "How is this possible?" Is it maybe a
>>>> classloader issue?
>>>>
>>>> Chris
>>>>
>>>>
>>>> On 01/12/2010 12:05, Wolfgang Laun wrote:
>>>>> A closer look:
>>>>>
>>>>> On 1 December 2010 11:13, Chris Selwyn <chris(a)selwyn-family.me.uk
>>>>> <mailto:chris@selwyn-family.me.uk>> wrote:
>>>>>
>>>>>
>>>>>
>>>>> I can see the facts being inserted in the log as follows
>>>>>
[#|2010-12-01T09:19:16.723+0000|INFO|sun-appserver2.1|javax.enterprise.system.stream.out|_ThreadID=70;_ThreadName=p:
>>>>> thread-pool-1; w: 7;|
>>>>> OBJECT ASSERTED value:QuestionValidation( mandatory=true,
>>>>> state=Closed,
>>>>> question=B2B_METER_SERIAL_NUMBER, type=string ) factId: 39|#]
>>>>>
>>>>>
>>>>> Shouldn't the package name appear in the log as well?
>>>>>
>>>>>
>>>>> However, if I insert a QuestionValidation fact created in
>>>>> Java by using
>>>>> kbase.getFactType("SWWB2BOutgoing",
"QuestionValidation")
>>>>> then the Test
>>>>> rule gets fired just fine.
>>>>>
>>>>>
>>>>> That is, apparently "SWWB2BOutgoing".
>>>>>
>>>>> -W
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> rules-users mailing list
>>>>> rules-users(a)lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>>>
>>>>>
>>>>> No virus found in this message.
>>>>> Checked by AVG - www.avg.com <http://www.avg.com>
>>>>> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date:
>>>>> 11/30/10
>>>>>
>>>>
>>>> _______________________________________________
>>>> rules-users mailing list
>>>> rules-users(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>>
>>>>
>>>> No virus found in this message.
>>>> Checked by AVG - www.avg.com <http://www.avg.com>
>>>> Version: 10.0.1170 / Virus Database: 426/3290 - Release Date: 11/30/10
>>>>
>>>
>>> _______________________________________________
>>> rules-users mailing list
>>> rules-users(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>
>>
>> _______________________________________________
>> rules-users mailing list
>> rules-users(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/rules-users
>>
>>
>> No virus found in this message.
>> Checked by AVG - www.avg.com <http://www.avg.com>
>> Version: 10.0.1170 / Virus Database: 426/3291 - Release Date: 12/01/10
>>
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
5109
days inactive
5109
days old
8 comments
3 participants
participants (3)
-
Chris Selwyn -
Tihomir Surdilovic -
Wolfgang Laun