9:16 p.m.
It's great when people attach unit tests to jiras, it would be ever
better if they are just cut nad paste single files, rather than entire
projects, which take more time for us to fiddle with. Also please do try
and pair your actual program "main" examples down to minimal unit tests
of the actual problem. If you suspect the problem is MVEL, then please
do a minimal MVEL test, without Drools.
Put static classes in the same file, put rules in a concatenated string.
This means testing is as easy as cut and paste for us, and it can be
easily added to existing test classes. For example, from MiscTest.
public static class A {
private String field1;
private String field2;
public A(String field1,
String field2) {
this.field1 = field1;
this.field2 = field2;
}
public String getField1() {
return field1;
}
public void setField1( String field1 ) {
this.field1 = field1;
}
public String getField2() {
return field2;
}
public void setField2( String field2 ) {
this.field2 = field2;
}
public String toString() {
return "A) " + field1 + ":" + field2;
}
}
@Test
public void testExistsIterativeModifyBug() {
// JBRULES-2809
// This bug occurs when a tuple is modified, the remove/add
puts it onto the memory end
// However before this was done it would attempt to find the
next tuple, starting from itself
// This meant it would just re-add itself as the blocker, but
then be moved to end of the memory
// If this tuple was then removed or changed, the blocked was
unable to check previous tuples.
String str = "";
str += "package org.simple \n";
str += "import " + A.class.getCanonicalName() + "\n";
str += "global java.util.List list \n";
str += "rule xxx \n";
str += "when \n";
str += " $f1 : A() \n";
str += " exists A(this != $f1, eval(field2 ==
$f1.getField2())) \n";
str += " eval( !$f1.getField1().equals(\"1\") ) \n";
str += "then \n";
str += " list.add($f1); \n";
str += "end \n";
KnowledgeBase kbase = loadKnowledgeBaseFromString( str );
StatefulKnowledgeSession ksession =
kbase.newStatefulKnowledgeSession();
List list = new ArrayList();
ksession.setGlobal( "list",
list );
A a1 = new A( "2",
"2" );
A a2 = new A( "1",
"2" );
A a3 = new A( "1",
"2" );
FactHandle fa1 = (FactHandle) ksession.insert( a1 );
FactHandle fa2 = (FactHandle) ksession.insert( a2 );
FactHandle fa3 = (FactHandle) ksession.insert( a3 );
// a2, a3 are blocked by a1
// modify a1, so that a1,a3 are now blocked by a2
a1.setField2( "1" ); // Do
ksession.update( fa1,
a1 );
a1.setField2( "2" ); // Undo
ksession.update( fa1,
a1 );
// modify a2, so that a1,a2 are now blocked by a3
a2.setField2( "1" ); // Do
ksession.update( fa2,
a2 );
a2.setField2( "2" ); // Undo
ksession.update( fa2,
a2 );
// modify a3 to cycle, so that it goes on the memory end, but
in a previous bug still blocked a1
ksession.update( fa3,
a3 );
a3.setField2( "1" ); // Do
ksession.update( fa3,
a3 );
ksession.fireAllRules();
assertEquals( 1,
list.size() ); // a2 should still be blocked by
a1, but bug from previous update hanging onto blocked
ksession.dispose();
}
2:44 a.m.
I totally agree, except for the inline DRL :)
If it's a big enough DRL (let's say +3 rules), isn't it easier to have
something on the lines of the existing integration tests inside
drools-compiler? With the drl in the resources folder.
I know that people say "a test is not a unit test if it touches the
filesystem", but as long as Java doesn't support a proper way of inlining
XML </> (or DRL in this case, or different text formats in the general
case), like Scala does, I personally think it's easier and more readable to
have a separate file.
My 5 cents.
Leo.
On Tue, May 17, 2011 at 4:16 AM, Mark Proctor <mproctor(a)codehaus.org> wrote:
It's great when people attach unit tests to jiras, it would be
ever
better if they are just cut nad paste single files, rather than entire
projects, which take more time for us to fiddle with. Also please do try
and pair your actual program "main" examples down to minimal unit tests
of the actual problem. If you suspect the problem is MVEL, then please
do a minimal MVEL test, without Drools.
Put static classes in the same file, put rules in a concatenated string.
This means testing is as easy as cut and paste for us, and it can be
easily added to existing test classes. For example, from MiscTest.
public static class A {
private String field1;
private String field2;
public A(String field1,
String field2) {
this.field1 = field1;
this.field2 = field2;
}
public String getField1() {
return field1;
}
public void setField1( String field1 ) {
this.field1 = field1;
}
public String getField2() {
return field2;
}
public void setField2( String field2 ) {
this.field2 = field2;
}
public String toString() {
return "A) " + field1 + ":" + field2;
}
}
@Test
public void testExistsIterativeModifyBug() {
// JBRULES-2809
// This bug occurs when a tuple is modified, the remove/add
puts it onto the memory end
// However before this was done it would attempt to find the
next tuple, starting from itself
// This meant it would just re-add itself as the blocker, but
then be moved to end of the memory
// If this tuple was then removed or changed, the blocked was
unable to check previous tuples.
String str = "";
str += "package org.simple \n";
str += "import " + A.class.getCanonicalName() + "\n";
str += "global java.util.List list \n";
str += "rule xxx \n";
str += "when \n";
str += " $f1 : A() \n";
str += " exists A(this != $f1, eval(field2 ==
$f1.getField2())) \n";
str += " eval( !$f1.getField1().equals(\"1\") ) \n";
str += "then \n";
str += " list.add($f1); \n";
str += "end \n";
KnowledgeBase kbase = loadKnowledgeBaseFromString( str );
StatefulKnowledgeSession ksession =
kbase.newStatefulKnowledgeSession();
List list = new ArrayList();
ksession.setGlobal( "list",
list );
A a1 = new A( "2",
"2" );
A a2 = new A( "1",
"2" );
A a3 = new A( "1",
"2" );
FactHandle fa1 = (FactHandle) ksession.insert( a1 );
FactHandle fa2 = (FactHandle) ksession.insert( a2 );
FactHandle fa3 = (FactHandle) ksession.insert( a3 );
// a2, a3 are blocked by a1
// modify a1, so that a1,a3 are now blocked by a2
a1.setField2( "1" ); // Do
ksession.update( fa1,
a1 );
a1.setField2( "2" ); // Undo
ksession.update( fa1,
a1 );
// modify a2, so that a1,a2 are now blocked by a3
a2.setField2( "1" ); // Do
ksession.update( fa2,
a2 );
a2.setField2( "2" ); // Undo
ksession.update( fa2,
a2 );
// modify a3 to cycle, so that it goes on the memory end, but
in a previous bug still blocked a1
ksession.update( fa3,
a3 );
a3.setField2( "1" ); // Do
ksession.update( fa3,
a3 );
ksession.fireAllRules();
assertEquals( 1,
list.size() ); // a2 should still be blocked by
a1, but bug from previous update hanging onto blocked
ksession.dispose();
}
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
10 a.m.
On 17/05/2011 08:44, Leonardo Gomes wrote:
I totally agree, except for the inline DRL :)
If it's a big enough DRL (let's say +3 rules), isn't it easier to have
something on the lines of the existing integration tests inside
drools-compiler? With the drl in the resources folder.
Once the rule(s) gets larger
yes. The problem we have now though is look
at the integration tests resource folder, it gets big pretty quick. And
ofcourse there is no automatic button to go from the java test to the
drl test file, so it's extra key presses. When you are doing a lot of
reviewing, or checking old regressions, it can can be nice to have
everything in a single place. Personally I just wish Java had multi-line
strings and then this wouldn't be an issue :(
I know that people say "a test is not a unit test if it touches the
filesystem", but as long as Java doesn't support a proper way of
inlining XML</> (or DRL in this case, or different text formats in the
general case), like Scala does, I personally think it's easier and
more readable to have a separate file.
It's nothign to do with being a proper
unit test, and more just making
edson and my life easier.
Mark
My 5 cents.
Leo.
On Tue, May 17, 2011 at 4:16 AM, Mark Proctor <mproctor(a)codehaus.org
<mailto:mproctor@codehaus.org>> wrote:
It's great when people attach unit tests to jiras, it would be ever
better if they are just cut nad paste single files, rather than entire
projects, which take more time for us to fiddle with. Also please
do try
and pair your actual program "main" examples down to minimal unit
tests
of the actual problem. If you suspect the problem is MVEL, then please
do a minimal MVEL test, without Drools.
Put static classes in the same file, put rules in a concatenated
string.
This means testing is as easy as cut and paste for us, and it can be
easily added to existing test classes. For example, from MiscTest.
public static class A {
private String field1;
private String field2;
public A(String field1,
String field2) {
this.field1 = field1;
this.field2 = field2;
}
public String getField1() {
return field1;
}
public void setField1( String field1 ) {
this.field1 = field1;
}
public String getField2() {
return field2;
}
public void setField2( String field2 ) {
this.field2 = field2;
}
public String toString() {
return "A) " + field1 + ":" + field2;
}
}
@Test
public void testExistsIterativeModifyBug() {
// JBRULES-2809
// This bug occurs when a tuple is modified, the remove/add
puts it onto the memory end
// However before this was done it would attempt to find the
next tuple, starting from itself
// This meant it would just re-add itself as the blocker, but
then be moved to end of the memory
// If this tuple was then removed or changed, the blocked was
unable to check previous tuples.
String str = "";
str += "package org.simple \n";
str += "import " + A.class.getCanonicalName() + "\n";
str += "global java.util.List list \n";
str += "rule xxx \n";
str += "when \n";
str += " $f1 : A() \n";
str += " exists A(this != $f1, eval(field2 ==
$f1.getField2())) \n";
str += " eval( !$f1.getField1().equals(\"1\") ) \n";
str += "then \n";
str += " list.add($f1); \n";
str += "end \n";
KnowledgeBase kbase = loadKnowledgeBaseFromString( str );
StatefulKnowledgeSession ksession =
kbase.newStatefulKnowledgeSession();
List list = new ArrayList();
ksession.setGlobal( "list",
list );
A a1 = new A( "2",
"2" );
A a2 = new A( "1",
"2" );
A a3 = new A( "1",
"2" );
FactHandle fa1 = (FactHandle) ksession.insert( a1 );
FactHandle fa2 = (FactHandle) ksession.insert( a2 );
FactHandle fa3 = (FactHandle) ksession.insert( a3 );
// a2, a3 are blocked by a1
// modify a1, so that a1,a3 are now blocked by a2
a1.setField2( "1" ); // Do
ksession.update( fa1,
a1 );
a1.setField2( "2" ); // Undo
ksession.update( fa1,
a1 );
// modify a2, so that a1,a2 are now blocked by a3
a2.setField2( "1" ); // Do
ksession.update( fa2,
a2 );
a2.setField2( "2" ); // Undo
ksession.update( fa2,
a2 );
// modify a3 to cycle, so that it goes on the memory end, but
in a previous bug still blocked a1
ksession.update( fa3,
a3 );
a3.setField2( "1" ); // Do
ksession.update( fa3,
a3 );
ksession.fireAllRules();
assertEquals( 1,
list.size() ); // a2 should still be blocked by
a1, but bug from previous update hanging onto blocked
ksession.dispose();
}
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org <mailto:rules-dev@lists.jboss.org>
https://lists.jboss.org/mailman/listinfo/rules-dev
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
11:40 a.m.
Would be good to have that on a wiki page which is linked to when someone
goes to create a new bug jira
On Mon, May 16, 2011 at 7:16 PM, Mark Proctor <mproctor(a)codehaus.org> wrote:
It's great when people attach unit tests to jiras, it would be
ever
better if they are just cut nad paste single files, rather than entire
projects, which take more time for us to fiddle with. Also please do try
and pair your actual program "main" examples down to minimal unit tests
of the actual problem. If you suspect the problem is MVEL, then please
do a minimal MVEL test, without Drools.
Put static classes in the same file, put rules in a concatenated string.
This means testing is as easy as cut and paste for us, and it can be
easily added to existing test classes. For example, from MiscTest.
public static class A {
private String field1;
private String field2;
public A(String field1,
String field2) {
this.field1 = field1;
this.field2 = field2;
}
public String getField1() {
return field1;
}
public void setField1( String field1 ) {
this.field1 = field1;
}
public String getField2() {
return field2;
}
public void setField2( String field2 ) {
this.field2 = field2;
}
public String toString() {
return "A) " + field1 + ":" + field2;
}
}
@Test
public void testExistsIterativeModifyBug() {
// JBRULES-2809
// This bug occurs when a tuple is modified, the remove/add
puts it onto the memory end
// However before this was done it would attempt to find the
next tuple, starting from itself
// This meant it would just re-add itself as the blocker, but
then be moved to end of the memory
// If this tuple was then removed or changed, the blocked was
unable to check previous tuples.
String str = "";
str += "package org.simple \n";
str += "import " + A.class.getCanonicalName() + "\n";
str += "global java.util.List list \n";
str += "rule xxx \n";
str += "when \n";
str += " $f1 : A() \n";
str += " exists A(this != $f1, eval(field2 ==
$f1.getField2())) \n";
str += " eval( !$f1.getField1().equals(\"1\") ) \n";
str += "then \n";
str += " list.add($f1); \n";
str += "end \n";
KnowledgeBase kbase = loadKnowledgeBaseFromString( str );
StatefulKnowledgeSession ksession =
kbase.newStatefulKnowledgeSession();
List list = new ArrayList();
ksession.setGlobal( "list",
list );
A a1 = new A( "2",
"2" );
A a2 = new A( "1",
"2" );
A a3 = new A( "1",
"2" );
FactHandle fa1 = (FactHandle) ksession.insert( a1 );
FactHandle fa2 = (FactHandle) ksession.insert( a2 );
FactHandle fa3 = (FactHandle) ksession.insert( a3 );
// a2, a3 are blocked by a1
// modify a1, so that a1,a3 are now blocked by a2
a1.setField2( "1" ); // Do
ksession.update( fa1,
a1 );
a1.setField2( "2" ); // Undo
ksession.update( fa1,
a1 );
// modify a2, so that a1,a2 are now blocked by a3
a2.setField2( "1" ); // Do
ksession.update( fa2,
a2 );
a2.setField2( "2" ); // Undo
ksession.update( fa2,
a2 );
// modify a3 to cycle, so that it goes on the memory end, but
in a previous bug still blocked a1
ksession.update( fa3,
a3 );
a3.setField2( "1" ); // Do
ksession.update( fa3,
a3 );
ksession.fireAllRules();
assertEquals( 1,
list.size() ); // a2 should still be blocked by
a1, but bug from previous update hanging onto blocked
ksession.dispose();
}
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
--
Michael D Neale
home: www.michaelneale.net
blog: michaelneale.blogspot.com
4972
days inactive
4972
days old
3 comments
3 participants
participants (3)
-
Leonardo Gomes -
Mark Proctor -
Michael Neale