11.3 --> 11.7
After converting to 11.7 the following statement seems to give an error:
User-defined function 'getMsgQueueProperty' invoked dynamically but could not be found. (5639) Procedure:"ProcessMessageReadyIn oomsg.msgQueue"
the statement is:
dtzDocumentDate = datetime-tz(dynamic-function("getMsgQueueProperty" in source-procedure,msgQueue.msgQueueGID,'DocumentDate'))
-------------
This is how it is called:
* msgQueue.cls has the Method getMsgQueueProperty.
in msgQueue there is another method that calls:
oDocument = NEW msgDocument(oMeasuringTicket,msgQueue.msgQueueGID).
in msgDocument I need to Call the method getMsgQueueProperty and untill now it has worked using dynamic-function(....
How would I do this now in 11.7 ? I want to Call a method in the Source-procedure (witch is a class).
|
I have looked at it, but how I am not sure how to Call the Source-procedure ... How do I refere to the Source-procedure's class? Is Source-procedure the handle to the class?
//Geir Otto
Source-procedure is a procedural construct and should not be used for classes ... when calling into a class method, you need to use a reference variable.
So If I have two classes:
a.cls
b.cls
and in a.cls I call b.cls, how can I refere to class a? Too much vacation here....
Hmm, this seems to imply that source-procedure albeit a procedure handle did pointed to an object instance (class) before… probably not supported, still it does seems to confirm my thought of initial OO implementation being just an ‘abstraction’ over existent procedural constructs, namely functions. This not working anymore might mean something changed there so an object instance it’s just not a persistent procedure with a makeup :)
Thanks Mike :-) I allways forget this statement...
If I add THIS-OBJECT as a parameter, it gives me an error : Invalid handle. Not initialized or Points to a deleted Object.... How can that be?
ooo, by the way Mike, I want to Call a Method in A.cls FROM B.cls.....
I tried in A.cls
oB = New b (this-object).
and in b.cls:
constructor ....(oA as class a):
test = oA:theMethod('something,222).
I get the error saying that oA is invalid....
test = oA:theMethod('something,222).
yes the test = oA:method(....) is done within the constructor of B. So, I have to define a variable x as class A ?.
something like this?
constructor of b (oA as class A):
x = oA.
x:method(....). // seems a bit strange to do it like this....
end constructor.
I have no idea what you are doing wrong, but this works in 11.3 both from a constructor and a method in a I can instantiate b and run a method in a from it's constructor. I tested it because we had this-object bugs in the past (but that was in 10.2B):
class a:
constructor a ():
new b(this-object, "from constructor").
end constructor.
method public void test(input icMessage as character ):
message icMessage
view-as alert-box info buttons ok.
end method.
method public void newbFromMethod():
new b(this-object, "from method").
end method.
end class.
class b:
constructor b(input oa as a, input icMessage as character):
oa:test(icMessage).
end constructor.
end class.
define variable oa as a no-undo.
oa = new a().
oa:newbFromMethod().
It is a bit strange for me as well. It is a bit tricky to test it since it is in a live Production system running on Linux x64 bit OS, so I can't test it for the moment. I have fixed the problem by doing a workaround....
The only diff from what you sent is that New b is not running from the constructor of class A, but from a Method in class A....
> The only diff from what you sent is that New b is not running from the constructor of class A, but from a Method in class A....
That's what oa:newbFromMethod() does.
I also tried making newbFromMethod static since then I would expect THIS-OBJECT to be valid, but then it doesn't even compile so that can't be what you're doing (or something changed in 11.7.)
You need to post some code. I no longer know what you're doing.
Here it is:
Two classes, A and B.
Class A:
def property AGuid as char ....
constructor ()
Method Public logical doStuff ():
:
: some stuff that sets the AGUID property....
:
oB = New B (THIS-OBJECT).
Return somestuff.
END.
Method Public character getStuff (input icGID as char, input icName as char):
:
Return theReturn.
END.
Class B:
constructur (input oA as class A):
find somestuff where iID = oA:AGuid...... no-lock ....
aVariable = oA:getStuff('Something',aintegerfromthefind).
end.
I am not sure if this made it more informativ....:-) but I need to Reference the class A from within class B.
Obviously, there is SomeStuff missing, but one of the questions was the scope of oB, which I don't see defined anywhere.
Personally, I avoid doing things in constructors that could possibly fail, e.g., newing another class. If the constructor fails, you are left with an undefined reference. You might try moving such things to an Initialize method, both for this reason and because it might making it more clear.
Is somestuff a temp-table or DB table or what? What is it's scope? (I encapsulate TTs within a class).
In fact, I meet the same issues here, dynamic-function not work for the source-procedure of a class,
I meet the same issue here. We use the source-procedure handle for a class, Dynamic-function(method name, sour-procedure for a class). This works before 11.7, but it fails in 11.7. Could anyone give me some advice? Otherwise, we might change too many codes in our application
Sorry Thomas, there is a lot of stuff going on, and the class A is started from another program with
CLASS A INHERITS Ahelper USE-WIDGET-POOL:.....
FINALLY:
DELETE OBJECT B NO-ERROR.
END FINALLY.
The same is within class B, everything is deleted.
The somestuff is a Find that sends back the result to B. I have done a workaround by putting the same method in B so that it will work, but I would like to have that method in A....
..Just want to know, has the original issue that dynamic-function not work for a source-procedure of a class been resolved..
Thanks!
|
a
Ups... sorry Mike, the FINALLY Block is done in the Method...
OK. Still confused as there are too many moving parts coming and going here. I hope you realize that a FINALLY block is NOT a destructor block. It will run immediately after the main body of the block completes. Could that be your problem?
You could try actually debugging this! Step through each statement and watch the value of your variables. Or use the Dynamic Object monitor and look to see what objects are coming and going at each statement. I'm pretty convinced it is something in your code that is causing the problem, and not a bug.
Update from Progress Community
Laura Stern OK. Still confused as there are too many moving parts coming and going here. I hope you realize that a FINALLY block is NOT a destructor block. It will run immediately after the main body of the block completes. Could that be your problem?
You could try actually debugging this! Step through each statement and watch the value of your variables. Or use the Dynamic Object monitor and look to see what objects are coming and going at each statement. I'm pretty convinced it is something in your code that is causing the problem, and not a bug.
You received this notification because you subscribed to the forum. To unsubscribe from only this thread, go here.
Flag this post as spam/abuse.
Update from Progress Community
Peter Judge You should NOT use SOURCE-PROCEDURE for classes. You should not use it to determine what the instantiating class is. It may do what you want today, but it is not intended to do that.If you want to know what class instantiated/NEW’ed this class, then pass that value in via a property or constructor.If you want to publish something to the whole session, do a PUBLISH “event” IN a session super-procedure.You received this notification because you subscribed to the forum. To unsubscribe from only this thread, go here.
Flag this post as spam/abuse.
It seems that in the 11.7 progress, source-procedure for class is not supported. Our codes works well before 11.7 progress, we use the Dynamic-Function(method name, source-procedure for a class) in all our codes... Passing the instance will be a big trouble to change our codes... Source-procedure is only changed in 11,7, I am not sure whether progress is intended to change the Source-procedure in 11.7 or only by mistakes during the 11.7
This Community post has discussed several aspects of your issue and has identified an object-oriented approach for resolving it. However, to your question as to whether the ABL supports the use of source-procedure for classes the answer is, no it does not.
During the initial design of the object-oriented extensions to the ABL, the development team discussed whether the invocation syntax should work for both the procedural ABL and the object-oriented ABL. This approach was rejected. It was for that reason that an ABL application RUNs an external procedure and NEWs a class and invokes an internal procedure via the “RUN IN” syntax and invokes a method via object-reference: method () syntax. The intent was to allow both coding models to work seamlessly together but with different syntax.
Your approach of using SOURCE-PROCEDURE to get access to the procedure handle underlying our implementation of an object instance took advantage of a bug in our implementation. It was never our intent to support for such a facility. The regression you observed was caused by a performance related change we made in 11.7.0.
The OpenEdge development team needs to investigate this issue to close this loop hole in the invocation logic ensuring that an application cannot invoke a method using procedural constructs.
