[vwnc] class and method generation oddities

> -----Original Message-----
> From: [hidden email] [mailto:[hidden email]] On Behalf Of Steffen Märcker
> Sent: Monday, January 18, 2010 9:38 AM
> To: vwnc
> Subject: [vwnc] class and method generation oddities
>
> Hi,
>
> I am generating some classes and methods programmatically and got some
> questions:
>
> First:
> I define(*) classes A and B in two different namespaces, say NA and NB.
> Then I change class NB.B to have NA.A as superclass (using #superclass:).
> When I query the subclasses of NA.A via #subclasses I get an empty
> collection. The RB doesn't show them either in the hierarchy view if NA.A
> is selected. Is this behavior desired and how can I query subclasses from
> all namespaces?

> Second:
> I define an instance variable a in class A (using #addInstVarName:) and
> afterwards accessors via #compile:classified. If I browse these methods,
> the RB surprisingly insists on variable a being undefined, although the
> generated code works like a charm. Do you have an idea what's causing this?
>
> Perhaps these methods I am using during generation are not meant for this
> purpose? If so, which should I use instead?
>
> Thanks in advance,
> Steffen
>
>
>
>
> (*) using
> #defineClass:superclass:indextedType:private:instanceVariableNames:classInstanceVariableNames:imports:
> category:
> _______________________________________________
> vwnc mailing list
> [hidden email]
> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

> -----Original Message-----
> From: [hidden email] [mailto:[hidden email]] On Behalf Of Steffen Märcker
> Sent: Monday, January 18, 2010 9:38 AM
> To: vwnc
> Subject: [vwnc] class and method generation oddities
>
> Hi,
>
> I am generating some classes and methods programmatically and got some
> questions:
>
> First:
> I define(*) classes A and B in two different namespaces, say NA and NB.
> Then I change class NB.B to have NA.A as superclass (using #superclass:).
> When I query the subclasses of NA.A via #subclasses I get an empty
> collection. The RB doesn't show them either in the hierarchy view if NA.A
> is selected. Is this behavior desired and how can I query subclasses from
> all namespaces?

>Hi,
>
>I am generating some classes and methods programmatically and got some  
>questions:
>
>First:
>I define(*) classes A and B in two different namespaces, say NA and NB.  
>Then I change class NB.B to have NA.A as superclass (using #superclass:).  
>When I query the subclasses of NA.A via #subclasses I get an empty  
>collection. The RB doesn't show them either in the hierarchy view if NA.A  
>is selected. Is this behavior desired and how can I query subclasses from  
>all namespaces?
>
>Second:
>I define an instance variable a in class A (using #addInstVarName:) and  
>afterwards accessors via #compile:classified. If I browse these methods,  
>the RB surprisingly insists on variable a being undefined, although the  
>generated code works like a charm. Do you have an idea what's causing this?
>
>Perhaps these methods I am using during generation are not meant for this  
>purpose? If so, which should I use instead?
>
>Thanks in advance,
>Steffen
>
>
>
>
>(*) using  
>#defineClass:superclass:indextedType:private:instanceVariableNames:classInstanceVariableNames:imports:category:
>_______________________________________________
>vwnc mailing list
>[hidden email]
>http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
>
>
>  
>

>
>> -----Original Message-----
>> From: [hidden email] [mailto:[hidden email]] On  
>> Behalf Of Steffen Märcker
>> Sent: Monday, January 18, 2010 9:38 AM
>> To: vwnc
>> Subject: [vwnc] class and method generation oddities
>>
>> Hi,
>>
>> I am generating some classes and methods programmatically and got some
>> questions:
>>
>> First:
>> I define(*) classes A and B in two different namespaces, say NA and NB.
>> Then I change class NB.B to have NA.A as superclass (using  
>> #superclass:).
>> When I query the subclasses of NA.A via #subclasses I get an empty
>> collection. The RB doesn't show them either in the hierarchy view if  
>> NA.A
>> is selected. Is this behavior desired and how can I query subclasses  
>> from
>> all namespaces?
>
> You have a bug. I think you should use #assignSuperclass: instead of  
> #superclass:.
> Browse the messages and senders that manipulate the superclass and  
> subclasses variables.
>
> You can use the RB and code reader as guides to learn how to  
> programmatically build and
> manipulate classes and their composition.
>
>> Second:
>> I define an instance variable a in class A (using #addInstVarName:) and
>> afterwards accessors via #compile:classified. If I browse these methods,
>> the RB surprisingly insists on variable a being undefined, although the
>> generated code works like a charm. Do you have an idea what's causing  
>> this?
>>
>> Perhaps these methods I am using during generation are not meant for  
>> this
>> purpose? If so, which should I use instead?
>>
>> Thanks in advance,
>> Steffen
>>
>>
>>
>>
>> (*) using
>> #defineClass:superclass:indextedType:private:instanceVariableNames:classInstanceVariableNames:imports:
>> category:
>> _______________________________________________
>> vwnc mailing list
>> [hidden email]
>> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
>
>
> Terry
> ===========================================================
> Terry Raymond
> Crafted Smalltalk
> 80 Lazywood Ln.
> Tiverton, RI  02878
> (401) 624-4517      [hidden email]
> <http://www.craftedsmalltalk.com>
> ===========================================================
>

> -----Original Message-----
> From: Steffen Märcker [mailto:[hidden email]]
> Sent: Monday, January 18, 2010 12:52 PM
> To: Terry Raymond; 'vwnc'
> Subject: Re: [vwnc] class and method generation oddities
>
> Thank you for pointing out the wrong method.
>
> However, one strange bug remains: Why does #compile:classified: not
> respect already defined instance variables in my case? I am doing
> something like:
>
> aClass := aNamespace defineClass:supe ...
> aClass addInstVarName: #myVar. "an inspector confirms the existence of
> myVar"
> aClass
> compile:
> 'myVar
> ^myVar'
> classified:
> 'accessing'.
>
> The method is created but refers to an undeclared variable 'myVar' instead
> of the already defined instance variable.
> Do you have a clue what my mistake might be?
>
> THX,
> Steffen
>
>
> Am 18.01.2010, 16:14 Uhr, schrieb Terry Raymond
> <[hidden email]>:
>
> >
> >> -----Original Message-----
> >> From: [hidden email] [mailto:[hidden email]] On
> >> Behalf Of Steffen Märcker
> >> Sent: Monday, January 18, 2010 9:38 AM
> >> To: vwnc
> >> Subject: [vwnc] class and method generation oddities
> >>
> >> Hi,
> >>
> >> I am generating some classes and methods programmatically and got some
> >> questions:
> >>
> >> First:
> >> I define(*) classes A and B in two different namespaces, say NA and NB.
> >> Then I change class NB.B to have NA.A as superclass (using
> >> #superclass:).
> >> When I query the subclasses of NA.A via #subclasses I get an empty
> >> collection. The RB doesn't show them either in the hierarchy view if
> >> NA.A
> >> is selected. Is this behavior desired and how can I query subclasses
> >> from
> >> all namespaces?
> >
> > You have a bug. I think you should use #assignSuperclass: instead of
> > #superclass:.
> > Browse the messages and senders that manipulate the superclass and
> > subclasses variables.
> >
> > You can use the RB and code reader as guides to learn how to
> > programmatically build and
> > manipulate classes and their composition.
> >
> >> Second:
> >> I define an instance variable a in class A (using #addInstVarName:) and
> >> afterwards accessors via #compile:classified. If I browse these methods,
> >> the RB surprisingly insists on variable a being undefined, although the
> >> generated code works like a charm. Do you have an idea what's causing
> >> this?
> >>
> >> Perhaps these methods I am using during generation are not meant for
> >> this
> >> purpose? If so, which should I use instead?
> >>
> >> Thanks in advance,
> >> Steffen
> >>
> >>
> >>
> >>
> >> (*) using
> >>
> #defineClass:superclass:indextedType:private:instanceVariableNames:classInstanceVariableNames:imports:
> >> category:
> >> _______________________________________________
> >> vwnc mailing list
> >> [hidden email]
> >> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
> >
> >
> > Terry
> > ===========================================================
> > Terry Raymond
> > Crafted Smalltalk
> > 80 Lazywood Ln.
> > Tiverton, RI  02878
> > (401) 624-4517      [hidden email]
> > <http://www.craftedsmalltalk.com>
> > ===========================================================
> >