SV: StringResourceModels doesn't seem to detach properly

2012-01-04 Thread Einar Bjerve
Hi,

Well, StringResourceModel#detach() is called, but since the SRM isn't attached, 
StringResourceModel#onDetach() is never called and the result is that the 
property substitution model isn't detached. I'll demonstrate with a bit of code:


IModel substitutionModel1 = ...;
IModel substitutionModel2 = ...;

add( new Label( id1, new StringResourceModel( resourceKey1, getPage(), 
substitutionModel1) ) ); // substitutionModel1 detaches properly
add( new Label( id2, new StringResourceModel( resourceKey2, 
substitutionModel2) ) ); // substitutionModel2 never detaches


In the above snippet, substitutionModel1 will be detached, but 
substitutionModel2 will not be detached when its label is detached. In 1.4.x 
both substitutionmodels would be detached

The reason this happens is that StringResourceModel is wrapped on assignment in 
a StringResourceModel.AssignmentWrapper. This assignmentwrapper never calls 
StringResourceModel#getObject() if the component constructor argument is null. 
In 1.4 StringResourceModel#getObject() would always be called (which means 
load() is called, and the enclosing SRM is attached).

The result is that the AssignmentWrapper is attached, but the enclosing 
StringResourceModel isn't. When StringResourceModel.AssignmentWrapper#detach() 
attempts to calls StringResourceModel#detach() the if(!attached) test (in 
StringResourceModel #detach()) returns false, which results in that 
StringResourceModel#onDetach() is never called.

Since detaching of property substitution models and detachable parameters is 
done in StringResourceModel#onDetach(), these objects won't be detached.


I believe this is a bug, as it was changed between 1.4 and 1.5. In 1.4 it was 
consistent, in 1.5 it isn't.


Best regards
Einar Bjerve


-Opprinnelig melding-
Fra: Martin Grigorov [mailto:mgrigo...@apache.org] 
Sendt: 3. januar 2012 16:36
Til: users@wicket.apache.org
Emne: Re: StringResourceModels doesn't seem to detach properly

Hi,

It is intended.
IModel#detach() is called only if this model is assigned to a component.
Since no one calls #detach() on the StringResourceModel it has no idea
when to call detach() on its inner model.
You need to call SRM.detach() in #onDetach() in the component where it
is used. Then it will detach its inner model too.

On Tue, Jan 3, 2012 at 5:30 PM, Einar Bjerve einar.bje...@visma.no wrote:
 If a StringResourceModel contains a model for property substitutions, and 
 there has not been assigned a component it is relative to on construction 
 time, it will not detach the property substitution model.

 If there is assigned a component on construction time, the property 
 substitution model is detached with the StringResourceModel.

 This behavior seems kind of inconsistent to me. Is this intended and correct?


 Best regards
 Einar Bjerve



-- 
Martin Grigorov
jWeekend
Training, Consulting, Development
http://jWeekend.com

-
To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org
For additional commands, e-mail: users-h...@wicket.apache.org



Re: StringResourceModels doesn't seem to detach properly

2012-01-04 Thread Martin Grigorov
Hi,

Explained this way it looks like a bug indeed.
File a ticket. If you find the differences in SRM in 1.4 vs. 1.5 then
please attach a patch too.
Thanks!

On Wed, Jan 4, 2012 at 10:44 AM, Einar Bjerve einar.bje...@visma.no wrote:
 Hi,

 Well, StringResourceModel#detach() is called, but since the SRM isn't 
 attached, StringResourceModel#onDetach() is never called and the result is 
 that the property substitution model isn't detached. I'll demonstrate with a 
 bit of code:


 IModel substitutionModel1 = ...;
 IModel substitutionModel2 = ...;

 add( new Label( id1, new StringResourceModel( resourceKey1, getPage(), 
 substitutionModel1) ) ); // substitutionModel1 detaches properly
 add( new Label( id2, new StringResourceModel( resourceKey2, 
 substitutionModel2) ) ); // substitutionModel2 never detaches


 In the above snippet, substitutionModel1 will be detached, but 
 substitutionModel2 will not be detached when its label is detached. In 1.4.x 
 both substitutionmodels would be detached

 The reason this happens is that StringResourceModel is wrapped on assignment 
 in a StringResourceModel.AssignmentWrapper. This assignmentwrapper never 
 calls StringResourceModel#getObject() if the component constructor argument 
 is null. In 1.4 StringResourceModel#getObject() would always be called (which 
 means load() is called, and the enclosing SRM is attached).

 The result is that the AssignmentWrapper is attached, but the enclosing 
 StringResourceModel isn't. When 
 StringResourceModel.AssignmentWrapper#detach() attempts to calls 
 StringResourceModel#detach() the if(!attached) test (in StringResourceModel 
 #detach()) returns false, which results in that 
 StringResourceModel#onDetach() is never called.

 Since detaching of property substitution models and detachable parameters is 
 done in StringResourceModel#onDetach(), these objects won't be detached.


 I believe this is a bug, as it was changed between 1.4 and 1.5. In 1.4 it was 
 consistent, in 1.5 it isn't.


 Best regards
 Einar Bjerve


 -Opprinnelig melding-
 Fra: Martin Grigorov [mailto:mgrigo...@apache.org]
 Sendt: 3. januar 2012 16:36
 Til: users@wicket.apache.org
 Emne: Re: StringResourceModels doesn't seem to detach properly

 Hi,

 It is intended.
 IModel#detach() is called only if this model is assigned to a component.
 Since no one calls #detach() on the StringResourceModel it has no idea
 when to call detach() on its inner model.
 You need to call SRM.detach() in #onDetach() in the component where it
 is used. Then it will detach its inner model too.

 On Tue, Jan 3, 2012 at 5:30 PM, Einar Bjerve einar.bje...@visma.no wrote:
 If a StringResourceModel contains a model for property substitutions, and 
 there has not been assigned a component it is relative to on construction 
 time, it will not detach the property substitution model.

 If there is assigned a component on construction time, the property 
 substitution model is detached with the StringResourceModel.

 This behavior seems kind of inconsistent to me. Is this intended and correct?


 Best regards
 Einar Bjerve



 --
 Martin Grigorov
 jWeekend
 Training, Consulting, Development
 http://jWeekend.com

 -
 To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org
 For additional commands, e-mail: users-h...@wicket.apache.org




-- 
Martin Grigorov
jWeekend
Training, Consulting, Development
http://jWeekend.com

-
To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org
For additional commands, e-mail: users-h...@wicket.apache.org



Re: StringResourceModels doesn't seem to detach properly

2012-01-04 Thread Einar Bjerve
Done

WICKET-4323

Best regards
Einar

Den 4. jan. 2012 kl. 09:49 skrev Martin Grigorov mgrigo...@apache.org:

 Hi,
 
 Explained this way it looks like a bug indeed.
 File a ticket. If you find the differences in SRM in 1.4 vs. 1.5 then
 please attach a patch too.
 Thanks!
 
 On Wed, Jan 4, 2012 at 10:44 AM, Einar Bjerve einar.bje...@visma.no wrote:
 Hi,
 
 Well, StringResourceModel#detach() is called, but since the SRM isn't 
 attached, StringResourceModel#onDetach() is never called and the result is 
 that the property substitution model isn't detached. I'll demonstrate with a 
 bit of code:
 
 
 IModel substitutionModel1 = ...;
 IModel substitutionModel2 = ...;
 
 add( new Label( id1, new StringResourceModel( resourceKey1, getPage(), 
 substitutionModel1) ) ); // substitutionModel1 detaches properly
 add( new Label( id2, new StringResourceModel( resourceKey2, 
 substitutionModel2) ) ); // substitutionModel2 never detaches
 
 
 In the above snippet, substitutionModel1 will be detached, but 
 substitutionModel2 will not be detached when its label is detached. In 1.4.x 
 both substitutionmodels would be detached
 
 The reason this happens is that StringResourceModel is wrapped on assignment 
 in a StringResourceModel.AssignmentWrapper. This assignmentwrapper never 
 calls StringResourceModel#getObject() if the component constructor argument 
 is null. In 1.4 StringResourceModel#getObject() would always be called 
 (which means load() is called, and the enclosing SRM is attached).
 
 The result is that the AssignmentWrapper is attached, but the enclosing 
 StringResourceModel isn't. When 
 StringResourceModel.AssignmentWrapper#detach() attempts to calls 
 StringResourceModel#detach() the if(!attached) test (in StringResourceModel 
 #detach()) returns false, which results in that 
 StringResourceModel#onDetach() is never called.
 
 Since detaching of property substitution models and detachable parameters is 
 done in StringResourceModel#onDetach(), these objects won't be detached.
 
 
 I believe this is a bug, as it was changed between 1.4 and 1.5. In 1.4 it 
 was consistent, in 1.5 it isn't.
 
 
 Best regards
 Einar Bjerve
 
 
 -Opprinnelig melding-
 Fra: Martin Grigorov [mailto:mgrigo...@apache.org]
 Sendt: 3. januar 2012 16:36
 Til: users@wicket.apache.org
 Emne: Re: StringResourceModels doesn't seem to detach properly
 
 Hi,
 
 It is intended.
 IModel#detach() is called only if this model is assigned to a component.
 Since no one calls #detach() on the StringResourceModel it has no idea
 when to call detach() on its inner model.
 You need to call SRM.detach() in #onDetach() in the component where it
 is used. Then it will detach its inner model too.
 
 On Tue, Jan 3, 2012 at 5:30 PM, Einar Bjerve einar.bje...@visma.no wrote:
 If a StringResourceModel contains a model for property substitutions, and 
 there has not been assigned a component it is relative to on construction 
 time, it will not detach the property substitution model.
 
 If there is assigned a component on construction time, the property 
 substitution model is detached with the StringResourceModel.
 
 This behavior seems kind of inconsistent to me. Is this intended and 
 correct?
 
 
 Best regards
 Einar Bjerve
 
 
 
 --
 Martin Grigorov
 jWeekend
 Training, Consulting, Development
 http://jWeekend.com
 
 -
 To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org
 For additional commands, e-mail: users-h...@wicket.apache.org
 
 
 
 
 -- 
 Martin Grigorov
 jWeekend
 Training, Consulting, Development
 http://jWeekend.com
 
 -
 To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org
 For additional commands, e-mail: users-h...@wicket.apache.org
 

-
To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org
For additional commands, e-mail: users-h...@wicket.apache.org



StringResourceModels doesn't seem to detach properly

2012-01-03 Thread Einar Bjerve
If a StringResourceModel contains a model for property substitutions, and there 
has not been assigned a component it is relative to on construction time, it 
will not detach the property substitution model.

If there is assigned a component on construction time, the property 
substitution model is detached with the StringResourceModel.

This behavior seems kind of inconsistent to me. Is this intended and correct?


Best regards
Einar Bjerve


Re: StringResourceModels doesn't seem to detach properly

2012-01-03 Thread Martin Grigorov
Hi,

It is intended.
IModel#detach() is called only if this model is assigned to a component.
Since no one calls #detach() on the StringResourceModel it has no idea
when to call detach() on its inner model.
You need to call SRM.detach() in #onDetach() in the component where it
is used. Then it will detach its inner model too.

On Tue, Jan 3, 2012 at 5:30 PM, Einar Bjerve einar.bje...@visma.no wrote:
 If a StringResourceModel contains a model for property substitutions, and 
 there has not been assigned a component it is relative to on construction 
 time, it will not detach the property substitution model.

 If there is assigned a component on construction time, the property 
 substitution model is detached with the StringResourceModel.

 This behavior seems kind of inconsistent to me. Is this intended and correct?


 Best regards
 Einar Bjerve



-- 
Martin Grigorov
jWeekend
Training, Consulting, Development
http://jWeekend.com

-
To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org
For additional commands, e-mail: users-h...@wicket.apache.org