Re: WG: Re: proposed change

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Re: WG: Re: proposed change

Thomas Brodt
That would be another option, but certainly introduces more changes to the behavior of aspects in general, while modified aspects in specs keep the changes "in place."

Thomas


[hidden email] wrote

Personally, I'd prefer if the builder optionally generated aspect adaptor code into the #initialize method.  That would be easier to look at, and, imho, easier to maintain.

Sent from my iPad

On Dec 9, 2010, at 1:12 PM, Thomas Brodt <[hidden email]> wrote:

Dave is so right. Every so often we have the same Problem. Especially because lens makes extensive use of it. 

So please Count my vote As at least +3 ;-)

Thomas

Am 09.12.2010 um 18:15 schrieb "Stevenson, Dave (Contr)" <[hidden email]>:

Currently, specifying an aspect adaptor in the UIPainter saves the model definition in the spec as a concatenated symbol:

                #model: #'someValueModel someAspect’

 

Unfortunately, the result is that the reference to #someAspect is now dynamic. That is, the concatenated symbol is parsed at run time, but the original source reference in the spec method is not statically identifiable to the refactoring engine (a rename refactoring will not consider the concatenated symbol to be a reference to #someAspect, and the reference will not be changed).

 

It seems that small changes to the UIPainter, UIBuilder and related code could continue to handle the old style for backward compatability, but use an array for newly generated specs that could be visible to rename refactorings:

                #model: #(#someValueModel #someAspect)

 

A script could also be provided that would convert the old style to the new in existing spec methods, if the user desires. Will Cincom consider such a change?

 

Stevenson, Dave (contr) <[hidden email]>

Smalltalk Developer

Manufacturing Execution System (MES) / Shop Floor Control
El Segundo, CA

 

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc