Re: JavaMemberNode optimisation

On Dec 13, 2010, at 5:58, Thomas Leonard wrote:

> This change caused a slight side-effect: causeway does this:
>
>  def editorShellListener :ShellListener := def _ { ... }
>
>  editorShell.addShellListener(editorShellListener)
>  ...
>  editorShell.removeShellListener(editorShellListener)
>
> This relies on editorShellListener being the same (deflector) object
> each time. However, E.as always unwraps the deflector and creates a  
> new
> one.


(Note that as a general invariant, it must be impossible to  
distinguish between objects which are considered the same by  
Equalizer. Deflectors and resolved promises are both wrappers that can  
cause trouble in these cases.)

Merely avoiding making a new deflector would still result in this  
being a problem:

  def editorShellListener := def _ { ... }
  editorShell.addShellListener(editorShellListener :ShellListener)
  editorShell.removeShellListener(editorShellListener :ShellListener)

since a deflector is constructed twice.

The fix which comes to mind is to use a weak table to ensure that  
there is at most one deflector for each combination of E-object and  
Java interface.

Or, we could make Deflectors not considered E-same as their referents;  
this would be safe since it happens at the moment of explicit  
coercion, but less transparent.

-- 
Kevin Reid                                  <http://switchb.org/kpreid/>