On Mon, 14 Feb 2005 10:26:02 -0800, Craig McClanahan <[EMAIL PROTECTED]> wrote:
> On Mon, 14 Feb 2005 11:34:28 -0600, Hubert Rabago <[EMAIL PROTECTED]> wrote:
> >
> > formatting/parsing.  We still need that cache for invalid values (such
> > as the 1a3 entered for a numeric field).
> >
> 
> Not if you used JSF :-).  The components take care of that for you,
> which is how Shale can let you bind a text field to a property of type
> "int" and not lose the ability to redisplay invalid data.

True, but we're discussing Struts Classic here :-), which is on a
different path from Shale.  (Isn't imitation the greatest compliment?)
 Now that you mention it, I gotta remember to find a doc which would
explain to me how JSF decides how to format dates and numbers.  I'm
curious.

> 
> As written, the default implementation of the command execution logic
> does indeed require the executed objects to implement the Command
> interface.  While that could be changed, at some point I think you
> really start blurring the lines about what the boundaries of your
> design pattern are, and what semantic and timing expectations are
> expected.
> 
> One could argue, for example, that one of the best things (from a
> flexibility viewpoint) and one of the worst things (from a debugging
> viewpoint) about writing JavaScript code is that fact that type safety
> is not imposed.  Decoupling "command" from the Command interface would
> be a step in the direction of picking up this very two-edged sword.
> 

I think the benefit would outweigh the disadvantage, and allow
commons-chain to be used where it couldn't have before.

> Craig
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to