[TCLCORE] Re: TIP#181: Add a [namespace unknown] Command

[Neil M. <ne...@cs...>]

On 28 Jun 2005, at 18:15, dg...@us... wrote:
> I'd like to revise TIP 181 a bit.
>
> It currently proposes that the default
> [namespace unknown] handler for all
> namespaces is the command [::unknown]
> in order to keep compatibility.
>
> I'd rather have the default handler
> be the command [::tcl::DefaultUnknown]
> and have that command in turn check
> for the existence of [::unknown] and use
> it if present.

That could be done easily enough.

> This change would avoid the creation of
> the global [::unknown] command by Tcl
> itself.  Besides being cleaner in use
> of the global namespace, it's
> in fact easier to write a proc
> [::tcl::DefaultUnknown] than it is to
> write [::unknown]  (have a look at the
> current monster!), because one can make
> use of namespace variables freely without
> worrying about polluting the global namespace.
>
> This change should still retain compatibility,
> but will also allow the current [unknown]
> proc to be simplified quite a bit.

Sounds like a good idea. It can be incorporated into the TIP (and  
patch) easily enough. Feel free to edit the TIP directly to add this.  
The implementation would be simple enough (changing "::unknown" to  
"::tcl::DefaultUnknown" in a couple of places, and then writing the  
new proc itself).

>
> One other nit in the patch documentation
> the syntax is described:
>
>     namespace unknown ?script?
>
> This would be better described
>
>     namespace unknown ?cmdPrefix?
>
> as was done for [interp bgerror].  Experience
> has taught that single callback commands
> are as powerful, and far less trouble than
> supporting entire scripts as callbacks.  A
> single command is what the patch supports,
> so the docs should clearly describe that.

OK, looks like a bug in my patch, I'll get that fixed. I'll hold off  
updating the patch until after a vote on this TIP, in case any other  
changes are requested. The TIP itself will be kept up-to-date, of  
course. AIUI, a reference implementation is not actually required for  
a vote on a TIP, so I'd rather not spend too much more time on  
updating and testing the patch until I know the TIP is accepted. TIP  
2 states:

"The TIP should be reviewed and accepted before a reference  
implementation is begun, unless a reference implementation will aid  
people in studying the TIP."

So my 4 patches so far are probably premature anyway!

Cheers,

-- Neil