Alex,
Take a look at Joe's extra/literals. You can do something like:
CONSTANT: morse-code-table $[ H{ { CHAR: a ".-" } } >biassoc ]
Doug
On Apr 19, 2009, at 9:18 PM, Alex Chapman wrote:
> Hi All,
>
> Sometimes I want to define an object in my program which is intended
> to be a constant, but isn't a literal. For example, in the morse
> vocab I want to define morse-code-table, a biassoc which allows
> lookups of ascii characters to morse codes and vice versa. One way I
> could do this:
>
> : morse-code-table ( -- table )
> H{
> { CHAR: a ".-" }
> { CHAR: b "-..." }
> ...
> { CHAR: \s "/" }
> } >biassoc ;
>
> But this is inefficient, because this creates a new biassoc tuple
> from the literal hashtable every time [1]. Maybe instead we could
> define it in a global variable:
>
> DEFER: morse-code-table
> H{
> { CHAR: a ".-" }
> { CHAR: b "-..." }
> ...
> { CHAR: \s "/" }
> } >biassoc \ morse-code-table set-global
>
> : morse-code-table ( -- table )
> \ morse-code-table get-global ;
>
> This works, but what I really want to do is express that this is a
> value to be created at parse time and never modified, something that
> the recent CONSTANT: word does. I can define this as a constant like
> this:
>
> DEFER: morse-code-table
> \ morse-code-table
> H{
> { CHAR: a ".-" }
> { CHAR: b "-..." }
> ...
> { CHAR: \s "/" }
> } >biassoc define-constant
>
> It would be nice if I didn't have to do this manually, but could use
> a modified CONSTANT: syntax like this:
>
> CONSTANT: morse-code-table
> H{
> { CHAR: a ".-" }
> { CHAR: b "-..." }
> ...
> { CHAR: \s "/" }
> } >biassoc ;
>
> Is there a reason CONSTANT: only takes an object literal and not a
> block of code to evaluate, other than the convenience of having a
> shorter syntax in the most common case?
>
> On a related issue, if I want morse-code-table to be constant, how
> can I make sure that it isn't modified? I know the factor philosophy
> (and implementation) is not to hide anything, but I'd like to at
> least throw an error if someone ignorantly tried to modify the
> table. How about this:
>
> ERROR: morse-code-table-is-constant ;
> PREDICATE: is-morse-code-table < biassoc morse-code-table eq? ;
> M: is-morse-code-table set-at morse-code-table-is-constant new throw ;
>
> What I'm doing is defining a class which consists only of the morse-
> code-table object, and throwing an error if anyone tries to change
> it using set-at. I don't think I'll actually bother doing this in
> the real morse vocab.
>
> Alex
>
> [1] If all I wanted was a read-only hashtable then there would be no
> problem here, because the literal hashtable is created only once at
> parse time.
> ------------------------------------------------------------------------------
> Stay on top of everything new and different, both inside and
> around Java (TM) technology - register by April 22, and save
> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
> 300 plus technical and hands-on sessions. Register today.
> Use priority code J9JMT32.
> http://p.sf.net/sfu/p_______________________________________________
> Factor-talk mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/factor-talk
------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today.
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Factor-talk mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/factor-talk
