[bitc-dev] Replacing typealias

Swaroop Sridhar swaroop at cs.jhu.edu
Wed Feb 9 23:20:36 EST 2005



On Wed, 9 Feb 2005, Jonathan S. Shapiro wrote:

> At present, we have a defining construct
>
> 	(typealias id type-decl)
>
> the purpose of this was to give a convenience name for otherwise unnamed
> types.
>
> The keyword TYPEALIAS stands out as inconsistent. All other defining
> forms begin with "def". I initially resisted DEFTYPE because the form
> doesn't actually define a new type -- it only binds a name to an
> existing type.
>
> However, I am coming to the conclusion that DEFTYPE is visually
> preferable, and the possible interpretation that
>
> 	(deftype nm type)
>
> would introduce a new, named type that was incompatible with the
> existing one seems to have faded out of the idea space in the language.
>
> Unless there are objections, I will rename TYPEALIAS to DEFTYPE.
>

My guess is that it is OK, especially because:

   C says typedef and does typealias
BitC says deftype and does typealias.

Swaroop.



More information about the bitc-dev mailing list