On Thu, Dec 13, 2012 at 3:06 AM, Roland Mainz <[email protected]> wrote:
> On Wed, Dec 12, 2012 at 4:02 PM, David Korn <[email protected]> wrote:
>> cc: [email protected]
>> Subject: Re: Re: [ast-developers] [rfe] typeset -B to define a (C99-like)  
>> <stdbool.h> datatype
>> --------
>>
>> Why bother with an alias?  Why not just create Bool as a predefined type?
>
> Erm... I had a couple of reasons:
> 1. All basic ksh93 types modelled after C types (e.g. integer, float,
> compound) are aliases to the real type definitions
> 2. C99 (see the <stdbool.h> manpage quote in my earlier email) says
> "... bool - Expands to _Bool..." ... which IMO makes sense since C99
> allows to |#undef bool| and redefine it to something else. We can do
> the same with $ unalias bool #.
> 3. Irek suggested the same $ unalias bool # for the (very) unlikely
> case ([a1]) that someone already uses "bool" as a command name

Yes, that is IMO a useful feature to migrate legacy code and IMO
mandatory until your proposal for an unset -T (to unset type
definitions) and unset -E (to unset enums) gets implemented. Aside
from that I like the "consistency" argument - your point 1. above.

Irek
_______________________________________________
ast-developers mailing list
[email protected]
http://lists.research.att.com/mailman/listinfo/ast-developers

Reply via email to