On 9 Dec 2008, at 11:27, Damien Pollet wrote:

> On Sun, Dec 7, 2008 at 15:16, David Chisnall <[EMAIL PROTECTED]>  
> wrote:
>> On 7 Dec 2008, at 13:41, Günther Noack wrote:
>>>      * true and false should be defined
>>
>> I think Truls was going to do this - they just need mapping to 1  
>> and 0
>> integer literals in the front end.
>
> T_T

Yes, it's not pure Smalltalk, but it done to make integration with  
Objective-C easier.  We can easily add a Boolean object with True and  
False subclasses and as long as they implement -intValue and - 
boolValue returning 0 and 1 respectively they can be used  
interchangeably with the ObjC values.

>>>      * hexadecimal escape characters in strings
>>
>> I am not sure what this is useful for - Smalltalk code is unicode so
>> you rarely need escape sequences.  Patches welcome, of course.
>
> That would be standard library (smalltalk) code:
> '\\ \t \n' backslashEscaped
>
> After a quick glance in Squeak, I found #expandMacrosWithArguments:
> but it's a full-blown sprintf-like DSL,
> unescapePercentsWithTextEncoding: which is more like an hex-sequence
> expander, and encoding conversions squeak-to-mac, squeak-to-utf8, or
> #encodeToHTTP.
>
> In any case, the C notation is just a string encoding among others,
> and should be treated as such. The only escape sequence handled by the
> parser is the quote doubling.

That makes sense.  NSString already has methods for doing most of  
these expansions, so there is no reason for putting them in the parser  
other than speed (and since it slows down parsing speed...).

David
_______________________________________________
Etoile-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/etoile-discuss

Répondre à