D still has macros ..

[James Dunne]

Hasan Aljudy wrote:
> Isn't __FILE__ really just a macro? it's processed before tokenizing! 
> according to the specc, if the lexer sees __FILE__ or __LINE__ or such, 
> it should convert it to an appropriate string/number/whatever before 
> tokenizing it. Hence, it's a macro.
> 

__FILE__ is processed (like __LINE__) as a special reserved identifier 
which tokenizes to a string representing the current source file being 
tokenized.

It is _not_ processed _before_ tokenization as you say; it is processed 
during tokenization and is converted to an equivalent string-literal token.

> This implies that __FILE__ and such cannot be used in mixins or 
> templates to mimic things like the following C macro:

I'm not sure how you concluded that... Are you asking to which source 
file the __FILE__ reserved identifier represents when using a mixin 
across file boundaries?

> 
> #define log(s) logprintf(__FILE__ ": " __FUNCTION__ ": %s", s)
> 
> Shouldn't these things become templates (or something like that) so that 
> we can use them like the above C macro, but with D templates?

Perhaps you're hinting at a new set of reserved words like 
__INSTANTIATED_FILE__ and __INSTANTIATED_FUNCTION__ so that they expand 
to the file/line number where the template/mixin is instantiated at?

-- 
Regards,
James Dunne