Re: [boost] Review : format2, yet another 'printf class'

From: Samuel Krempp (krempp_at_[hidden])
Date: 2001-07-01 13:19:18

>  >Is it a (different) problem that dollar-sign is not part
>  >of the language's basic source character set (2.2/1) ?
>  >If that character is unavailable on, say, an Icelandic
>  >keyboard, then there's no trigraph for it either.
>
> Ouch!  Good catch, Greg.

Reading this paragraph of the norm,
I guess the 'Universal character' system could make it possible to use a
'$' on such a keyboard.

really, do you believe there will ever be a platform where it is not
possible to type '$' ?

>  > Would
>  >a library that needs dollar-sign be acceptable in ISO C++?
>
> I doubt it.
>
> IMO, we need to stick to the basic character set.

for this, I replaced $ by # in some parts of the syntax.

currently, to use printf-like direcives, you still need $ to make
positional directives ( %5$x  == 5-th argument in hexa )

this makes another point to consider for the final design of the syntax.
I havnt fixed my opinion yet, but the current state is nearly
satisfactory.

Maybe we can extend the short-style notation so that usual people dont
ever need to use printf-like directives.
Or not.

anyway, it's not a crucial design issue.
The parsing code can be modified with little impact on the rest.

-- 
Sam