Re: [boost] String formatting library: interest?

From: Samuel Krempp (krempp_at_[hidden])
Date: 2002-05-15 02:56:02

Le lun 13/05/2002 à 16:13, Mattias Flodin a écrit :
> have the patience to follow that debate, but I assume it either ran
> out of steam when people got tired of arguing, or that consensus was
> reached. Depending on which, % overloading is 1. already agreed on as
> being OK, or 2. something that is so difficult to agree on that we're
> going to annoy people no matter which path we choose, so we might
> aswell just leave it be.

it was 2. 
Most people objecting to using % will always ask for an other solution
(e.g. your 'arg(..)' member function, or simply N-ary function call
(fmt, arg1, arg2, .., argN)  )

I favored '%' other those 2 solutions, and it seemed the majority was ok
with it, as a consensus.
I mean, although not the best choice for the majority, '%' would be the
winner if we had asked people to vote for the different solutions, using
condorcet system. (which means : '%' wins when facing each of the other
solutions in 1vs1 elections)
[ the fact that we had a very strange election recently in France
explains why I'm inclined to look at things with a voting-system angle..
;-) ]

> > - It doesn't substitute/enhance std::string, but it acts as an external
> > library. You can either feed the result of format() to a stream, or use the
> > member .str() to get the string.
> 
> Yes, whether this should be so may be an interesting point to discuss.

In fact this request was never raised when debating, so I believed that
all participants preferred having the library clearly distinct from
string.

Personnally, I see no benefit to placing the library so close to string.

-- 
Sam