Re: [boost] regex problem

From: Pavol Droba (droba_at_[hidden])
Date: 2003-09-21 10:04:32

On Sun, Sep 21, 2003 at 12:41:33PM +0100, John Maddock wrote:
> > I'm aware of truncation. The code I'm working on is the string_algo
> > library.
> > And part of it wraps regex in a specific way. In the version v3 of regex,
> > there
> > was no restriction in the char types of regex and the container. I only
> > need
> > to know if this feature is still supported or deprecated (I see that is
> > should
> > be supported), so I can adjust interfaces accordingly.
> > 
> I guess you could call it unofficially supported - there are two other
> problem you will face:
> 
> A ton of compiler warnings.
> Signed/unsigned conversion issues.

I see. That's fine with me. I don't want to encourage this kind of usage, 
but I don't want to make unnecessary restriction too.

> 
> If this is a string lib shouldn't it be using character types?

This is partialy true. However, there is a problem of strictly defining 
what is a characted type. For some unicode encodings it can be char,
wchar_t, int of even something else. That's the reson why there is no
strict restriction on the characted type in the string_algo lib.

Pavol.