Boost logo

Boost Users :

Subject: Re: [Boost-users] Unique parameter values for interfaces
From: Ryan McConnehey (mccorywork_at_[hidden])
Date: 2009-05-08 10:55:33


Peter Bartlett wrote:
> In addition to tagging, you can also look into whether true_typedef is
> suitable. Matthew Wilson's STLSoft has an implementation.

true_typedef seems like a really good choice. I'm unclear what the
STLSOFT_GEN_OPAQUE macro really provides. The macro is defined as follows.

#define STLSOFT_GEN_OPAQUE (type) typedef struct
__stlsoft_htype##type{ int i; } const* type;

So to use the STLSoft version I have to use the macro and then provide
the typedef. It would look like this.

STLSOFT_GEN_OPAQUE (ConfigName_u)
typedef stlsoft::true_typedef<std::string, ConfigName_u> ConfigName;

How is this better than just declaring a struct in the typedef itself.
It would look like this.

typedef stlsoft::true_typedef<std::string, struct ConfigName_u>
ConfigName;

What does having a pointer to a struct provide that a normal struct
doesn't?

Ryan



Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net