This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: shadowed rel_ops operators


On Sat, Jan 20, 2001 at 05:50:35PM -0600, amep@softhome.net wrote:
> From: Nathan Myers <ncm@nospam.cantrip.org>
> > 
> > The rel_ops operators are not useful.  They were a mistake.  If you 
> > insist on using them, you will have trouble.  Nobody will care, because 
> > you asked for it.  
> 
> 1) I noticed that some of the backwards compatibility .h's bring these
> operators into std::. This makes it so you can't use libraries that
> include these headers along with code that uses the new features of
> libstdc++-v3. Is there a way around this? It seems to me that in most
> cases all the backwards compatibility. you need is mapping iterator.h
> to iterator, ... . Maybe there should be a #define that turns off all
> the extra backwards compatibility stuff. Just a thought. And maybe a
> bad one.

The backward-compatibility stuff has languished.  Yes, it would be
good to make it play better with the standard stuff, to make an
easier transition.  Can you send patches?
 
> 2) It seems to me that rel_ops is a very good idea that went wrong
> somehow. Is there anything that replaces them in anyway? It seems to
> me they could be very useful if they where designed so they could be
> confined easily.

The rel_ops operators are a good idea for some language other than C++.
To my knowledge, nobody has been able to make them work right given the
rules for template overloading in C++.  The good news is that they would 
only have been a notational convenience, and we can live without them.

Nathan Myers
ncm at cantrip dot org

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]