Re: [PATCH] Fix libstdc++/6642

[Phil Edwards <phil at jaj dot com>]

On Sun, May 26, 2002 at 09:32:35PM +0200, Paolo Carlini wrote:
> 2- I didn't wrap operator- in #ifdef _GLIBCPP_RESOLVE_LIB_DEFECTS since, 
> otherwise, all the other non-member comparison operators should be 
> wrapped too and provided also in the member form #ifndef. More 
> generally, I seem to remember some mailings from Gaby maintaining that 
> in fact the <Ready> DR are in fact part of the standard for all 
> practical purposes.

At one point I think the idea was to support both "a C++ library as
specified by the Standard," and also, "a C++ library as specified by
the Standard with the latest round of committee fixes and suggestions."
Defining this macro, or not, would switch between them.

However, if _GLIBCPP_RESOLVE_LIB_DEFECTS is not defined, we cannot actually
compile the library.  Therefore there seems little point in continuing to
guard changes.

So, it seems that this name has changed from something that switches between
"strict compliance" and "strict compliance with bugfixes," to something
that we can search for when asking the question, "which committee changes
have we already implemented?"  And that seems to be okay; nobody has ever
complained about not being able to build libstdc++ after removing the
#define from c++config.

As an example, look in include/std/std_bitset.h.


My suggestion would be to add _GLIBCPP_RESOLVE_LIB_DEFECTS in a comment
along with the DR number in the code you add.  For major bonus points,
update docs/html/ext/howto.html#5 before I get around to it.  :-)

Thanks very much for working on this, Paolo.  The iterator/const_iterator 
mess has always been a thorn in my side.


Phil

-- 
If ye love wealth greater than liberty, the tranquility of servitude greater
than the animating contest for freedom, go home and leave us in peace.  We seek
not your counsel, nor your arms.  Crouch down and lick the hand that feeds you;
and may posterity forget that ye were our countrymen.            - Samuel Adams