Copy constructor access check while initializing a reference

4 msg	SPEC / testsuite results for disabling SFTs and...
2 msg	Re: static array with constant size
4 msg	Help with GCC on Cygwin
3 msg	[4.3/4.4]: PATCH: PR target/35453: nmmintrin.h ...
18 msg	Constrain valid arguments to BIT_FIELD_REF
3 msg	Re: Interoperability of Fortran array and C vec...
2 msg	Information regarding issue with While Loop wit...
1 msg	Getting GCC to always dllimport vtables on X86?
3 msg	Swing replacements

Copy constructor access check while initializin...
\ Peter A. Felvegi (4 Mar 2008)
. \ Gabriel Dos Reis (4 Mar 2008)

34 msg	atomic accesses
44 msg	gcc-4.1-20080303 is now available
3 msg	compiler slowdown in 4.3 development
5 msg	GCC 4.4 schedule
3 msg	GCC 4.3.0 Status Report (2008-03-03)
4 msg	Google Summer of Code 2008
2 msg	start.encap target in gcc/Makefile.in
2 msg	Release planning for GCC 4.4?
2 msg	Stan Shebs appointed maintainer for Darwin, Obj...
2 msg	Invalid address after reload

Subject:	Copy constructor access check while initializing a reference
Group:	Gcc
From:	Peter A. Felvegi
Date:	4 Mar 2008

hello,

i've found this in the known non-bugs list
(http://gcc.gnu.org/bugs.html#known), after running into the issue. gcc
3.4-4.2 gives a compile error, but 4.3 compiles it. is this a
regression, or the rules were relaxed somewhat (c++0x?) ? i checked the
changelog, but couldn't find any relevant entries there.

could someone explain the rationale behind this extra access check,
please? i'm puzzled, since the copy ctor is not needed and is not called
from the generated code.

regards, p