[c++] bug with new inlining mechanism on v-3, commentary

[Nathan Sidwell]

Gabriel Dos_Reis wrote:
> 
> Benjamin Kosnik <bkoz@cygnus.com> writes:
> | I'm trying to compile the CVS libstdc++-3 sources with
> | -O2, and getting an ICE with both gcc-2.95.2 and CVS egcs (after your
> | enable inlining patch went in Mark). This is hosted on x86/linux.
> 
> I saw similar behaviour without inling-on-tree enabled a few days ago.
I just tried it after Mark's reenabled inlining, and it still fails :-(
Both linux-i686 & sparc-solaris2.7, as one might expect for a front
end problem.

nathan82.C: In function `int ExtendFoos ()':
nathan82.C:27: Internal compiler error in `expand_fixup_region_end', at except.c:1602
Please submit a full bug report.

Here's the test case, I'd like to install it

ok?

nathan
-- 
Dr Nathan Sidwell :: Computer Science Department :: Bristol University
Never hand someone a gun unless you are sure where they will point it
nathan@acm.org  http://www.cs.bris.ac.uk/~nathan/  nathan@cs.bris.ac.uk
// Build don't link:

// Copyright (C) 1999 Free Software Foundation, Inc.
// Contributed by Nathan Sidwell 29 Nov 1999 <nathan@acm.org>

// Special g++ Options: -ansi -pedantic-errors -O2
// crash test - XFAIL *-*-*

struct Foo
{
  inline ~Foo ();
};

inline void Wibble (int) throw ()
{
}

inline Foo::~Foo ()
{
  Wibble (6);
}

int ExtendFoos ()
{
  Foo  tmp;
  return 0;
} // ICE here