On Mon, Nov 25, 2002 at 12:33:11AM +0100, Krister Walfridsson wrote:
  | On Mon, 25 Nov 2002, Luke Mewburn wrote:
  | 
  | > Does this change now make the sun3 build generate C++ binaries identical
  | > to the atari/amiga/... group or the luna68k/... group, or is the sun3 still
  | > "different"?
  | 
  | Actually, the grouping is due to a compiler bug, and is not static (i.e.
  | you can change the groups by adding/removing some lines of code in a way
  | that interact with whitespace differences in the header files...)

Wow!  Good diagnosis.


  | I'm working on the compiler bug now...

Excellent.

Is this an m68k specific bug, or would it also help explain the
similar problem seen on mipsel and mipseb (PRs 18316 and 18317)?

Luke.