analysis: alpha bootstrap failure on mainline

[David Edelsohn]

>>>>> Roger Sayle writes:

Roger> Digging deeper it turns out that the bug isn't
Roger> actually in the preprocessor code at all, but somewhere else in the
Roger> compiler resulting in cppexp.o being miscompiled during bootstrap.
Roger> Removing this file and rebuilding with the host compiler, allows
Roger> stage3 to complete.

Roger> Given that working like this, its going to take me a while, I was
Roger> wondering whether gdb works on alpha/Linux in which case they'd be
Roger> able to find the culprit much faster, or alternatively could someone
Roger> with an alpha start the binary search to identify the patch that
Roger> introduced the failure.

	Have you thought about using Janis's tools to automatically narrow
in on the patch which introduced the failure?

David