Weekly Bootstrapping Report: 46% Failures

[James Dennett]

Gabriel Dos Reis wrote:
> 
> Geoff Keating <geoffk@geoffk.org> writes:
> 
> | Jeffrey Oldham <oldham@codesourcery.com> writes:
> |
> | > GCC 3.1     i686-pc-linux-gnu       mips-sgi-irix6.5
> | > Apr02       failure                 failure
> | > Apr03       failure                 failure
> | > Apr04                               failure
> | > Apr05                               failure
> | > Apr06       failure                 failure
> | > Apr07       failure                 failure
> | > Apr08       failure                 failure
> | > ------------------------------------------------
> | >     43% failure             50% failure
> |
> | That's odd.  The regression tester (which runs on the mainline on
> | i686-linux) reports bootstrap success since Apr 3.
> 
> Yes that is weird because I'm not the only one seeing a bootstrap
> failure on mainline for about a week.

I've been unable to bootstrap the 3.0 branch for a similar
length of time on Linux (Mandrake 7.1 + upgrades) using an
older 3.0 for the bootstrap compiler.  Looks like problems
in the fields of config/makefile stuff.  I might try a
clean checkout before diving in to try to work out what
has been broken, or whether I need a binutils update.

-- James Dennett