Subject: Re: problem compiling egcs cpp
To: Perry E. Metzger <>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: current-users
Date: 09/30/1998 15:53:13
I hit this too, trying to bootstrap egcs for a mips.  I think its
going to be a generic problem for bootstrapping, so we should have a
hook to fix it.

How about we put Johnny's patch to turn off -Werror in egcs/cpp, but
only if USE_EGCS is not set, viz:

    # avoid spurious warning in cexp.y
    .ifndef USE_EGCS
    CWARNFLAGS+=	-Wno-error

works for me with gcc  Anyone object?

>Strangely enough, egcs doesn't make it die. Bizarre.

It looks awfully similar to a very longstanding warning "bug" that
binutils and/or gdb used to generate for mips targets.  The
flow-optimizer didnt work out that a shift would actually always be
legitimate, so it it complained.  

My guess is that egcs wins by doing smarter optimztaion, or maybe
just by doing optimizations in a different ordera