Re: fix bootstrap comparison failures

[Tim Josling <tej at melbpc dot org dot au>]

> Well, some of them anyway.
> 
> When starting from gcc 2.96 (or some other non-recent or non-gcc), we
> don't have a C99 compliant boolean.  What we've got is regular integral
> type typedefed to bool.  Which means that any time we rely on the
> properties of a C99 boolean, we're wrong.
> 
> Alex Oliva suggested enhancing -Wtraditional to catch this kind of thing.
> I'm inclined to agree, since this sort of error will become more common
> over time.  I have no intention of actually implementing this at the
> moment though.
> 
> 
> r~
> 
> 

There is a similar problem I get sometimes (4-5 times so far) when doing lots
of changes and using make quickstrap, on i686-pc-gnu-linux.

Sometimes I get a problem where if there are any warnings the cc1 terminates
with non-zero return code. I tracked this down to different effective
definitions of the bool data type betwen toplev.c and diagnostic.c, but I
could dnever work out what is going on. A re-bootstrap fixes it.

Tim Josling