This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: 6 GCC regressions, 6 new, with your patch on 2002-01-29T19:24:37Z.
> Date: Tue, 29 Jan 2002 23:38:08 +0000
> From: Neil Booth <neil@daikokuya.demon.co.uk>
> Geoff Keating wrote:-
>
> > No, it's not a debug info issue. Look at the log file at
> > <http://people.redhat.com/geoffk/gcc-regression/native-logsum/test-gdb/gdb.log.gz>; you'll see it's actually a bunch of
> >
> > bitfields.c:83: warning: overflow in implicit constant conversion
> >
> > warnings. I'm not sure whether the warnings are legitimate, although
> > at first glance they may be.
>
> I downloaded a log file, but seem to have gotten the wrong one as I
> couldn't see anything wrong. I was just in the process of trying to
> build GDB locally.
Does your file contain the string 'overflow in implicit constant
conversion'? Mine does.
> I'm fairly sure the warnings are legit; getting the right type for
> bitfields naturally leads to stricter overflow warnings. I'll fix the
> test cases; hopefully in the next 24 hours.
Hmmm. Actually, the first test, bitfields.exp, produces the warnings;
but the other testcases are other kinds of problems. call-ar-st and
call-rt-st seem to be small changes in GDB's output produced by the
different type information; I can't tell if the change is correct.
setvar.exp appears to be a genuine bug introduced by your patch
relating to output of debug information for enumeration bitfields, in
a structure like:
struct senum_field {enum senum field:2; } sef;
it seems to be testing that when you say 'print sef.field', you get
the symbolic name for the value.
--
- Geoffrey Keating <geoffk@geoffk.org> <geoffk@redhat.com>