This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Someone broke bootstrap with gfortran, again!


On Sat, 23 Jul 2005, 15:53:21 +0200, Daniel Berlin wrote:
> On Fri, 2005-07-22 at 17:53 -0700, Steve Kargl wrote:
> > On Fri, Jul 22, 2005 at 05:44:44PM -0700, Jerry DeLisle wrote:
> > > Steve Kargl wrote:
> > > >Does this look familiar to anyone?
> > > >
> > > I was having troubles doing a build after a cvs update.  I had to delete 
> > > everything in the build directory and rerun configure and then it would 
> > > build ok. Not sure its the same problem you are seeing, but it happened 
> > > today.  I am running on i686-pc-linux-gnu.
> > > 
> > 
> > I always remove the contents in the build directory, run configure,
> > then do a "gmake bootstrap".  The only exception to this process
> > is when I'm making small changes to gfortran files where I know
> > a "gmake bubblestrap" will not run into problems.
> > 
> > I suspect the commit that broken gfortran is 
> > 
> > 2005-07-22  Manfred Hollstein  <mh@suse.com>
> > 
> >         * tree-ssa-structalias.c (merge_graph_nodes): Fix uninitialised
> >         warnings.
> >         (int_add_graph_edge): Likewise.
> >         (collapse_nodes): Likewise.
> >         (process_unification_queue): Likewise.
> > 
> > I'll know with certainty in an hour or so.
> 
> I should simply note that the te warnings are invalid.
> We don't *use* the weights uninitialized (this is easily verifiable).

Correct, but unfortunately -Werror errors out at -O3 due to these
warnings... That's why I suggested the fix.

Cheers.

l8er
manfred


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]