Re: Patch for max number of errors make gfortran.dg/gomp/reduction1.f90 fail

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Thu, Mar 08, 2007 at 11:58:11AM -0800, Brooks Moses wrote:
> Brooks Moses wrote:
> >Fran?ois-Xavier Coudert wrote:
> >>The subject says it all...
> >
> >I don't see how that can be the -fmax-error patch, though.  When I run 
> >the testcase manually, the error I get is a linker error:
> >
> > > ~/bin-4.2/bin/gfortran reduction1.f90
> >/home/brooks/bin-4.2/lib/gcc/i686-pc-linux-gnu/4.2.0/libgfortranbegin.a(fmain.o)(.text+0x23): 
> >In function `main':
> >../../../svn-source-4.2/libgfortran/fmain.c:18: undefined reference to 
> >`MAIN__'
> >collect2: ld returned 1 exit status
> 
> Sorry, brainglitch -- the same one I had when I thought this was 
> unrelated to my patch when I first regtested it.  Supplying -fopenmp on 
> the compilation line shows the problem, and the linker error is because 
> this is only a dg-compile testcase.
> 
> I'll submit a patch (to add -fmax-error=50 to the compilation options) 
> as soon as it regtests.
> 
> Mea culpa.  Thanks for catching that!
> 

Mainline has

gfortran.dg/gomp/reduction1.f90

! { dg-do compile }
! { dg-options "-fopenmp -fmax-errors=100" }
! { dg-require-effective-target tls }

while 4.2 has

/gfortran.dg/gomp/reduction1.f90

! { dg-do compile }
! { dg-require-effective-target tls }



-- 
Steve