fortran testcase fix
Dave Love
d.love@dl.ac.uk
Fri Sep 11 10:55:00 GMT 1998
I checked in this to fix failures in this new test. I'd lost the
idate one somehow from the version I'd tested before and the real
number comparision shows up randomly. I don't think it merited
approval and there are notwork problos here even before tonight's
predicted Death Of The Net, so I took the opportunity of a period when
I could get through.
1998-09-11 Dave Love <d.love@dl.ac.uk>
* g77.f-torture/execute/u77-test.f: Fix bad consistency checks.
Index: u77-test.f
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/testsuite/g77.f-torture/execute/u77-test.f,v
retrieving revision 1.1
retrieving revision 1.2
diff -p -c -r1.1 -r1.2
*** u77-test.f 1998/09/08 14:58:01 1.1
--- u77-test.f 1998/09/11 04:57:47 1.2
***************
*** 80,90 ****
c consistency-check etime vs. dtime for first call
r1 = etime (tarray1)
- if (r1.ne.tarray1(1)+tarray1(2)) then
- write (6,*) '*** ETIME didn''t return sum of the array: ',
- + r1, ' /= ', tarray1(1), '+', tarray1(2)
- call abort
- end if
r2 = dtime (tarray2)
if (abs (r1-r2).gt.1.0) then
write (6,*)
--- 80,85 ----
*************** c now try to get times to change enough
*** 141,148 ****
call idate (idat)
write (6,*) 'IDATE d,m,y: ',idat
print *, '... and the VXT version: ', i,j,k
! if (i/=idat(2) .or. j/=idat(1) .or. k/=idat(3)) then
! print *, '*** vxy and u77 versions don''t agree'
call abort
end if
call time(line(:8))
--- 136,143 ----
call idate (idat)
write (6,*) 'IDATE d,m,y: ',idat
print *, '... and the VXT version: ', i,j,k
! if (i/=idat(2) .or. j/=idat(1) .or. k/=mod(idat(3),100)) then
! print *, '*** vxt and u77 versions don''t agree'
call abort
end if
call time(line(:8))
More information about the Gcc-patches
mailing list