gfortran (snapshot as of 12/17/04) compiled code gives a runtime error when reading formatted complex*16 values using the following program:
program testIO implicit none integer i complex*16 values(5) open(7,file='test1.dat',status='old') read ( 7, '(5E15.8)' ) ( values (i), i = 1, 5 ) print *, ( values (i), i = 1, 5 ) end The text file "test1.dat" contains: 2.00000000E+00 0.00000000E+00 2.52505826E-01 0.00000000E+00 1.44380768E-04 -1.11464849E-18 1.01002561E-01 0.00000000E+00-2.52507036E-02 0.00000000E+00 The code was compiled and executed with "gfortran bug.f; a.out", and produced the following error message: At line 8 of file bug2.f Fortran runtime error: End of file The code works correctly when compiled with g77 and ifort (Intel Fortran), and produces the output (g77): (2.,0.) (0.252505826,0.) (0.000144380768,-1.11464849E-18) (0.101002561,0.) (-0.0252507036,0.) The code and data was extracted from the Harwell-Boeing collection of sparse matrices, which is a commonly used format for sparse matrix computations. See http://math.nist.gov/MatrixMarket/collections/hb.html. Additional information: "gfortran -v" gives: Reading specs from /auto/home0/rmunk/temp_gfortran/irun/bin/../lib/gcc/i686-pc-linux-gnu/4.0.0/specs Configured with: ../gcc/configure --enable-languages=c,f95 --prefix=/usr/work/2004352/irun Thread model: posix gcc version 4.0.0 20041217 (experimental) Best regards, Rasmus Munk Larsen, Stanford University -- Summary: runtime error when reading complex*16 using formatted I/O Product: gcc Version: 4.0.0 Status: UNCONFIRMED Severity: normal Priority: P2 Component: libfortran AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: rmunk at quake dot stanford dot edu CC: gcc-bugs at gcc dot gnu dot org http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19114