http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58975
--- Comment #3 from Steve Kargl <sgk at troutmask dot apl.washington.edu> --- On Sun, Nov 03, 2013 at 12:24:10AM +0000, arthur.j.odwyer at gmail dot com wrote: > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58975 > > --- Comment #2 from Arthur O'Dwyer <arthur.j.odwyer at gmail dot com> --- > You're right, adding > > EXTERNAL RAN > > to the top of the program seems to "fix" it. However, for gfortran to break > code that has worked since the 1980s (in a dialect with "1977" in the name, no > less!) seems... gratuitously annoying. > > gfortran doesn't try to maintain f77 compatibility? > Your question does not make sense. F77 is a subset of F90, which is a subset of F95, which is a subset of F2003, which is a subset of F2008. RAN is not an intrinsic subprogram required by any of the Fortran standards. The standard, however, does not forbid a compiler from including additional intrinsic procedures. gfortran elected to include RAN as an intrinsic procedure to be backwards compatiable with g77. I forgot to mention that you could uses -std=f95 to prevent RAN from being considered to be an intrinsic subprogram. I suspect that given the age of your code, that this would lead to other problems.
