https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67883
--- Comment #9 from Paul Thomas <pault at gcc dot gnu.org> --- (In reply to Gerhard Steinmetz from comment #2) > Another group of examples. > First case is working in a sufficient manner. > Concatenating two empty hulls (zero len and size, respectivly) > gives an empty hull as result again. > > > $ cat zz1.f90 > program p > character(*), parameter :: c1(*) = [character(*) :: 'a'] // [character(*) > :: 'z'] > character(*), parameter :: c2(*) = [character(0) ::] // [character(0) ::] > character(*), parameter :: c3(*) = [character(0) ::] // [character(*) ::] > character(*), parameter :: c4(*) = [character(0) ::] > print *, 'c1: ', len(c1), size(c1), c1 > print *, 'c2: ', len(c2), size(c2), c2 > print *, 'c3: ', len(c3), size(c3), c3 > print *, 'c4: ', len(c4), size(c4), c4 > end > > > $ gfortran-6 -g -O0 -Wall -fcheck=all zz1.f90 > $ a.out > c1: 2 1 az > c2: 0 0 > c3: 0 0 > c4: 0 0 > > --- > > > Next case is questionable : no error, unexpected results. > > > $ cat zz2.f90 > program p > character(*), parameter :: c5(*) = [character(*) :: 'a'] // [character(0) > ::] > character(*), parameter :: c6(*) = [character(0) ::] // [character(*) :: > 'z'] > print *, 'c5: ', len(c5), size(c5), c5 > print *, 'c6: ', len(c6), size(c6), c6 > end > > > $ gfortran-6 -g -O0 -Wall -fcheck=all zz2.f90 > $ a.out > c5: 0 0 > c6: 0 0 > > > > Errors detected with alternative code : > > $ cat zz3.f90 > program p > character(*), parameter :: b1(*) = [character(*) :: 'a'] > character(*), parameter :: b2(*) = [character(0) ::] > character(*), parameter :: b3(*) = b1 // b2 > ! character(*), parameter :: b4(*) = b2 // b1 > print *, 'b1: ', len(b1), size(b1), b2 > print *, 'b2: ', len(b2), size(b2), b2 > print *, 'b3: ', len(b3), size(b3), b3 > ! print *, 'b4: ', len(b4), size(b4), b4 > end > > > $ gfortran-6 -g -O0 -Wall -fcheck=all zz3.f90 > zz3.f90:4:37-43: > > character(*), parameter :: b3(*) = b1 // b2 > 1 2 > Error: Shapes for operands at (1) and (2) are not conformable > zz3.f90:8:24: > > print *, 'b3: ', len(b3), size(b3), b3 > 1 > Error: 'string' argument of 'len' intrinsic at (1) must be CHARACTER > zz3.f90:8:34: > > print *, 'b3: ', len(b3), size(b3), b3 > 1 > Error: 'array' argument of 'size' intrinsic at (1) must be an array All these yield the error: "Type-spec at (1) cannot contain an asterisk for a type parameter" on trunk. Again, I will check previous branches in a week or so. Paul
