missing(?) intrinsics for real(10)

[Dominique Dhumieres]

> There isn't an INTEGER(10), so _gfortran_transpose_i10 doesn't
> make sense. ...

May be! But nevertheless I get:

[scala] lin/source> gfc -O3 induct_10.f90
/tmp/ccFhMjK3.o(.text+0x129f9): In function `__m_quad_m__quad_mutual_inductance':
: undefined reference to `_gfortran_transpose_i10'
/tmp/ccFhMjK3.o(.text+0x12c4b): In function `__m_quad_m__quad_mutual_inductance':
: undefined reference to `_gfortran_transpose_i10'
/tmp/ccFhMjK3.o(.text+0x134be): In function `__m_quad_m__quad_mutual_inductance':
: undefined reference to `_gfortran_transpose_i10'
/tmp/ccFhMjK3.o(.text+0x13710): In function `__m_quad_m__quad_mutual_inductance':
: undefined reference to `_gfortran_transpose_i10'
/tmp/ccFhMjK3.o(.text+0x13eb2): In function `__m_quad_m__quad_mutual_inductance':
: undefined reference to `_gfortran_transpose_i10'
/tmp/ccFhMjK3.o(.text+0x147f0): more undefined references to `_gfortran_transpose_i10' follow
collect2: ld returned 1 exit status

where induct_10.f90 is induct.f90 I have replaced all the occurences of

(1)

integer, parameter, private :: longreal = selected_real_kind(15,90)

by

integer, parameter, private :: longreal = selected_real_kind(18,90)

(2)

      integer, parameter :: LONGreal = selected_real_kind(15,90)

by

      integer, parameter :: LONGreal = selected_real_kind(18,90)

(3)

      real (kind = longreal), parameter :: pi = 3.141592653589793_longreal

by

      real (kind = longreal), parameter :: pi = 3.141592653589793238463_longreal

and (4)

real (kind = longreal), parameter, private :: pi = 3.141592653589793_longreal

by

real (kind = longreal), parameter, private :: pi = 3.141592653589793238463_longreal

> IMHO, -fdefault-{integer,real}-* are a HUGE mistake.

I don't agree; I think it's very convenient for testing the
effects of the different kind's even if some care must be
taken to avoid side effects.

Cheers

Dominique