------- Comment #2 from anlauf at gmx dot de 2008-01-16 22:41 ------- (In reply to comment #1) > I don't think you can disable these types. If your target does not support > them, something is wrong.
The target (if you refer to it as the underlying processor) does support these types; at least they are supported in C but not in Fortran. > How is char defined for your target? sizeof(char) needs to be 1 to be > complaint to C. C is most likely fine, but I don't care too much about C. Fortran on that platform (i.e. (sx)f90) does not support INTEGER(kind=1). In fact, it does not need to. If you look into the F2003 draft standard: 15.1.1 Named constants and derived types in the module [...] The values of C_SHORT, C_LONG, C_LONG_LONG, C_SIGNED_CHAR, C_SIZE_T, C_INT8_T, C_INT16_T, C_INT32_T, C_INT64_T, C_INT_LEAST8_T, C_INT_LEAST16_T, C_INT_LEAST32_T, C_INT_LEAST64_T, C_INT_FAST8_T, C_INT_FAST16_T, C_INT_FAST32_T, C_INT_FAST64_T, C_INTMAX_T, and C_INTPTR_T shall each be a valid value for an integer kind type parameter on the processor or shall be -1 if the companion C processor defines the corresponding C type and there is no interoperating Fortran processor kind or -2 if the C processor does not define the corresponding C type. I checked some of these parameters (using sxf90 -f2003 -dw -dW): C_SIGNED_CHAR = -1 C_INT8_T = -1 C_INT_LEAST8_T = -2 C_SHORT = 2 C_INT = 4 C_LONG = 8 C_LONG_LONG = 8 I am a bit concerned about the value of C_INT_LEAST8_T, but the others appear ok. NEC's f90 is somewhat broken, although in other respect. Printing selected_int_kind (i) for some values of i, I get: i selected_int_kind (i) 2 2 4 2 6 4 8 4 10 8 12 8 14 8 16 -1 18 -1 I am told that the last two lines are a known bug. But by reading the standard I do not see a reason why there should be a 1-byte integer type. (Don't get me wrong: it would save me a lot of trouble if NEC would support integer*1 ...) -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34804
