This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [rfc] some intrinsics changes
On Mon, Nov 24, 2003 at 10:31:11PM +0000, Paul Brook wrote:
> _gfortran_foo_r4 is the form that should be used, although we sometimes drop
> the "r" if it's implicit in foo.
Personally, I think consistency is worth a lot. Dropping the one letter
doesn't seem to be worth it.
> Anything that isn't prefixed by _gfortran_ should be.
00008f00 T __all_l4
00009400 T __any_l4
00009900 T __count_4_l4
0002fe40 T __cshift0_4
00020a50 T __dot_product_c4
00030270 T __eoshift0_4
00021990 T __matmul_c4
0000a480 T __maxloc0_4_i4
0000d6a0 T __maxloc1_4_i4
00011120 T __maxval_i4
00012ce0 T __minloc0_4_i4
00015f00 T __minloc1_4_i4
00019980 T __minval_i4
0000a6d0 T __mmaxloc0_4_i4
0000d940 T __mmaxloc1_4_i4
000113a0 T __mmaxval_i4
00012f30 T __mminloc0_4_i4
000161a0 T __mminloc1_4_i4
00019c00 T __mminval_i4
0001d440 T __mproduct_c4
0001fc50 T __msum_c4
000309a0 T __pack
0001d190 T __product_c4
00031b90 T __reshape
00022840 T __shape_4
00030ce0 T __spread
0001f9c0 T __sum_c4
00032350 T __transpose
000327b0 T __unpack0
00032490 T __unpack1
> My current plan is to (eventually) add an extension for this,
> or maybe an attribute or compiler option that prefixes symbols with _.
I would suggest an attribute that allows an arbitrary string,
much like the C/C++ FE does with
int foo (void) __asm__("bar");
r~