On Fri, Apr 23, 2010 at 08:22:40PM -0500, lylelyle wrote: > > No, I can't do this in gdb. > > (gdb) print a(4) > Wrong number of subscripts > > (gdb) print a(6,1) > no such vector element > > lyle >
I'm, quite sorry. As gdb relies on language features, you might either examine the memory directly (calculating the address by hand): (showing the second step): (gdb) x/5f &a 0x7fffffffd360: 1 0 -9.05608908e+33 4.59163468e-41 0x7fffffffd370: -4.65473378e+33 (gdb) x/5f 0x7fffffffd374 0x7fffffffd374: 2 0 0 -nan(0x7fd4c8) 0x7fffffffd384: 4.59163468e-41 or more convenient switch the language to C: be carefull: the colummn numbers are calculated from 1 up, as the first element a[0][0] is skipped internally. I once knew why, but I can't remember why. (gdb) set language c (gdb) p a $32 = {{1, 4, -9.05608908e+33, 4.59163468e-41, -4.65473378e+33}, {2, 5, 0, -nan(0x7fd4c8), 4.59163468e-41}, {3, 6, 1.40129846e-45, 0, -4.65251539e+33}} (gdb) p a[2][2] $33 = 6 of course you can always get back to fortran again: (gdb) set language fortran (gdb) p a $36 = (( 1, 4, -9.05608908e+33, 4.59163468e-41, -4.65473378e+33) ( 2, 5, 0, -nan(0x7fd4c8), 4.59163468e-41) ( 3, 6, 1.40129846e-45, 0, -4.65251539e+33) ) bye ingo _______________________________________________ bug-gdb mailing list bug-gdb@gnu.org http://lists.gnu.org/mailman/listinfo/bug-gdb