https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116148
--- Comment #10 from Kewen Lin <linkw at gcc dot gnu.org> --- (In reply to John David Anglin from comment #9) > These two are reversed: > Breakpoint 2 at 0x105a8: file > /home/dave/gnu/gcc/gcc/gcc/testsuite/c-c++-common/fam-in-union-alone-in- > struct-2.c, line 49. > (gdb) p/x with_fam_3_v.a[0] > $13 = 0x1f > (gdb) p/x with_fam_3_v.a[7] > $14 = 0x8f Thanks for the information, trying with a cross build with target hppa-linux, it defines #define __ORDER_LITTLE_ENDIAN__ 1234 #define __FLOAT_WORD_ORDER__ __ORDER_BIG_ENDIAN__ #define __ORDER_PDP_ENDIAN__ 3412 #define __BYTE_ORDER__ __ORDER_BIG_ENDIAN__ #define __ORDER_BIG_ENDIAN__ 4321 , so it's BE and would adopt: +#define WITH_FAM_3_V_A0 0x1f +#define WITH_FAM_3_V_A7 0x8f for with_fam_3_v.a[0] and with_fam_3_v.a[7], which looks matching the above gdb outputs, did I miss something? Or the above output isn't from a hppa-linux debugging (the actual output there are reversed)?
