This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
9 bcopys (5 top dir, 4 cp dir), please help identify memmoves
- To: gcc-patches at gcc dot gnu dot org, gcc at gcc dot gnu dot org
- Subject: 9 bcopys (5 top dir, 4 cp dir), please help identify memmoves
- From: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- Date: Sun, 11 Mar 2001 12:41:11 -0500 (EST)
There are only a handful of bcopy calls left, I'd like to eliminate
them and poison the token. However I'm not entirely sure by reading
the code which of the remaining ones need to be memmove vs memcpy. In
fact I'm pretty sure at least some are memmoves through some across
the board bootstrap testing of using only memcpy.
I'd rather not guess or get it wrong simply because a bootstrap fails
to indicate a latent problem. Also I'd rather not weasel out by
making them all memmove.
Can anyone provide insight into these?
Thanks,
--Kaveh
./collect2.c:3279: bcopy ((char *) load_hdr, (char *) (obj + offset), size);
./combine.c:2031: bcopy ((char *) &old->elem[0], (char *) XVEC (newpat, 0)->elem,
./genattrtab.c:2013: bcopy ((char *) unit_ops[unit->num], (char *) &op_array[i],
./sbitmap.c:99: bcopy ((PTR) src->elms, (PTR) dst->elms,
./toplev.c:1668: bcopy ((char *) handler, (char *) float_handler, sizeof (float_handler));
cp/class.c:1220: bcopy ((PTR) &TREE_VEC_ELT (method_vec, 0),
cp/class.c:1259: bcopy ((PTR) &TREE_VEC_ELT (method_vec, slot),
cp/decl.c:3694: bcopy ((char *) newdecl + sizeof (struct tree_common),
cp/decl.c:3732: bcopy ((char *) newdecl + sizeof (struct tree_common),
--
Kaveh R. Ghazi Engagement Manager / Project Services
ghazi@caip.rutgers.edu Qwest Internet Solutions