Re: advance warning

[Alan Modra <amodra at bigpond dot net dot au>]

On Wed, Sep 04, 2002 at 11:25:51PM +0930, Alan Modra wrote:
> I intend to apply the gas reloc patch tomorrow morning, ie. probably

OK, it's in.  Differences from the last posted patch are:

- Made subsegs.c change the proper way, ie. using symbol_get_bfdsym
rather than including struc-symbol.h and fiddling with structure
fields.

- Some tc-ppc.c changes went in prior to this commit.

- config/obj-coff.h gets this extra define

/* We can't use the predefined section symbols in bfd/section.c, as
   COFF symbols have extra fields.  See bfd/libcoff.h:coff_symbol_type.  */
#ifndef obj_sec_sym_ok_for_reloc
#define obj_sec_sym_ok_for_reloc(SEC)	((SEC)->owner != 0)
#endif

- config/obj-elf.h comment for the above revised to say
/* BFD wants to write the udata field, which is a no-no for the
   predefined section symbols in bfd/section.c.  They are read-only.  */

- mips testsuite fixed to not use a global label for 16 bit branch
  targets.

- I snuck in an xfail for rs6000-*-* on the macro strings test.

I don't intend to attack bfd_install_relocation for quite a while, as
I think this one needs to settle in first.

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre