Re: Assume solib.h

[Mark Kettenis <kettenis@gnu.org>]

   Date: Tue, 16 Nov 2004 11:13:54 -0500
   From: Andrew Cagney <cagney@gnu.org>

   >>and how my change breaks it?
   > 
   > 
   > This part I don't know the details about.  Mark said that it does
   > break, and I spoke on the assumption that you agree with the fact of
   > breakage,

   Unfortunatly your assumption is wrong.  As I stated to Mark, and 
   contrary to his assertion, powerpc-elf passes this sniff test:

   $ gcc -g -static src/gdb/testsuite/gdb.base/break{,1}.c
   cagney@to-dhcp51$ ./X-powerpc-elf/gdb/gdb ./a.out
   warning: A handler for the OS ABI "GNU/Linux" is not built into this 
   configuration
   of GDB.  Attempting to continue with the default powerpc:common settings.
   (gdb) target sim
   Connected to the simulator.
   (gdb) load
   (gdb) run
   Starting program: /home/scratch/PENDING/YYYY-MM-DD-solib/a.out
   do_call() unimplemented call settimeofday
   (gdb) disassemble
   No frame selected.
   <oops>
   (gdb) x/i $pc
   0x10008a10 <uname+4>:   sc

Yup, I'm wrong.  Not *every* embedded target will break.  Some of them
include solib-svr4.o or some other solib-xxx.o; powerpc-elf is one of
them.  I'm also wrong that it breaks vax-dec-openbsd* for pretty much
the same reason.  However, there are plenty of embedded targets for
which I'm pretty certain that my analysis is true: arm-elf, mips-elf,
i386-elf are among them.

Mark