On Wed, Mar 21, 2012 at 10:23 PM, Iain Sandoe
<develo...@sandoe-acoustics.co.uk> wrote:
>> objc/
>>        * objc-act (objc_build_ivar_assignment): Do not call
>> assemble_external.
>>        (objc_build_global_assignment): Likewise.
>>        (objc_build_strong_cast_assignment): Likewise.
>>        * objc-next-runtime-abi-01.c: Cleanup commented-out
>> assemble_external.
>>        * objc-next-runtime-abi-02.c: Likewise.
>>        * objc-gnu-runtime-abi-01.c: Likewise.
>> <cleanup_objc_assemble_external.diff>
> ... this would allow us to close PR17982?

I believe so, but I have to look into it a but deeper to understand
what the remaining assemble_external calls are for. The ones in
config/* are OK, they are all for writing out multiple-inheritance
thunks. The ones that need checking are:

calls.c:1649:     assemble_external (fndecl);
expr.c:1422:      assemble_external (block_move_fn);
expr.c:2794:      assemble_external (block_clear_fn);
expr.c:7423:          assemble_external (exp);
expr.c:9022:      assemble_external (exp);
final.c:2745:             assemble_external (t);
final.c:3497:   assemble_external (t);
final.c:3568:   assemble_external (SYMBOL_REF_DECL (x));
toplev.c:489:      assemble_external (decl);

> ... and make progress on PR24777? (I'm not sure where exactly we need to go
> with this one - we have different sets of calls depending on the runtime)

The FIXMEs in that PR do not exist anymore. Perhaps this removed them:

2011-10-11  Michael Meissner  <meissner at linux dot vnet dot ibm dot com>

        * objc-next-runtime-abi-01.c (objc_build_exc_ptr): Delete old
        interface with two parallel arrays to hold standard builtin
        declarations, and replace it with a function based interface that
        can support creating builtins on the fly in the future.  Change
        all uses, and poison the old names.  Make sure 0 is not a
        legitimate builtin index.
        * objc-next-runtime-abi-02.c (objc_build_exc_ptr): Ditto.
        * objc-gnu-runtime-abi-01.c (objc_build_exc_ptr): Ditto.

In any case, if there's nothing left to fix for PR24777, I suppose it
can be closed as FIXED.


Reply via email to