[Patch] Fix bogus 'function does return' warning

Richard Guenther richard.guenther@gmail.com
Sat Oct 17 17:10:00 GMT 2009


On Sat, Oct 17, 2009 at 6:51 PM, Richard Henderson <rth@redhat.com> wrote:
> On 10/17/2009 04:46 AM, Richard Guenther wrote:
>>
>> As EH lowering complicates matters in your case (otherwise cfgcleanup
>> would fixup all issues) I think a proper place to address this is EH
>> lowering
>> (thus I cced rth).
>
> What would you suggest that EH lowering do?  I suppose we could always copy
> the eh path for finally.  It would eliminate the problematic case here.
>
> Really, I'd have preferred this be cleaned up in the remove_useless_stmts
> pass, which Matz recently removed.  But given a choice between complicating
> pass_lower_cf and pass_lower_eh, frankly I'd prefer pass_lower_cf -- eh
> lowering is already complicated enough.

Heh.  Well then - the original patch is ok.

Thanks,
Richard.



More information about the Gcc-patches mailing list