Re: [C++ patch] one more missed mark_used

[Andreas Jaeger <aj at suse dot de>]

Mark Mitchell <mark@codesourcery.com> writes:

> On Sat, 2003-08-02 at 13:54, Jason Merrill wrote:
>> On 02 Aug 2003 13:24:05 -0700, Mark Mitchell <mark@codesourcery.com> wrote:
>> 
>> > This code:
>> >
>> > +   if (type)
>> > +     {
>> > +       tree tinfo = get_tinfo_decl (type);
>> > +       mark_used (tinfo);
>> > +     }
>> >
>> > will penalize not-function-at-a-time mode compilation because it will
>> > result in generating additional typeinfos, even when the back end
>> > optimizes away the handlers.
>> 
>> I had the same concern, but then thought it was OK because DECL_NEEDED_P
>> won't be true unless we actually refer to the symobl.  Do you disagree?
>
> No, you're correct.  I hadn't taken that into account.  We'll still call
> assemble_external for each tinfo, but I guess that's reasonably
> harmless?  (And clearly the real fix for that has nothing to with Jan's
> stuff; that happens any time we call mark used.)
>
> I withdraw the objection.

Just for this part?  So, that means the patch is not approved since
you had also comments on another part.

Andreas
-- 
 Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj
  SuSE Linux AG, Deutschherrnstr. 15-19, 90429 Nürnberg, Germany
   GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126

Attachment: pgp00000.pgp
Description: PGP signature