This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Virtual table contents
- To: Mike Stump <mrs at windriver dot com>
- Subject: Re: Virtual table contents
- From: Jason Merrill <jason at redhat dot com>
- Date: 19 Dec 2000 15:11:39 +0000
- Cc: divx at euro dot ru, gcc at gcc dot gnu dot org
- References: <200012182210.OAA19103@kankakee.wrs.com>
>>>>> "Mike" == Mike Stump <mrs@windriver.com> writes:
>> Date: Mon, 18 Dec 2000 14:23:55 -0600
>> To: gcc@gcc.gnu.org
>> From: Eugene Kuznetsov<divx@euro.ru>
>> Is there an universal way ( i.e. the way that would be acceptable
>> for all compilers since egcs 2.91 and higher ) to tell compiler NOT
>> to use these two positions and to start function pointers
>> immediately from 1st position? Or such a way only for some specific
>> versions of compilers?
> No.
Well, there's __attribute__ ((com_interface)). And the new ABI will always
start from the 1st position (the RTTI stuff will be at negative offsets).
Jason