This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: ICE in gcc 2.95.2 [mingw]
- To: "Priem, Gregory L" <gregory dot l dot priem at intel dot com>
- Subject: Re: ICE in gcc 2.95.2 [mingw]
- From: Mumit Khan <khan at NanoTech dot Wisc dot EDU>
- Date: Mon, 10 Jan 2000 18:39:59 -0600 (CST)
- cc: "'gcc-bugs at gcc dot gnu dot org'" <gcc-bugs at gcc dot gnu dot org>
On Mon, 10 Jan 2000, Priem, Gregory L wrote:
> I am trying to use COM using gcc 2.95.2 under mingw [packaged by Mumit
> Khan], and am getting an ICE.
>
> [I hope I RTFM'd enough to not waste anyone's time!]
Yes, you did. The kind of bug reports that doesn't elicit the usual `yes,
please do send ...' and `yes, we do need ...' responses from Alexandre ;-)
[ ... ]
> I assume that the ICE is somehow due to the COM class not being listed first
> in the inheritance chain. Looking through the archives, I saw
> http://gcc.gnu.org/ml/gcc-patches/1999-06/msg00274.html which I thought
> would give me an error message if the COM interface were not the left-most,
> so perhaps this is not the issue. [Or maybe this made it into the mainline,
> but not the 2.95 branch?]
Hmmm ... as far as I can see, the C++ front-end should issue an error
when it sees this code since the leftmost bases must be all COM. Wonder
if I goofed when I added the change that allows for *multiple* COM base
classes, as long as all of those are leftmost. I'll double check with
my local tree and the mainline one.
Thanks for the well researched bug report.
Regards,
Mumit