[Bug c++/11376] [3.3/3.4 regression] mozilla-1.4 miscompiled

["kevin dot hendricks at sympatico dot ca" <gcc-bugzilla at gcc dot gnu dot org>]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11376



------- Additional Comments From kevin dot hendricks at sympatico dot ca  2003-07-08 18:29 -------
Subject: Re:  [3.3/3.4 regression] mozilla-1.4 miscompiled

Hi Franz,

Please try changing inlinedGetFrameAt() in your imgContainerGIF.ii

To look like the following (a rough hack job)

   inline nsresult inlinedGetFrameAt(PRUint32 index, gfxIImageFrame 
**_retval) {
     nsISupports ** rp = (nsISupports**) _retval;
     nsISupports *_elem = mFrames.ElementAt(index);
     if (!_elem) return ((nsresult) 0x80004005L);
     *rp = _elem;
     return 0;
}

This casts back the gfxIImageFrame** _retval to what is really is which 
is a nsISupports**

I tried this with your test case under gcc 3.3 and get the following:

[kbhend@base1 huh]$ g++ -O2 -fno-exceptions -o junk imgContainerGIF.ii 
-L/src1/mozilla-1.3/obj-ppc/dist/bin -lxpcom -lgkgfx
[kbhend@base1 huh]$ ./junk
rv 0
done
[kbhend@base1 huh]$ gcc -dumpversion
3.3

So it fixes the problem (no segfault!)

This is not the right fix.  The nsCOMPtr<> template type and the
many of the methods should be passing around nsISupports** instead
of gfxIImageFrame** in many places.

But this supports Mark's claim that the code is not aliasing safe.

Kevin


On Tuesday, July 8, 2003, at 01:50 PM, sirl at gcc dot gnu dot org 
wrote:

> PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* 
> gcc-bugs@gcc.gnu.org.
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11376
>
>
>
> ------- Additional Comments From sirl at gcc dot gnu dot org  
> 2003-07-08 17:50 -------
> Well, since the generated assembly looks now quite different than for 
> the full
> testcase, I would just like to hear another opinion whether the full 
> testcase
> was correctly minimized. If yes, I agree that this bug shoud be closed 
> and one
> against mozilla should be opened.
> Hmm, Kevin, could you create a "fixed" (like you just did for the 
> simplified
> one) full testcase (diff?)? If that fixes the full testcase I'll be 
> happy :-).
>
>
>
>
>
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug, or are watching someone who is.
>
>
----
Kevin B. Hendricks
Associate Professor of Operations and Information Technology
Richard Ivey School of Business, University of Western Ontario
London Ontario, CANADA  N6A 3K7
khendricks@ivey.uwo.ca