c++/8183: ICE when compiling classes with virtual base class on sparc64-sun-solaris2.8
Robert Schiele
rschiele@uni-mannheim.de
Thu Jan 9 13:06:00 GMT 2003
The following reply was made to PR c++/8183; it has been noted by GNATS.
>From rschiele@uni-mannheim.de Mon Jan 13 00:43:38 2003
From: Robert Schiele <rschiele@uni-mannheim.de>
To: Christian =?iso-8859-1?Q?J=F6nsson?= <c.christian.joensson@telia.com>
Cc: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, gcc-bugs@gcc.gnu.org,
ghazi@caip.rutgers.edu
Subject: Re: c++/8183: ICE when compiling classes with virtual base class on sparc64-sun-solaris2.8
Date: Thu, 9 Jan 2003 14:02:33 +0100
--PEIAKu/WMn1b1Hv9
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
On Thu, Jan 09, 2003 at 01:12:24PM +0100, Christian J=F6nsson wrote:
> however, if I just build the compiler, i.e., not doing a bootstrap, I do=
=20
> not get the ICE.
Well, this is perfectly clear, because if you just built the compiler,
you didn't compile anything with such a class construction, as the
compiler is written in C, not C++.
But I would expect the such built compiler to also fail building my
test case, doesn't it?
> So, I wonder what the analysis is?
I can't say very much about that, but it seems so that there is
something broken in the C++ frontend on 64 bit systems. I assume some
false assumptions about size of specific data types that are only
correct on some platforms.
I don't have tried on other 64 bit platforms, but for Solaris I can
say that I am able to build both 32 and 64 bit versions of the
testcase successfully with the compiler built itself in 32 bit mode.
But with a compiler built in 64 bit mode, the bug is triggered
independent of the target being 32 or 64 bit.
Robert
--=20
Robert Schiele Tel.: +49-621-181-2517
Dipl.-Wirtsch.informatiker mailto:rschiele@uni-mannheim.de
--PEIAKu/WMn1b1Hv9
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)
iQEVAwUBPh1y6cQAnns5HcHpAQFlSwf/Rs1ZzsM3hEpD/dyG8ynNW5XLYLyG97z3
Yt5PQGTWGye0sBI/Q1o0dOX852mZsUer4tSoUG36ful48tYzgO1lbZi4a56CqOmy
mdV59pcLkf0IRHohpi1vMEcaUh20/oXuGJEbmHxw4qYLQGPiEHHmf2TdqxBTeIAZ
ds9oHRDduJZfvH5Mkog12RnNSbvdF+TR5sGgVU3pT5BiA34zy8ndrbu+561SKUlf
p0Rm05pgu+E0NgEx1akn90z2klzxNtrBTMD/B11AL+1nI0SFe4Fp7xrO95RFPULX
9um0TK22/vrwYmeqtHTh/emLjERypVz2wtM2g52YLA7AoNyn8BzgTQ==
=HRsk
-----END PGP SIGNATURE-----
--PEIAKu/WMn1b1Hv9--
More information about the Gcc-prs
mailing list