This is the mail archive of the
gcc-prs@gcc.gnu.org
mailing list for the GCC project.
c++/8762: Internal compiler error.
- From: bothie at t-online dot de
- To: gcc-gnats at gcc dot gnu dot org
- Date: 30 Nov 2002 01:01:02 -0000
- Subject: c++/8762: Internal compiler error.
- Reply-to: bothie at t-online dot de
>Number: 8762
>Category: c++
>Synopsis: Internal compiler error.
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: unassigned
>State: open
>Class: rejects-legal
>Submitter-Id: net
>Arrival-Date: Fri Nov 29 17:06:01 PST 2002
>Closed-Date:
>Last-Modified:
>Originator: bothie@t-online.de
>Release: gcc version 2.95.3 20010315 (SuSE)
>Organization:
>Environment:
(i386 architecture:) AMD Athlon XP 1700+ (1.466Mhz) / 512MB ram
Linux kernel 2.4.10 (SuSE (de) Linux 7.3 Distribution)
(Note: My kernel is a bit patched, but that should not affect the compiler - it still does a very good job for me without errors all the other time - wanted only to tell ...)
>Description:
/*
$ g++ -v
Reading specs from /usr/lib/gcc-lib/i486-suse-linux/2.95.3/specs
gcc version 2.95.3 20010315 (SuSE)
$ g++ gcc-bug.cc -o gcc-bug
gcc-bug.cc: In function `int any_function()':
gcc-bug.cc:$LINE(buggy=malloc()): Internal compiler error.
gcc-bug.cc:$LINE(buggy=malloc()): Please submit a full bug report.
gcc-bug.cc:$LINE(buggy=malloc()): See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
$ g++ gcc-bug.cc -o gcc-bug.o -mno-built-in
cc1plus: Invalid option `no-built-in'
gcc-bug.cc: In function `int any_function()':
gcc-bug.cc:$LINE(buggy=malloc()): confused by earlier errors, bailing out
-> The latter call (with argument -mno-built-in) was the try to enter
'-fno-builtin' - in the hope the problem was the builtin version of malloc()
but that (-fno-builtin) didn't work (same result as the *first* try).
??? Why cased the argument -mno-built-in, which was not supported (and thusly
hopefully ignored) the compiler to act in another way as without?
My System: (i386 architecture) AMD Athlon XP 1700+ (1.466Mhz) / 512MB ram
Linux kernel 2.4.10 (SuSE (de) Linux 7.3 Distribution)
(Note: My kernel is a bit patched, but that should not
affect the compiler - it still does a very good
job for me without errors all the other time)
*/
class any_class {
public:
friend any_class any_friend();
friend any_class malloc();
};
int any_function() {
any_class clean,buggy;
clean=any_friend(); /* Works well */
buggy=malloc(); /* Causes the compiler to fail! */
/* Whatever comes here seems to doesn't affect anything */
>How-To-Repeat:
>Fix:
Please drop me a mail, as soon as you found and corrected the bug (I think it will be added to the official distribution then) so I can download the newest version of gcc (without this bug). I *really* need the friend called malloc!
>Release-Note:
>Audit-Trail:
>Unformatted: