Thread View: gnu.g++.bug
1 messages
1 total messages
Started by piggy
Fri, 01 Sep 1989 15:38
Some g++ bugs (sorry, no fixes...)
Author: piggy
Date: Fri, 01 Sep 1989 15:38
Date: Fri, 01 Sep 1989 15:38
220 lines
6211 bytes
6211 bytes
We're running g++ 1.35.1 on Sun 3's with Sun OS 3.5 and on Sun 4's with
Sun OS 4.0.1. The machine gargoyle is a Sun 3 and the machine
tartarus is a Sun 4.
On the Sun 3's tm.h was tm-sun3+.h and md was m68k.md.
On the Sun 4's tm.h was tm-sparc+.h and md was sparc.md.
If you have questions, direct them to keenan@gargoyle.uchicago.edu,
but we would appreciate a cc: to csstaff@gargoyle.uchicago.edu.
-- La Monte H. Yarroll
Preferred: piggy@gargoyle.uchicago.edu
Permanent: piggy@chinet.chi.il.us
Forwarded messages follow -------------------------------------------------
(Message inbox:14)
Received: by gargoyle.uchicago.edu (5.59/1.14)
id AA18281; Fri, 25 Aug 89 09:36:39 198
Date: Fri, 25 Aug 89 09:36:39 198
From: "Tom Barron" <thos>
Message-Id: <8908251436.AA18281@gargoyle.uchicago.edu>
To: piggy
Subject: [keenan@gargoyle.uchicago.edu: g++ crash]
Return-Path: <keenan@gargoyle.uchicago.edu>
Date: Mon, 10 Jul 89 15:06:46 198
From: "Philip Keenan" <keenan@gargoyle.uchicago.edu>
To: bugs@gargoyle.uchicago.edu
Subject: g++ crash
Cc: keenan@gargoyle.uchicago.edu
The following source code, which is OK on other versions of g++, crashes
the compiler, dumping core.
Run with:
g++ -c pde.cc
****************************************
/* pde.cc
test program : why does it crash the compiler when the PDE declared?
Phil Keenan 7-10-89
*/
#include <stream.h>
#include "p.h"
main()
{
cout << "Start";
PDE L;
cout << "End";
}
****************************************
Here is the header file:
//p.h
struct point {
double x ;
double y ;
double z ;
point( ) { x =y =z =0 ;}
point( double x0 , double y0 = 0 , double z0 = 0 ) { x =x0 ;y =y0 ;z =z0 ;}
point( point & p ) { x =p .x ;y =p .y ;z =p .z ;}
void operator =( point & p ) { x =p .x ;y =p .y ;z =p .z ;}
int operator == ( point & p ) ;
}
;
typedef double ( * DF ) ( point & ) ;
typedef double ( * DFV ) ( point & , void * ) ;
class funct {
void * fn ;
int i ;
void * data ;
public :
funct( ) { i =0 ;}
funct( DF Fn ) { i =-1 ;fn =Fn ;data =0 ;}
funct( DFV Fn , void * Data ) { i =1 ;fn =Fn ;data =Data ;}
funct( funct & f ) { fn =f .fn ;i =f .i ;data =f .data ;}
operator =( funct & f ) { fn =f .fn ;i =f .i ;data =f .data ;}
virtual double value ( point & p ) ;
}
;
struct PDE {
int dim ;
funct a [ 3 ] [ 3 ] ;
funct aderiv [ 3 ] [ 3 ] ;
funct b [ 3 ] ;
funct bderiv [ 3 ] ;
funct c [ 3 ] ;
funct e ;
funct f ;
funct p ;
funct pp ;
}
;
extern double oneF ( point & p ) ;
*****************************************
The PDE object uses vectors of "funct"s, but that is OK on other
systems. But this version crashes.
This is actually a code fragment from a larger project I am porting to
gargoyle, so if you can find what is wrong I would greatly appreciate it.
Thanks!
Phil Keenan
(Message inbox:15)
Received: by gargoyle.uchicago.edu (5.59/1.14)
id AA18285; Fri, 25 Aug 89 09:36:57 198
Date: Fri, 25 Aug 89 09:36:57 198
From: "Tom Barron" <thos>
Message-Id: <8908251436.AA18285@gargoyle.uchicago.edu>
To: piggy
Subject: [keenan%tartarus@gargoyle.uchicago.edu: crash of g++ while compiling on tartarus/gargoyle]
Return-Path: <keenan%tartarus@gargoyle.uchicago.edu>
Date: Thu, 13 Jul 89 15:59:16 CDT
From: keenan%tartarus@gargoyle.uchicago.edu (Philip Keenan)
To: bugs%tartarus@gargoyle.uchicago.edu,
csstaff%tartarus@gargoyle.uchicago.edu
Subject: crash of g++ while compiling on tartarus/gargoyle
Here is a much smaller test program than the previous one I sent you,
which still causes g++ to core-dump while compiling, yet which compiles
fine with CC (on gargoyle only). It makes use of a 2-dimensional array,
but that should not be unreasonable.
Phil Keenan
**************************
/* bug.cc
This very short program still crashes the g++ compiler, causing
a core dump, when it gets to defining 'p' in 'main'.
Phil Keenan 7-13-89
*/
struct funct {
int i;
funct() {}
};
struct pde {
funct a[1][1];
funct b;
};
main()
{
pde p;
}
*************************
Thanks!
(Message inbox:16)
Received: by gargoyle.uchicago.edu (5.59/1.14)
id AA18321; Fri, 25 Aug 89 09:40:47 198
Date: Fri, 25 Aug 89 09:40:47 198
From: "Tom Barron" <thos>
Message-Id: <8908251440.AA18321@gargoyle.uchicago.edu>
To: piggy
Subject: [keenan@tartarus.UChicago.EDU: GNU C++ compiler crash -- another bug]
Date: Fri, 28 Jul 89 13:36:22 CDT
From: keenan@tartarus.UChicago.EDU (Philip Keenan)
To: bugs@tartarus.uchicago.edu
Subject: GNU C++ compiler crash -- another bug
Cc: keenan@tartarus.uchicago.edu
This time I have a program which compiles fine with the command
g++ in.cc
but which crashes the compiler, causing a core dump, with the command
g++ -finline-functions in.cc
As usual, the program I am including is very short, so short that
inlining functions is hardly necessary, but the compiler crash
initially occured in a larger program of mine in which I do wish
to use inlining.
*********************************************
/* in.cc
This is a test of the -finline-functions option which crashes the
GNU C++ compiler. Compile with:
g++ -finline-functions in.cc
Note, it works fine if you just use
g++ in.cc
Phil Keenan 7-28-89
*/
int smallFn(int l,int f)
{
int buf[4];
buf[0]=0;
return buf[0];
}
int callerFn(int l, int f)
{
return smallFn(l,f);
}
int main()
{
return callerFn(0,0);
}
*********************************************
Thanks -- let me know if you need more info!
Phil Keenan
------- End of Forwarded Message
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads