int compiler error bootstrapping on openserver

int compiler error bootstrapping on openserver

[Robert Lipe]

OpenServer worked pretty well oh, about three days ago.  I just did
an update (and did another since seeing this just in case things were
wierded out temporarily) and a make.  I did another make just to be sure
this wasn't a fleeting (RAM error) problem.   It seems repeatable.

host==target==i486-pc-sco3.2v5.0.4

The make proceeds OK until we get to:

/home/play/negcs/gcc/xgcc -B/home/play/negcs/gcc/ -O2   -DIN_GCC    -g -O2 -Wall  -I./include   -g1  -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -fexceptions    -I. -I../../egcs/gcc -I../../egcs/gcc/config \
  -c ../../egcs/gcc/cp/tinfo.cc
../../egcs/gcc/cp/tinfo.cc: In method `void * __class_type_info::dcast(const class type_info &, int, void *, const class type_info * = 0, void * = 0) const':
../../egcs/gcc/cp/tinfo.cc:80: warning: comparison between signed and unsigned
include/typeinfo: At top level:
include/typeinfo:64: Internal compiler error.
include/typeinfo:64: Please submit a full bug report to `egcs-bugs@cygnus.com'.
make[4]: *** [tinfo.o] Error 1
make[3]: *** [libgcc2.a] Error 1
make[2]: *** [stmp-multilib-sub] Error 2
make[1]: *** [stmp-multilib] Error 1
make: *** [all-gcc] Error 2

I just did a 'cvs update' and resovled even the generated files
(configure, parse,etc.) against those in the tree and still get the
same.

I know we've seen some ICE's while building tinfo.cc before in quarters
past, but I really can't recall the details.

This isn't during the coff build, so it should be pretty close to the same
thing the linux crowd would be doing.    Are you guys seeing that fail?

 
RJL

Re: int compiler error bootstrapping on openserver

[Manfred Hollstein]

On Fri, 10 April 1998, 18:45:43, robertl@dgii.com wrote:

 > OpenServer worked pretty well oh, about three days ago.  I just did
 > an update (and did another since seeing this just in case things were
 > wierded out temporarily) and a make.  I did another make just to be sure
 > this wasn't a fleeting (RAM error) problem.   It seems repeatable.
 > 
 > host==target==i486-pc-sco3.2v5.0.4
 > 
 > The make proceeds OK until we get to:
 > 
 > /home/play/negcs/gcc/xgcc -B/home/play/negcs/gcc/ -O2   -DIN_GCC    -g -O2 -Wall  -I./include   -g1  -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -fexceptions    -I. -I../../egcs/gcc -I../../egcs/gcc/config \
 >   -c ../../egcs/gcc/cp/tinfo.cc
 > ../../egcs/gcc/cp/tinfo.cc: In method `void * __class_type_info::dcast(const class type_info &, int, void *, const class type_info * = 0, void * = 0) const':
 > ../../egcs/gcc/cp/tinfo.cc:80: warning: comparison between signed and unsigned
 > include/typeinfo: At top level:
 > include/typeinfo:64: Internal compiler error.
 > include/typeinfo:64: Please submit a full bug report to `egcs-bugs@cygnus.com'.
 > make[4]: *** [tinfo.o] Error 1
 > make[3]: *** [libgcc2.a] Error 1
 > make[2]: *** [stmp-multilib-sub] Error 2
 > make[1]: *** [stmp-multilib] Error 1
 > make: *** [all-gcc] Error 2
 > 
 > I just did a 'cvs update' and resovled even the generated files
 > (configure, parse,etc.) against those in the tree and still get the
 > same.
 > 
 > I know we've seen some ICE's while building tinfo.cc before in quarters
 > past, but I really can't recall the details.
 > 
 > This isn't during the coff build, so it should be pretty close to the same
 > thing the linux crowd would be doing.    Are you guys seeing that fail?

Yep, same here on i586-linux-gnulibc1.

manfred

Re: int compiler error bootstrapping on openserver

[Jean-Pierre Radley]

Robert Lipe telecommunicated (on 10Apr):
| OpenServer worked pretty well oh, about three days ago.  I just did
| an update (and did another since seeing this just in case things were
| wierded out temporarily) and a make.  I did another make just to be sure
| this wasn't a fleeting (RAM error) problem.   It seems repeatable.
| 
| host==target==i486-pc-sco3.2v5.0.4
| 
| The make proceeds OK until we get to:
| 
| /home/play/negcs/gcc/xgcc -B/home/play/negcs/gcc/ -O2   -DIN_GCC    -g -O2 -Wall  -I./include   -g1  -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -fexceptions    -I. -I../../egcs/gcc -I../../egcs/gcc/config \
|   -c ../../egcs/gcc/cp/tinfo.cc
| ../../egcs/gcc/cp/tinfo.cc: In method `void * __class_type_info::dcast(const class type_info &, int, void *, const class type_info * = 0, void * = 0) const':
| ../../egcs/gcc/cp/tinfo.cc:80: warning: comparison between signed and unsigned
| include/typeinfo: At top level:
| include/typeinfo:64: Internal compiler error.
| include/typeinfo:64: Please submit a full bug report to `egcs-bugs@cygnus.com'.
| make[4]: *** [tinfo.o] Error 1
| make[3]: *** [libgcc2.a] Error 1
| make[2]: *** [stmp-multilib-sub] Error 2
| make[1]: *** [stmp-multilib] Error 1
| make: *** [all-gcc] Error 2
| 
| I just did a 'cvs update' and resovled even the generated files
| (configure, parse,etc.) against those in the tree and still get the
| same.
| 
| I know we've seen some ICE's while building tinfo.cc before in quarters
| past, but I really can't recall the details.
| 
| This isn't during the coff build, so it should be pretty close to the same
| thing the linux crowd would be doing.    Are you guys seeing that fail?

OSR worked fine for me yesterday, but not a few hours ago.
I got an error requesting me to submit a bug report too, but didn't do so.
I think it was NOT the one you saw.

I've done a total refetch, and it barfs, pretty much like yours does:

make[4]: Entering directory `/s/tools/egcs/obj/gcc'
/s/tools/egcs/obj/gcc/xgcc -B/s/tools/egcs/obj/gcc/ -O2   -DIN_GCC    -O3 -I./in
clude   -g1  -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -fexceptions    -I. -I../../g
cc -I../../gcc/config \
  -c ../../gcc/cp/tinfo.cc
include/typeinfo:64: Internal compiler error.
include/typeinfo:64: Please submit a full bug report to `egcs-bugs@cygnus.com'.


|  
| RJL

-- 
Jean-Pierre Radley <jpr@jpr.com>  XC/XT Custodian   Sysop, CompuServe SCOForum

Re: int compiler error bootstrapping on openserver [also onm68k-next-nextstep3]

[Toon Moene]

Robert wrote:

> host==target==i486-pc-sco3.2v5.0.4

> The make proceeds OK until we get to:

>  include/typeinfo: At top level:
>  include/typeinfo:64: Internal compiler error.
>  include/typeinfo:64: Please submit a full bug report to
>  `egcs-bugs@cygnus.com'.

Same over here, after updating this morning (ca. 10:00 UTC), on  
m68k-next-nextstep3, so it's probably pretty generic:

(gdb) run tinfo.ii  -quiet -dumpbase tinfo.cc -g -g1 -O2 -O2 -O2 -W  
-Wall -version -fexceptions -o /usr/tmp/cc027905.s
Starting program:  
/Users/toon/Unix/compilers/snapshots/obj/gcc/cc1plus tinfo.ii   
-quiet -dumpbase tinfo.cc -g -g1 -O2 -O2 -O2 -W -Wall -version  
-fexceptions -o /usr/tmp/cc027905.s
GNU C++ version egcs-2.91.22 980404 (gcc2 ss-980401 experimental)  
(m68k-next-nextstep3) compiled by GNU C version egcs-2.91.22 980404  
(gcc2 ss-980401 experimental).
.../../egcs/gcc/cp/tinfo.cc: In method `void *  
__class_type_info::dcast(const class type_info &, int, void *, const  
class type_info * = 0, void * = 0) const':
.../../egcs/gcc/cp/tinfo.cc:80: warning: comparison between signed  
and unsigned
Program generated(1): Memory access exception on address 0x10  
(protection failure).
Reading in symbols for ../../../egcs/gcc/cp/decl2.c...done.
0x2a0f2 in mark_vtable_entries (decl=void) at  
.../../../egcs/gcc/cp/decl2.c:2359
2359          tree fnaddr = (flag_vtable_thunks ? TREE_VALUE  
(TREE_CHAIN (entries))
(gdb) p flag_vtable_thunks
$1 = 0
(gdb) list
2354    {
2355      tree entries = CONSTRUCTOR_ELTS (DECL_INITIAL (decl));
2356
2357      if (flag_rtti)
2358        {
2359          tree fnaddr = (flag_vtable_thunks ? TREE_VALUE  
(TREE_CHAIN (entries))
2360                         : FNADDR_FROM_VTABLE_ENTRY (TREE_VALUE  
(entries)));
2361          tree fn = TREE_OPERAND (fnaddr, 0);
2362          TREE_ADDRESSABLE (fn) = 1;
2363          mark_used (fn);
(gdb) p *entries
Attempt to take contents of a non-pointer value.
(gdb) p entries
$2 = void
(gdb) bt
Reading in symbols for ../../../egcs/gcc/cp/parse.c...done.
Reading in symbols for ../../egcs/gcc/toplev.c...done.
#0  0x2a0f2 in mark_vtable_entries (decl=void) at  
.../../../egcs/gcc/cp/decl2.c:2359
#1  0x2a46e in finish_vtable_vardecl (prev=void, vars=void) at  
.../../../egcs/gcc/cp/decl2.c:2534
#2  0x2a51e in walk_vtables (typedecl_fn=0, vardecl_fn=0x2a3fc  
<finish_vtable_vardecl>) at ../../../egcs/gcc/cp/decl2.c:2600
#3  0x2afa2 in finish_file () at ../../../egcs/gcc/cp/decl2.c:3138
#4  0x37a3e in yyparse () at parse.y:336
#5  0x5d026 in compile_file (name=0x3fffbe5 "tinfo.ii") at  
.../../egcs/gcc/toplev.c:2519
#6  0x5ef96 in main (argc=16, argv=0x1, envp=0x3fffb80) at  
.../../egcs/gcc/toplev.c:4340
(gdb)

Hope someone can make something out of this - it's C++, so my NIH  
alarm immediately closed down all relevant brain capacity ...

Regards,
Toon.