Undefined symbols make ing GCC 3.2.2 on Linux

[SHolton]

Linux 2.4.17, Dell Dimension pIII, 750MHz, 384MB

Building GCC 3.2.2 (from 'stable' release download at gnu.org)

configure shows

... (other stuff listed here)

checking for limits.h... no

... (more stuff listed here)



make bootstrap gets:

... (other stuff listed here)

else true; fi

gcc -c -DHAVE_CONFIG_H -g -O2 -I. -I../../gcc-3.2.2/libiberty/../
include -W -Wall -Wtraditional -pedantic
../../gcc-3.2.2/libiberty/fibheap.c

../../gcc-3.2.2/libiberty/fibheap.c: In function
`fibheap_delete_node':

../../gcc-3.2.2/libiberty/fibheap.c:285: `LONG_MIN' undeclared
(first use in this function)

../../gcc-3.2.2/libiberty/fibheap.c:285: (Each undeclared
identifier is reported only once

../../gcc-3.2.2/libiberty/fibheap.c:285: for each function it
appears in.)

make[1]: *** [fibheap.o] Error 1

make[1]: Leaving directory `/alt/Source/GCCobj86/libiberty'

make: *** [all-libiberty] Error 2

... (no more stuff listed after this...)



I have limits.h in /usr/include, and it defines LONG_MIN.

I also found limits.h in the gcc-3.2.2 directory tree (in
following, '.' is gcc-3.2.2 dir

since I issued the find from there...)

# find . -name limits.h -print

./gcc/fixinc/tests/base/sys/limits.h

./libstdc++-v3/config/cpu/generic/bits/limits.h

./libstdc++-v3/include/c_shadow/limits.h

./libstdc++-v3/include/c_compatibility/limits.h

./include/limits.h



...none of which define LONG_MIN directly, but several include
<climits>.

cd to /usr finds 2 climits:

# find . -name 'climit*' -print

./include/g++/climits

./local/include/g++-3/climits



both of which have no defs of their own (so no LONG_MIN), but
include <limits.h>... kind of circular??

So, apparently make is --not-- looking in /usr/include (or it
would have found LONG_MIN

contained within the limits.h which appears to be

correct (from gcc 2.95.3 installation from too long ago..) and I
am cooked??

I am issuing configure and make in a directory different from the

source. (source != target - source is in gcc-3.2.2 tree and
configure/make issued from a

clean dir - GCCobj86, which is

separate and not part of the gcc-3.2.2 tree.



I googled 'gcc LONG_MIN' and found several threads discussing
this issue, and several

even mentioned that they had built a patch to correct this
problem, but no mention

where I might get this patch.

I am lost - any ideas? I really don't wish to stay on 2.95.3 ....

Thanks


Nothing in this world can take the place of persistence. Talent
will not; nothing is more common than
unsuccessful people with talent. Genius will not; un-rewarded
genius is almost a proverb. Education will not;
the world is full of educated derelicts. Persistence and
determination alone are omnipotent. The slogan
“Press On” has solved and always will solve the problems of the
human race.
—Calvin Coolidge



     _/_/_/    _/
   _/        _/_/_/_/    _/_/    _/      _/    _/_/
    _/_/      _/      _/_/_/_/  _/      _/  _/_/_/_/
       _/    _/      _/          _/  _/    _/
_/_/_/        _/_/    _/_/_/      _/        _/_/_/



    _/    _/            _/    _/
   _/    _/    _/_/    _/  _/_/_/_/    _/_/    _/_/_/
  _/_/_/_/  _/    _/  _/    _/      _/    _/  _/    _/
 _/    _/  _/    _/  _/    _/      _/    _/  _/    _/
_/    _/    _/_/    _/      _/_/    _/_/    _/    _/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://gcc.gnu.org/pipermail/gcc-help/attachments/20030418/54d0d5ca/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: (Vcard) HoltonConsultingLtd..vcf
Type: text/x-vcard
Size: 481 bytes
Desc: not available
URL: <https://gcc.gnu.org/pipermail/gcc-help/attachments/20030418/54d0d5ca/attachment.vcf>