[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: bison error on OSX: "error_one_per_line" (?)
From: |
R Blake |
Subject: |
Re: bison error on OSX: "error_one_per_line" (?) |
Date: |
Mon, 01 Jul 2002 09:04:52 -0700 |
Yep: the CC to bug-bison.
ooops, sorry :-S
Other than that, I am the one who missed an important point: we
compile this file only when needed, and it appears it is not `needed'
on your platform, but it is not the same version as the usual one,
since it does not seem to provide `error_one_per_line'. That's weird.
Do you have a `man error_at_line' or something? Could you also look
for `error_at_line' in config.log? I have this on my machine:
unfortunately, no such man page (cref. http://www.hmug.org/man/e.html) on
the current release.
i don't understand your comment "but it is not the same version as the
usual one" .... which version of what?
ok, now i'm really confused ..... i re dl'd 1.49a, and started from scratch
.... still no error.o, BUT, when i'm a bit paranoid and start over with:
glibtoolize --force --copy; aclocal -I m4 ; autoconf ; autoheader ;
automake ; ./configure --prefix=/usr/local/bison
--exec-prefix=/usr/local/bison --mandir=/usr/local/man --disable-nls; make
all goes well, and NOW, i DO see error.o! nm gives:
address@hidden error.o
U ___sF
U ___swbuf
00000128 T _error
000001e4 T _error_at_line
00000488 S _error_message_count
0000048c S _error_one_per_line
00000484 S _error_print_progname
00000058 t _error_tail
U _exit
U _fflush
U _fprintf
0000047c b _old_file_name.15
00000480 b _old_line_number.16
00000000 t _print_errno_message
U _program_name
U _strcmp
U _strerror
U _vfprintf
U dyld_stub_binding_helper
address@hidden
configure:5418: checking for error_at_line
configure:5442: gcc -o conftest -g -O2 conftest.c >&5
configure:5445: $? = 0
configure:5448: test -s conftest
configure:5451: $? = 0
configure:5461: result: yes
and the relevant info from config.log is (fail on check):
configure:16667: checking for error_at_line
configure:16691: cc -o conftest -O3 -L/System/Library/Frameworks -L/usr/lib
-L/usr/local/lib -L/usr/X11R6/lib -L/usr/local/BerkeleyDB/lib
-fno-common -fno-rtti -fno-strict-aliasing
-I/System/Library/Frameworks -I/usr/local/include -I/usr/X11R6/include
-I/usr/local/BerkeleyDB/include -L/System/Library/Frameworks
-L/usr/lib -L/usr/local/lib -L/usr/X11R6/lib -L/usr/local/BerkeleyDB/lib
conftest.c >&5
/usr/bin/ld: Undefined symbols:
_error_at_line
configure:16694: $? = 1
configure: failed program was:
#line 16672 "configure"
#include "confdefs.h"
#ifdef F77_DUMMY_MAIN
# ifdef __cplusplus
extern "C"
# endif
int F77_DUMMY_MAIN() { return 1; }
#endif
int
main ()
{
error_at_line (0, 0, "", 0, "");
;
return 0;
}
configure:16710: result: no
and, after make install, bison --version NO LONGER gives me an error .....
address@hidden --version
bison (GNU Bison) 1.49a
Written by Robert Corbett and Richard Stallman.
Copyright (C) 2002 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
address@hidden
i'm not sure what fixed what, but it seems to work now .... any thoughts
as to what specifically changed?
richard
--On Monday, July 1, 2002 5:15 PM +0200 Akim Demaille <address@hidden> wrote:
"R" == R Blake <R> writes:
R> hi, hmmmm........
R> well i *tried* running nm on error.o, only to find that,
R> unfortunately, error.o is not created! here's the dir ls after my
R> 'successful' make ....
R> address@hidden /Volumes/ServerSystem/ports/bison-1.49a/lib
R> Makefile bitset.o getopt.c libbison.a quotearg.h strspn.c
R> Makefile.am bitsetv.c getopt.h malloc.c quotearg.o tempname.c
R> Makefile.in bitsetv.h getopt.o memchr.c readpipe.c xalloc.h
R> alloca.c bitsetv.o getopt1.c memrchr.c readpipe.o xmalloc.c
R> ansi2knr.1 dirname.c getopt1.o mkstemp.c realloc.c xmalloc.o
R> ansi2knr.c dirname.h gettext.h obstack.c sbitset.c xstrdup.c
R> basename.c dirname.o hash.c obstack.h sbitset.h xstrdup.o
R> basename.o ebitset.c hash.h obstack.o sbitset.o xstrndup.c
R> bbitset.h ebitset.h hash.o quote.c stpcpy.c xstrndup.o bitset-int.h
R> ebitset.o lbitset.c quote.h strchr.c bitset.c error.c lbitset.h
R> quote.o strnlen.c bitset.h error.h lbitset.o quotearg.c strnlen.o
R> wondering if, perhaps, i'd accidentally deleted it (not likeley,
R> tho) i started over with a clean dist & make. There's still no
R> error.o! (full make session below ....)
R> a quick check shows that there's no reference to error.c in
R> ./Makefile, but that there *is* in ./libs/Makefile.in, and that
R> Makefile does pick up the reference in the "ac_config_files=" def'n
R> .....
R> am i missing something here?
--------------------------------------
R Blake
blakers <at> mac.com
http://homepage.mac.com/blakers
--------------------------------------
MacOSX ..... because Windows sux.
--------------------------------------