RE: Problems configuring cfengine 2.1.17 on Solaris10

[Martin, Jason H]

The 'real' error is:
ld.so.1: ./conftest: fatal: libgcc_s.so.1: open failed: No such file or
directory
So, the problem isn't with Berkeley DB, except that perhaps BDB requires
libgcc. Is libgcc installed?

-Jason Martin



> -----Original Message-----
> From: Kurt Reimer [mailto:greimer@fccc.edu] 
> Sent: Wednesday, December 21, 2005 8:14 AM
> To: Martin, Jason H
> Cc: Kurt.Reimer@fccc.edu; help-cfengine@gnu.org; bug-cfengine@gnu.org
> Subject: RE: Problems configuring cfengine 2.1.17 on Solaris10
> 
> 
> >
> > What is the exact output from the config.log for when the problem
> > occurs?
> >
> > -Jason Martin
> >
> Jason,
>       Thanks for taking a look at this. Here's the relevant 
> part of the 
> log output from the original configure script, with my email client 
> inserting line breaks and me adding in backslashes:
> 
> ======================================================================
> configure:8582: checking Berkeley DB Version
> configure:8611: gcc -o conftest -D_REENTRANT -pthread -g -O2      \
> -I/usr/local/BerkeleyDB.4.4//include -D_REENTRANT -pthread        \
> -L/usr/local/BerkeleyDB.4.4//lib -R/usr/local/BerkeleyDB.4.4//lib \
> conftest.c -lnsl -lsocket -lm  -ldb >&5
> gcc: unrecognized option `-pthread'
> gcc: unrecognized option `-pthread'
> configure:8614: $? = 0
> configure:8616: ./conftest
> ld.so.1: ./conftest: fatal: libgcc_s.so.1: open failed: No such   \
>   file or directory
> ./configure: line 8617: 26012 Killed        ./conftest$ac_exeext
> configure:8619: $? = 137
> configure: program exited with status 137
> configure: failed program was:
> | /* confdefs.h.  */
> |
> | #define PACKAGE_NAME ""
> | #define PACKAGE_TARNAME ""
> | #define PACKAGE_VERSION ""
> | #define PACKAGE_STRING ""
> | #define PACKAGE_BUGREPORT ""
> | #define PACKAGE "cfengine"
> | #define VERSION "2.1.17"
> | #define HAVE_PTHREAD 1
> | #define YYTEXT_POINTER 1
> | #define STDC_HEADERS 1
> | #define HAVE_SYS_TYPES_H 1
> | #define HAVE_SYS_STAT_H 1
> | #define HAVE_STDLIB_H 1
> | #define HAVE_STRING_H 1
> | #define HAVE_MEMORY_H 1
> | #define HAVE_STRINGS_H 1
> | #define HAVE_INTTYPES_H 1
> | #define HAVE_STDINT_H 1
> | #define HAVE_UNISTD_H 1
> | #define HAVE_LIBKSTAT 1
> | #define HAVE_GETLOADAVG 1
> | #define HAVE_LIBM 1
> | #define HAVE_LIBSOCKET 1
> | #define HAVE_LIBNSL 1
> | #define USE_BERKELEY_DB 1
> | /* end confdefs.h.  */
> |
> | #include <db.h>
> | #include <stdio.h>
> | int main(void)
> | {
> |     printf("%d.%d.%d 
> ",DB_VERSION_MAJOR,DB_VERSION_MINOR,DB_VERSION_PATCH);
> |     if (DB_VERSION_MAJOR < 3 || (DB_VERSION_MAJOR == 3 &&    \
> DB_VERSION_MINOR < 2))
> |         exit(1);
> |     exit(0);
> | }
> |
> configure:8629: error: This release of cfengine requires       \ 
> BerkeleyDB 3.2 or later
> ======================================================================
> 
>       The second invocation of this program doesn't cause the 
> configure 
> script to abort, but it does define CF_OLD_DB, which 
> generates compile 
> errors during the make.
> 
>       After collecting this output, I tried running the gcc compile 
> command as it appears above on a copy of the test program (without the
> output redirection), and it does compile OK and generate the same
> error message when I run it. I find that I can make the program run
> successfully if I take the "-lbd" off the compile string OR if I
> go into /usr/local/BerkeleyDB.4.4/lib and remove all the shared 
> libraries. (I guess this forces use of the static libraries in that 
> directory. I know that when I remove them too the compile fails.)
> 
>       I'm not sure what all this means, but I now have less 
> confidence 
> that the cfengine executables I ended up with are worth anything. Does
> this mean that I built the Berkeley DB incorrectly?
> 
> Yours,
> 
> Kurt
> 
> >> -----Original Message----- >> 
> From: >> help-cfengine-bounces+jason.h.martin=cingular.com@gnu.org
> >> [mailto:help-cfengine-bounces+jason.h.martin=cingular.com@gnu.
> >> org] On Behalf Of Kurt Reimer
> >> Sent: Tuesday, December 20, 2005 1:56 PM
> >> To: help-cfengine@gnu.org; bug-cfengine@gnu.org
> >> Subject: Problems configuring cfengine 2.1.17 on Solaris10
> >>
> >>
> >>
> >> Hello,
> >>      Like previous posters I'm running into problems with the
> >> configure
> >> script not recognizing the BerkeleyDB version number, and it
> >> doesn't seem
> >> to have anything to do with LD_LIBRARY_PATH or LD_RUN_PATH:
> >>
> >> *******************
> >> bash-3.00# export \
> >> LD_LIBRARY_PATH=/usr/local/BerkeleyDB.4.4/:/usr/local/Berkeley
> >> DB.4.4/lib/
> >> bash-3.00# export \
> >> 
> LD_RUN_PATH=/usr/local/BerkeleyDB.4.4/:/usr/local/BerkeleyDB.4.4/lib/
> >> bash-3.00# export CFLAGS=-I/usr/local/BerkeleyDB.4.4/include
> >> bash-3.00# ./configure
> >> --with-berkeleydb=/usr/local/BerkeleyDB.4.4/ \
> >> --with-openssl=/usr/sfw checking build system type...
> >> sparc-sun-solaris2.10
> >> .   .   .   .   .   .   .   .   .   .   .   .   .   .
> >> checking for BerkeleyDB location in /usr/local/BerkeleyDB.4.4/...
> >> /usr/local/BerkeleyDB.4.4/
> >> checking Berkeley DB Version... configure: error: This
> >> release of cfengine
> >> requires BerkeleyDB 3.2 or later
> >> bash-3.00#
> >>
> >>        (I added the backslashes where my email client wrapped the
> >> cut-and-pasted command lines.)
> >> *******************
> >>
> >>       I've gone so far as to comment out all but the
> >> "exit(0);" return in
> >> the test program that is being run at this point in the
> >> configure script, with the same result:
> >>
> >>
> >> *******************
> >> #include <db.h>
> >> #include <stdio.h>
> >> int main(void)
> >> {
> >>      /*
> >>      printf("%d.%d.%d ",
> >>             DB_VERSION_MAJOR,DB_VERSION_MINOR,DB_VERSION_PATCH);
> >>      if (DB_VERSION_MAJOR < 3 ||
> >>          (DB_VERSION_MAJOR == 3 && DB_VERSION_MINOR < 2))
> >>          exit(1);
> >>      */
> >>      exit(0);
> >> }
> >> *******************
> >>
> >>       I can get cfengine to configure, make, and install if I
> >> edit the
> >> configure script to dummy out the two places where it runs the test
> >> program listed above. Note that the configure, make, and make
> >> install all
> >> succeed without the help of LD_LIBRARY_PATH, LD_RUN_PATH, or
> >> any special include directories defined in CFLAGS:
> >>
> >> *******************
> >> bash-3.00# diff configure.works configure
> >> 8618,8620c8618
> >> < ##############################
> >> < # ac_status=$? This test is broken
> >> <   ac_status=0
> >> ---
> >>>   ac_status=$?
> >> 8687,8690c8685,8687
> >> < ################################
> >> < #cat >>confdefs.h <<\_ACEOF Same program, still broken
> >> < ##define CF_OLD_DB 1
> >> < #_ACEOF
> >> ---
> >>> cat >>confdefs.h <<\_ACEOF
> >>> #define CF_OLD_DB 1
> >>> _ACEOF
> >> bash-3.00# cp configure.works configure
> >> bash-3.00# unset LD_LIBRARY_PATH
> >> bash-3.00# unset LD_RUN_PATH
> >> bash-3.00# unset CFLAGS
> >> bash-3.00# ./configure --with-berkeleydb=/usr/local/BerkeleyDB.4.4/
> >> --with-openssl=/usr/sfw
> >> checking build system type... sparc-sun-solaris2.10
> >> .   .   .   .   .   .   .   .   .   .   .   .   .   .
> >> checking for library containing socket... none required
> >> checking for BerkeleyDB location in /usr/local/BerkeleyDB.4.4/...
> >> /usr/local/BerkeleyDB.4.4/
> >> checking Berkeley DB Version... OK
> >> checking Berkeley DB API... checking for db_create in -ldb...
> >> yes checking for OpenSSL location... /usr/sfw
> >> .   .   .   .   .   .   .   .   .   .   .   .   .   .
> >> DONE: Configuration done. Run make/gmake to build cfengine.
> >>
> >> bash-3.00# make[1]: Entering directory
> >> `/usr/local/src/cfengine-2.1.17/pub'
> >> if gcc -DHAVE_CONFIG_H -I. -I. -I../src
> >> -I/usr/local/BerkeleyDB.4.4//include -I/usr/sfw/include  
> -D_REENTRANT
> >> -pthread  -D_REENTRANT -pthread -g -O2
> >> -I/usr/local/BerkeleyDB.4.4//include -D_LARGEFILE_SOURCE
> >> -D_FILE_OFFSET_BITS=64 -g -O2 -Wreturn-type -Wmissing-prototypes
> >> -Wuninitialized -D_REENTRANT -pthread -g -O2
> >> -I/usr/local/BerkeleyDB.4.4//include -D_LARGEFILE_SOURCE
> >> -D_FILE_OFFSET_BITS=64 -MT getopt.o -MD -MP -MF
> >> ".deps/getopt.Tpo" -c -o
> >> getopt.o getopt.c; \
> >> then mv -f ".deps/getopt.Tpo" ".deps/getopt.Po"; else rm -f
> >> ".deps/getopt.Tpo"; exit 1; fi
> >> gcc: unrecognized option `-pthread'
> >> gcc: unrecognized option `-pthread'
> >> gcc: unrecognized option `-pthread'
> >> if gcc -DHAVE_CONFIG_H -I. -I. -I../src
> >> -I/usr/local/BerkeleyDB.4.4//include -I/usr/sfw/include  
> -D_REENTRANT
> >> -pthread  -D_REENTRANT -pthread -g -O2
> >> -I/usr/local/BerkeleyDB.4.4//include -D_LARGEFILE_SOURCE
> >> -D_FILE_OFFSET_BITS=64 -g -O2 -Wreturn-type -Wmissing-prototypes
> >> -Wuninitialized -D_REENTRANT -pthread -g -O2
> >> -I/usr/local/BerkeleyDB.4.4//include -D_LARGEFILE_SOURCE
> >> -D_FILE_OFFSET_BITS=64 -MT getopt1.o -MD -MP -MF
> >> ".deps/getopt1.Tpo" -c -o
> >> getopt1.o getopt1.c; \
> >> then mv -f ".deps/getopt1.Tpo" ".deps/getopt1.Po"; else rm -f
> >> ".deps/getopt1.Tpo"; exit 1; fi
> >> .   .   .   .   .   .   .   .   .   .
> >>
> >> .....and so on....
> >>
> >> *******************
> >>
> >>       "make" and "make install" both succeed. I haven't yet
> >> begun to try to
> >> use cfengine, but I suspect the executables will work properly.
> >>
> >>       It would be nice if the configuration process would work
> >> out-of-the-box. If I'm dong something wrong please let me know.
> >>
> >> Yours,
> >>
> >> Kurt Reimer
> >>
> >>
> >> _______________________________________________
> >> Help-cfengine mailing list
> >> Help-cfengine@gnu.org
> >> http://lists.gnu.org/mailman/listinfo/help-> cfengine
> >>
> >
>