LINKING F90 programs leads to "multiply defined" and "undefined" CMA$TIS

LINKING F90 programs leads to "multiply defined" and "undefined" CMA$TIS

Post by Jerome LAUR » Fri, 19 May 2000 04:00:00



       I don't know when this happened but recently, I recompiled a program
with the F90 compiler and everything went OK as before. However, the LINK command

LINK/EXE=[.exe.VMS_ALPHA]thermalmodel.EXE [.obj.VMS_ALPHA]thermalmodel.OBJ ,[.li
b.VMS_ALPHA]libfreezer/LIB,[.lib.VMS_ALPHA]libhydro/LIB ,CERN_ROOT:[LIB]MATHLIB.
OLB/LIB,CERN_ROOT:[LIB]KERNLIB.OLB/LIB

        where libfreezer, libhydro are libraries I compiled myself and
the cern lib, a version taken since OpenVMS6.2 Alpha.
        In anycasethe above command leads to zillion of messages like below.

        Any idea of why I have this now ?

        If I compile and LINK one source code only (using ALLOCATE for example)
and LINK, it is fine.

        My platform

OpenVMS 7.2
Compaq Fortran V7.3-965-44A1I

        Any ideas ??

%LINK-W-MULDEF, symbol DFOR$ALLOCATE multiply defined
        in module FOR_VM file SYS$COMMON:[SYSLIB]STARLET.OLB;4
%LINK-W-MULDEF, symbol DFOR$ALLOC_ALLOCATABLE multiply defined
        in module FOR_VM file SYS$COMMON:[SYSLIB]STARLET.OLB;4
%LINK-W-MULDEF, symbol DFOR$DEALLOCATE multiply defined
        in module FOR_VM file SYS$COMMON:[SYSLIB]STARLET.OLB;4
%LINK-W-MULDEF, symbol DFOR$DEALLOC_ALLOCATABLE multiply defined
        in module FOR_VM file SYS$COMMON:[SYSLIB]STARLET.OLB;4
%LINK-W-MULDEF, symbol DFOR$SET_REENTRANCY multiply defined
        in module FOR_REENTRANCY file SYS$COMMON:[SYSLIB]STARLET.OLB;4
%LINK-W-NUDFSYMS, 18 undefined symbols:
%LINK-I-UDFSYM,         CMA$DECC$G_REENTRANCY
%LINK-I-UDFSYM,         CMA$TIS_ERRNO_GET_ADDR
%LINK-I-UDFSYM,         CMA$TIS_ERRNO_SET_ADDR
%LINK-I-UDFSYM,         CMA$TIS_ERRNO_SET_VALUE
%LINK-I-UDFSYM,         CMA$TIS_KEY_CREATE
%LINK-I-UDFSYM,         CMA$TIS_KEY_GET_CONTEXT
%LINK-I-UDFSYM,         CMA$TIS_KEY_SET_CONTEXT
%LINK-I-UDFSYM,         CMA$TIS_MUTEX_CREATE
%LINK-I-UDFSYM,         CMA$TIS_MUTEX_DELETE
%LINK-I-UDFSYM,         CMA$TIS_MUTEX_LOCK
%LINK-I-UDFSYM,         CMA$TIS_MUTEX_TRYLOCK
%LINK-I-UDFSYM,         CMA$TIS_MUTEX_UNLOCK
%LINK-I-UDFSYM,         CMA$TIS_ONCE
%LINK-I-UDFSYM,         CMA$TIS_VMSERRNO_GET_ADDR
%LINK-I-UDFSYM,         CMA$TIS_VMSERRNO_SET_ADDR
%LINK-I-UDFSYM,         CMA$TIS_VMSERRNO_SET_VALUE
%LINK-I-UDFSYM,         DFOR$$FILE_INFO_HASH_TABLE
%LINK-I-UDFSYM,         DFOR$$LUB_TABLE

--

       ,,,,,      Dept. of Chemistry
      ( o o )     Stony Brook NY 11794-3400
  ---m---U---m---------------------------------------------

  URL   : http://nucwww.chem.sunysb.edu/jlauret/jlauret.html

 
 
 

LINKING F90 programs leads to "multiply defined" and "undefined" CMA$TIS

Post by Steve Lione » Sat, 20 May 2000 04:00:00




>       I don't know when this happened but recently, I recompiled a program
>with the F90 compiler and everything went OK as before. However, the LINK command

>LINK/EXE=[.exe.VMS_ALPHA]thermalmodel.EXE [.obj.VMS_ALPHA]thermalmodel.OBJ ,[.li
>b.VMS_ALPHA]libfreezer/LIB,[.lib.VMS_ALPHA]libhydro/LIB ,CERN_ROOT:[LIB]MATHLIB.
>OLB/LIB,CERN_ROOT:[LIB]KERNLIB.OLB/LIB

>    where libfreezer, libhydro are libraries I compiled myself and
>the cern lib, a version taken since OpenVMS6.2 Alpha.
>    In anycasethe above command leads to zillion of messages like below.

[snip]

Interesting.  Do you have the LINK symbol defined as LINK/NOSYSSHR,
perhaps?  The symptom suggests that SYS$LIBRARY:IMAGELIB.OLB is not
getting searched before STARLET.OLB.  Try adding
SYS$LIBRARY:IMAGELIB/LIB to the end of your link commnd and see what
happens.


Fortran Engineering
Compaq Computer Corporation, Nashua NH

Compaq Fortran web site: http://www.compaq.com/fortran

 
 
 

LINKING F90 programs leads to "multiply defined" and "undefined" CMA$TIS

Post by Jerome LAUR » Sat, 20 May 2000 04:00:00



[...]
|>
|>Interesting.  Do you have the LINK symbol defined as LINK/NOSYSSHR,
|>perhaps?  The symptom suggests that SYS$LIBRARY:IMAGELIB.OLB is not
|>getting searched before STARLET.OLB.  Try adding
|>SYS$LIBRARY:IMAGELIB/LIB to the end of your link commnd and see what
|>happens.
|>
|>

        Hello Steve,

        First : thanks for your quick answer. Unfortunatly, your suggestion
has not solved the problem and I get the same messages all over again. I am
almost at the point of re-installing the FORtran compiler now just to see
what would happen. Shall I do that or is there any other tricks I can try ?
(like a lib/replace of some kind ?).

        Thanks in advance for any helps and tips,

--

       ,,,,,      Dept. of Chemistry
      ( o o )     Stony Brook NY 11794-3400
  ---m---U---m---------------------------------------------

  URL   : http://nucwww.chem.sunysb.edu/jlauret/jlauret.html

 
 
 

LINKING F90 programs leads to "multiply defined" and "undefined" CMA$TIS

Post by Steve Lione » Sat, 20 May 2000 04:00:00




>    First : thanks for your quick answer. Unfortunatly, your suggestion
>has not solved the problem and I get the same messages all over again. I am
>almost at the point of re-installing the FORtran compiler now just to see
>what would happen. Shall I do that or is there any other tricks I can try ?
>(like a lib/replace of some kind ?).

Did you, by any chance, upgrade VMS sometime after installing the
Fortran kit?  Please try reinstalling the Fortran RTL (FORRTL) kit and
see if that solves the problem. It may be that the Fortran-provided
version of the Run-Time Library got replaced with an older one.  Also,
make sure that you don't have any other libraries in a LNK$LIBRARY
path.  Examining a link map will tell you where everything is coming
from.


Fortran Engineering
Compaq Computer Corporation, Nashua NH

Compaq Fortran web site: http://www.compaq.com/fortran

 
 
 

LINKING F90 programs leads to "multiply defined" and "undefined" CMA$TIS

Post by Jerome LAUR » Tue, 23 May 2000 04:00:00



[...]
|>
|>Did you, by any chance, upgrade VMS sometime after installing the
|>Fortran kit?  Please try reinstalling the Fortran RTL (FORRTL) kit and

        No I did not. It is not the first time I am having problem with the
FORtran compiler/RTL etc ... since the 7.2 upgrade.

        But never mind, the re-installation of the FORtran RTL fixed the
problem.

        Thanks again for your quick response,

--

DEC AXPVMS FORRTL V7.3-1            Full LP     Install     22-MAY-2000 10:51:38
CPQ AXPVMS MOZILLA F5.0             Full LP     Install     14-MAY-2000 16:52:02
CPQ AXPVMS MOZILLA F5.0             Full LP     Install     14-MAY-2000 16:52:02
CPQ AXPVMS MOZILLA E5.0             Full LP     Remove      14-MAY-2000 16:52:02
CPQ AXPVMS MOZILLA E5.0             Full LP     Install     14-MAY-2000 16:41:54
CPQ AXPVMS MOZILLA E5.0             Full LP     Install     20-APR-2000 20:53:33
CPQ AXPVMS MOZILLA F5.0             Full LP     Remove      20-APR-2000 20:53:33
CPQ AXPVMS MOZILLA F5.0             Full LP     Install     20-APR-2000 20:40:17
CPQ AXPVMS MOZILLA E5.0             Full LP     Remove      20-APR-2000 20:40:17
CPQ AXPVMS MOZILLA E5.0             Full LP     Install     11-APR-2000 22:07:48
CPQ AXPVMS MOZILLA D5.0             Full LP     Remove      11-APR-2000 22:07:48
DEC AXPVMS VMS72_SYS V3.0           Patch       Install     11-APR-2000 20:41:58
DEC AXPVMS VMS72_TRACE V1.0         Patch       Install     04-APR-2000 00:02:43
DEC AXPVMS VMS72_PORTS V1.0         Patch       Install     04-APR-2000 00:01:14
DEC AXPVMS VMS72_SYS V2.0           Patch       Install     03-APR-2000 23:58:32
DEC AXPVMS FORTRAN V7.3-1           Full LP     Install     02-MAR-2000 13:55:45
DEC AXPVMS FORTRAN V7.1-1           Full LP     Remove      02-MAR-2000 13:55:45
DEC AXPVMS FORRTL V7.3-1            Full LP     Install     02-MAR-2000 13:40:01
DEC AXPVMS FORRTL V7.1-427          Full LP     Remove      02-MAR-2000 13:40:01
..
DEC AXPVMS OPENVMS V7.2             Platform    Install     24-FEB-2000 14:56:20


       ,,,,,      Dept. of Chemistry
      ( o o )     Stony Brook NY 11794-3400
  ---m---U---m---------------------------------------------

  URL   : http://nucwww.chem.sunysb.edu/jlauret/jlauret.html