[Paraview] Compile of 2.6.0 Fails on Tru64

Kent Eschenberg eschenbe at psc.edu
Wed Mar 14 13:46:28 EST 2007


There seems to be a problem defining ios routines under Tru64.

Sorry for the long message but I wanted to include all the 
information that might be useful. If you don't use Tru64 you 
probably don't want to read this!

======  Case 1

Set:
    CMAKE_CXX_FLAGS:STRING
    =-std gnu -pthread -verbose
The error:
    src/Servers/Filters/vtkSpyPlotReader.cxx, line 1172:
    class "ios" has no member "binary"

======  Case 2

Set:
    CMAKE_CXX_FLAGS:STRING
    =-std gnu -D__USE_STD_IOSTREAM -pthread -verbose
The error:
    cxx: Error: /usr/local/packages/paraview/2.6.0/src ...
    ... /Servers/Common/vtkCommandOptions.cxx, line 35:
    "ostream" is ambiguous (ambiguousname)
    vtkCxxRevisionMacro( ... );
    ^

======  Suspicious Difference

Before stopping at the above error the file 
vtkClientConnection.cxx compiled just fine (its in the same 
directory). Both use the macro vtkCxxRevisionMacro. So why 
did the macro work in one place but not the other?

The main difference is that the routine that failed includes 
vtksys/SystemTools.hxx which includes vtksys/ios/iosfwd 
which does some things with the io definitions and 
namespaces. It looks like ostream might be defined in more 
than one namespace hence the ambiguity.

======  System Settings

uname -a:
OSF1 rachel.psc.edu V5.1 2650 alpha

Compiler:
Compaq C++ V6.5-014 for Compaq Tru64 UNIX V5.1B (Rev. 2650)
Compiler Driver V6.5-014 (cxx) cxx Driver

This seems relevant: the file 
<build>/VTK/Utilities/vtksys/Configure.hxx contains these 
lines (along with many others):

#define vtksys_IOS_USE_ANSI 0
#define vtksys_IOS_USE_SSTREAM 0
#define vtksys_IOS_USE_STRSTREAM_H 1
#define vtksys_IOS_USE_STRSTREA_H 0
# define KWSYS_IOS_USE_ANSI vtksys_IOS_USE_ANSI
# define KWSYS_IOS_USE_SSTREAM vtksys_IOS_USE_SSTREAM
# define KWSYS_IOS_USE_STRSTREAM_H ...
    ... vtksys_IOS_USE_STRSTREAM_H
# define KWSYS_IOS_USE_STRSTREA_H vtksys_IOS_USE_STRSTREA_H

It might take me a couple of weeks to figure out how all 
these interact! Can anyone suggest a few settings to try?

TIA!
Kent
Pittsburgh Supercomputing Center


More information about the ParaView mailing list