[geos-devel] Exception Issues

[Norman Vine]

Martin Davis writes:
>
> Well, IMO the optimal sequence of fixes to try is:
>
> 1. Fix the compiler/linker/PostgresQL to eliminate the problem
> with aborting on exceptions in loaded libraries.
> 2. Hack GEOS to eliminate or optionalize (love that word!) the exceptions.
>

IMO I am not sure we really have any problem other then that

** C++ is not JAVA **  and GEOS is a very much an attempt at
a literal translation of a JAVA library to C++ where the underlying
JAVA machine handles much of what the C++ programmer has
to handle explicitly as is to be expected with a 'lower level' interface
to the 'bits themselves' there is some confusion as to how to handle
C++ exceptions, which are NOT the same thing as JAVA exceptions

Disclaimer I usually do not use the C++ exception mechanism
so buyer beware..

That said now that we have all of the PostGIS GEOS interaction
in a C++ not a C library we should be able to handle any "standard"
exception thrown by the GEOS code by including it within an appropriate
 try block

i.e

#include <exception>
#include <iostream>
#include <geos/util.h>

int wrapped_GEOS_CALL(..)
{
  try
  {
    GEOS_CALL(..);
  }
  // catch unhandled GEOS exceptions
  catch (GEOSException& ge)
  {
     std::cout << "GEOS Exception :" <<  ge.toString() <<std::endl;
     return X;
  }

  // Catch other unhandled standard exceptions
  catch ( std::exception& e)
  {
    std::cout << "Standard Exception: " << e.what() << std::endl;
    return X;
  }
  return(0);
}

Note this does nothing to try to ensure that the underlying
machine is left in an appropriate state for further computing
which I believe the JAVA machine will try to do for you

HTH

Norman