B
Bob Altman
Hi all,
We have some native C++ code that (inadvertently) performs a divide-by-zero
operation. This code is baked into two different executables, both built by
VS 2005. In one case, the divide-by-zero produces a result of positive
infinity (the debugger displays the value of the variable on the left side
of the equal sign as "#IND00"). In the other case, the divide-by-zero
results in a runtime exception.
We've looked through the project properties for some setting that might turn
divide-by-zero exceptions into "infinity" results, but the closest thing we
can find is /fp:except (which is disabled in both projects, and which would
only have the effect - if I understand it correctly - of deferring the
divide-by-zero exception until the next floating point operation).
Can someone give me a clue as to why the two projects behave differently?
TIA - Bob
We have some native C++ code that (inadvertently) performs a divide-by-zero
operation. This code is baked into two different executables, both built by
VS 2005. In one case, the divide-by-zero produces a result of positive
infinity (the debugger displays the value of the variable on the left side
of the equal sign as "#IND00"). In the other case, the divide-by-zero
results in a runtime exception.
We've looked through the project properties for some setting that might turn
divide-by-zero exceptions into "infinity" results, but the closest thing we
can find is /fp:except (which is disabled in both projects, and which would
only have the effect - if I understand it correctly - of deferring the
divide-by-zero exception until the next floating point operation).
Can someone give me a clue as to why the two projects behave differently?
TIA - Bob