Bjarne's comments about exception specification

Hello everyone,


How do you understand the Bjarne's comments about exception specification? Especially, "not required to be checked across compilation-unit" and "violations will not be caught at run time"?

section 14.6.1 Checking Exception Specifications

--------------------
Importantly, exception-specifications are not required to be checked exactly across compilation-unit boundaries. Naturally, an implementation could check. However, for many large and long-lived systems, it is important that the implementation does not -- or, if it does, than it carefully gives hard errors only where violations will not be caught at run time.
--------------------


thanks in advance,
George

Well I'm not sure exactly what he means, but I do know that exception specifications are completely useless and should be avoided whenever possible:
Chapter 75. Avoid exception specifications

It means that if exception specifications in two different compilation units disagree, that does not have to be a compile time error. It is however, undefined behavior.

Thanks cpjust,


I need to maintain some code which has exception specification, and I have to understand the concept. Looks like you are experienced at this topic, any ideas about my original question?

Quote:

---

Originally Posted by cpjust

Well I'm not sure exactly what he means, but I do know that exception specifications are completely useless and should be avoided whenever possible:
Chapter 75. Avoid exception specifications

---

regards,
George

Thanks King,


What means this part?

--------------------
However, for many large and long-lived systems, it is important that the implementation does not -- or, if it does, than it carefully gives hard errors only where violations will not be caught at run time.
--------------------

Quote:

---

Originally Posted by King Mir

It means that if exception specifications in two different compilation units disagree, that does not have to be a compile time error. It is however, undefined behavior.

---

regards,
George