Re: [boost] Looking for thoughts on a new smart pointer: shared_ptr_nonnull

On 08-10-2013 21:22, Nevin Liber wrote:

On 8 October 2013 13:17, Thorsten Ottosen wrote:

...

We are indeed going in circles. I can respect that some people don't want the overhead of the runtime check. I can also respect that people don't want to take down the whole server because of a bug in some subsystem.

Except you can't know if it is a bug in a subsystem or somewhere else. If you know the bug is from the subsystem, why did you put the bug in? And if you don't know the root cause of the bug, how can you possibly know the effect? All you've detected is a symptom. The rest of the "analysis" is nothing more than wishful thinking.

There are many types of bugs, some are severe and some are less severe. It's not exactly hard to put a catch around a call into a sub-system, in which case you know for a fact that the exception exposes itself in this sub-system. Could the bug be because of some other problem, working in conjunction with the sub-system to induce the bug? Sure. Does it matter? Nope. The bug is exposed in that particular subsystem. I don't see how you can pretend what the right behavior is to such an error is in other people's software. Killing a server application used by hundreds of people is just not an option for some people. Talk about wishful thinking. You seem to be completely obsessed by the point that a precondition or invariant violation must always be checked by compiled-away assertions. Your argument is that there is a bug in the program and that you should terminate anyway. Sadly, must programs have bugs, but we use them anyway. In practice it matters for some to avoid UB. It's a major benefit of using Java or C#. -Thorsten