Re: some questions re. Ada/GNAT from a C++/GCC user

[bobduff@world.std.com (Robert A Duff)]

In article <3160B2F1.260D@mcs.com>, Mike Young  <mikey@mcs.com> wrote:
>This appears somewhat frightening to me, that I might catch exceptions 
>from unknown and unspecified source at any given moment. 

No, it's not that bad.  The parent and child are synchronized at a known
point.  In any case, it's no worse than the current situation, which is
that the exception is simply ignored, and the task silently disappears.
At some later time ("any given moment!") some other task will probably
try to communicate with the dead task, and this will cause *another*
exception, or else a deadlock, depending on exactly how the
communication is done.

>...What exactly 
>should I do with this exception aside from terminating?

Nothing.  You would terminate, normally.  And then you would debug the
problem.  My point is just that debugging is a lot easier if you can
pinpoint the problem more closely to its true source.

This is no different from *most* cases of raising an exception -- the
normal response is to terminate the program.  Then the programmer fixes
the bug.  In *some* applications it makes sense to handle all exceptions
and try to be fault tolerant, but that's a very difficult thing to do
properly, and usually is not done.  For most exceptions in most programs
you simply want to terminate the program.

- Bob