From: Beman Dawes (bdawes_at_[hidden])
Date: 2004-01-04 08:30:50
At 06:19 PM 1/2/2004, Dan W. wrote:
>While talking about Invariants, DBC, assertions and exceptions in
>another thread, the subject came up of how useless assert() appears to
>be, in that it terminates the program.
>What I always do in my coding is to have a global function...
>
>void landing_strip()
>{
> nop();
> nop();
> nop();
>}
>
>... defining my own assert() that calls it upon evaluating to false, and
>placing a break in the middle nop(); (3 for breaking distance :)
>
>I thought of using a coded break within my assert, but opted not to
>because that might slow down execution in the debugger (having too many
>break addresses to compare at each step).
>
>In this way, instead of terminating, the code stops at the landing
>strip, and all I have to do is click on the previous stack level to get
>back to the caller. Works like a charm, and I bet I'm not the only one
>using this trick.
>
>Would it make sense to generalize this and offer it as an alternative to
>the standard assert()? Maybe call it "ensure()"?
How is what you are proposing different from BOOST_ASSERT?
See http://www.boost.org/libs/utility/assert.html
--Beman