Re: [Boost-testing] [jrp] Perplexed by VC-9 failures.

From: John Maddock (john_at_[hidden])
Date: 2007-11-20 06:27:08

John Pavel wrote:
> I'll use the release version on a real machine, which should obviate
> the need for extending timeouts.

Oh OK, thanks!

> Yes, the extra warnings are the result of invoking code analysis.
> Are any of the warnings about the core files legitimate?

I don't believe so: as far as I can tell it emits a warning whenever you use
sizeof with the result of an expression: a trick that is absolutely
ubiquitous in the type_traits library (and quite deliberate).  If I could
reproduce it here then I'd throw in some #pragma's to shut the compiler up:
otherwise folks are going to be very annoyed by this :-(

If you can let me have a full log of a rebuild of the type-traits tests with
code analysis turned on, I'll look through and see what can be fixed and
what suppressed with #pragma's.

Many thanks, John.