Re: [boost] Boost Policy on compiler warnings?

From: Douglas Gregor (gregod_at_[hidden])
Date: 2002-08-06 21:21:28

On Tuesday 06 August 2002 08:45 pm, Beman Dawes wrote:
>  >Warnings create, in essence, language variations. Today they're one of
>
> the
>
>  >biggest obstacles to portability.
>
> Sigh...  I'm afraid you're right.
>
> That would argue for either not reporting warnings at all, or only
> reporting warnings that are likely to be common to many platforms.
>
> I'd like to hear from others.  Or maybe nobody cares about warnings and we
> should forget the whole idea.
>
> --Beman

I don't entirely agree with Pete's assertion here. Good warnings create a 
subset of the language that excludes dumb behavior -- uninitialized 
variables, forgetting to return a value along some execution path, missing an 
enum value in a case statement, etc. -- and I think we should report and fix 
those warnings.

Bad warnings are, I think, what Pete is really referring to. These are 
warnings that are just plain wrong -- they complain about legal syntax, or 
whine about uninitialized variables that have been default constructed. Look 
at the warnings generated when compiling function_n_test.cpp in the win32 
regression logs for a particularly heinous example of Bad warnings. We should 
ignore Bad warnings. They should never show up and should be reported as bugs 
to the compiler vendor. If we can selectively turn off particular warnings 
for Boost code on the offending compilers, I think we should do so.

        Doug