Note that I ran into this same warning recently as part of upgrading to boost 1.42. Some research led me to write the following comment in my local patch:
// kab, 4/2/2010: workaround for gcc warnings: // comparison is always true due to limited range of data type // gcc 4.3 and later provide warning controls for this: controled by // -W[no-]type-limits, defaulting to disabled and enabled by -Wextra. // gcc versions prior to that provide NO CONTROL over this warning, which // first appeared circa gcc 3.3.2. see gcc bug 12963.
Frankly I'm hard-pressed to ever think of a situation where I'd want to enable this warning, but unfortunately there's no way to disable it for a significant range of gcc versions.
For expedience I changed these asserts to use boost.numeric.conversion facilities to perform the range check, as it carefully optimizes out tests guaranteed to succeed, and so avoids triggering the warning in question. I'm guessing that a real patch that Robert would accept would avoid adding the dependency on boost.numeric.conversion, but that was good enough for my immediate purposes. Specifically, my patch introduces the following helper function:
template<class T, class S> bool save_override_check(S s) { return boost::numeric::cInRange == boost::numeric::converter<T, S>::out_of_range(s); }
which is used as
assert(save_override_check<boost::int_least16_t>(t.t));
Well, this discussion makes me want to re-consider - at my leasure - what kind of checking - if any - should be done. It touches upon other issues as well, such as my imposed requirement that all archives should act identical. Feel free to open a track item on this so that it will hang around forever or until it's specifically considerd - whichever comes first. Robert Ramey