C++ PATCH: Disallow floating-point literals in integral-constant expressions

[Gabriel Dos Reis]

Mike Stump <mrs@apple.com> writes:

| On Jan 31, 2005, at 6:17 PM, Marcin Dalecki wrote:
| > But since you started this kind of "decision based on user polling"
| > argument, I just wanted to point you at the logical flaw behind it:
| > In C++ average programmers frequently use "features of the compiler"
| > without intention of doing so. You can't imagine how many people out
| > there just edit the code until the compiler let's it pass...
| 
| Yes, that is exactly my point.  In my world, users like to be able to
| get warnings for a release, while they fix things.

I don't believe it is reasonable to require that we emit diagnostics for
every single bug we fix.  

Removing a documented feature is different, and we do differently.

-- Gaby