[RFC] First 3.5 project: don't assign _S_terminal in string::c_str()

[Paolo Carlini]

Nathan Myers wrote:

> I say go for it. c_str() is used frequently enough that would
> certainly hear about it quickly if it got broken; and if it did,
> that would reveal latent bugs like the one in _M_clone. As it
> is, the assignment only interferes with us finding those bugs.

Agreed! The complete set of changes for mainline (would be 3.5) are
already in.

> Is 3.4 frozen and branched, now? I missed the announcement.
> Seems to me the _M_clone bug should be fixed in 3.4, too.

Ok, the _M_clone fix will go shortly into 3_4 too.

Thanks!
Paolo.