Silently passing wrong input for good one

[Paolo Carlini]

Martin Sebor wrote:

>> .... at variance
>> with EXAMPLE 3 that interestingly you mention, *all* the implementions
>> that you mentioned, *succesfully* extract -3.0,
>
> Which is incorrect. How funny!

An update: in the meanwhile I filed a glibc PR (libc/1765) about this
issue: if glibc fixes scanf and doesn't force us to inconsistent
behavior wrt to our reference libc implementation, I'm willing to change
libstc++-v3 to the stricter behavior even before working together in Berlin.

Paolo.

P.S. For your curiosity, Harbison & Steele, toward the end of page 383,
is also wrong. But in the updated Errata they added they correction,
which I had indicated a few months ago without receiving any feedback... ;)

    http://www.careferencemanual.com/errata.htm