Rethinking... (Re: RFC: fp printing speedup...)

[Matt Austern]

On Nov 14, 2003, at 11:29 AM, Nathan Myers wrote:

> On Fri, Nov 14, 2003 at 07:59:10PM +0100, Gabriel Dos Reis wrote:
>> Nathan Myers <ncm-nospam@cantrip.org> writes:
>>
>> | Any difference in behavior between the single-character and array
>> | forms of facet members is a user error.  If it doesn't say so
>> | clearly enough in the standard, that's a Defect.  Enter a Defect
>> | Report on the subject.  Mark 12988 as "not a bug", and reference
>> | the DR.
>>
>> A PR should be marked as not a defect if there is no ground in the
>> standard to do so.  And putting a reference to a non existing DR is
>> at best ill-advised.
>
> It should be noted in the PR that the example program evokes undefined
> behavior, because its facet's two widen members have different 
> semantics.
> Correcting the example program makes the problem go away, 
> irretrievably.
> Thus, "not a bug".
>
> The Defect is that it takes a too-close reading of the standard to
> realize that the supplied example program evokes undefined behavior.
> The bug resolution doesn't depend on the fix for that defect.

For what it's worth, I'm with Nathan here.  There are an awful lot
of holes you can find by an extra-close reading of clause 22.  We
really don't want to go there.  Better to implement the natural
reading, and the one that will benefit most users.

			--Matt