Braced initializers (again)

[Doug Gwyn]

In article <1741 at tkou02.enet.dec.com> diamond at tkou02.enet.dec.com (diamond at tkovoa) writes:
>In article <13002 at smoke.BRL.MIL> gwyn at smoke.BRL.MIL (Doug Gwyn) writes:
>>... interpretation of the standard; of course the
>>interpretation should be exactly that and not some wild contradiction
>>of what the standard actually says.
>Huh?  Some of the interpretations already wildly contradict what the
>standard actually says.  To be more precise, the standard says some
>things that are wildly different from what was intended, and the
>interpretations made a lot more sense.

Name one.  There has been NO information bulletin so far, and only six
official individual replies to interpretation requests, which I here
summarize (disclaimer: this posting is NOT an official document):

#1.  Still being discussed by committee; no resolution yet.

#2.  Was a request for changes, not for interpretation.

#3.  Was a request for changes, not for interpretation.

#4.  Was disambiguated by an editorial change in the final standard.

#5.  A strictly conforming program may not contain #pragma, because
as the standard explicitly states, a pragma causes the implementation
to behave in an implementation-defined manner.

#6.  strtoul() conversion of a string starting with a minus sign
reports a range error if and only if the rest of the subject sequence
would overflow the range of unsigned longs, otherwise the value is
negated as an unsigned long.  This is the only sensible way to
interpret what the standard states for this process.

So which of these "wildly contradict what the standard actually says"?