🚀 go-pugleaf

In article <1992Apr5.001734.26988@taumet.com> steve@taumet.com (Steve Clamage) writes:
>atbowler@thinkage.on.ca (Alan Bowler) writes:
>
>>In article <1992Apr1.175309.19833@taumet.com> steve@taumet.com (Steve Clamage) writes:
>>>    long *ip = &y.rem; /* legal but disallowed by this implementation */
>>the implementation might consider it legal.
>
>In other words, a compiler which claims conformance to the Standard
>would allow
>	int i;
>	long *ip = &i;	/* type conflict */
>I think this is much worse than making div_t and ldiv_t the same type.
>
Ooops. You are right.  The compiler is supposed to complain about this
type of assignment.

>>Even longer delays often occur with the site system administration not
>>bothering to install newer releases.  I would bet that in this case,
>>the version of the implementation simply predates the standard, and
>>newer release fix some of the problems outlined.
>
>This is certainly possible.  We would have to hear from the original
>poster (sorry, forgot who it was) whether it is the case.

Another message said that there were 1990 and 1991 dates in the
headers.  This is certainly very late to have the sort of errors
described in the original post.  I wrote my message shortly after
finding that the delivery pipeline for some of our customers is just
about to deliver the 5 year old version of some of our software.
Fortunately most customers have received newer versions.