On Wed, Oct 17, 2018 at 01:59:39PM -0600, Todd C. Miller wrote:
> On Wed, 17 Oct 2018 21:44:42 +0200, Martijn van Duren wrote:
>
> > The code before this commit has no resemblance to the field count
> > complexity we're addressing now, so it seems highly unlikely it's legacy
> > code from that era.
>
> I believe that code was written by michaels@ to fix the bugs
> referenced in:
>
> http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=1356
>
> Not just the crash but also the incorrect output.
>
> Unfortunately, the 4.4BSD Lite changes were merged in along with
> the other, unrelated fixes.
>
> - todd
My proposed diff produces the right output for the test case
mentioned in the PR. I'll add it to the regress tests.
-Otto
