Re: coreutils readutmp applies strchr to a non-string

[Paul Eggert]

Jim Meyering <address@hidden> writes:

> did you find a system on which the existing code misbehaved?

Not personally, but I vaguely recall hearing of one.

I just Googled for this and found James Kanze's 1997 reports here:

http://groups.google.com/groups?selm=m3afpggtwo.fsf%40gabi-soft.fr
http://groups.google.com/groups?selm=m3iv41k6gh.fsf%40gabi-soft.fr

That ancient (and unnamed) host implemented strchr (s, c) as the
equivalent of memchr (s, c, strlen (s) + 1), which is allowed by the
standard.