>Number:         10758
>Category:       bin
>Synopsis:       new kadmind build fails on sparc
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Aug 04 18:03:00 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     matthew green
>Release:        <NetBSD-current source date>today
>Organization:
people's front against (bozotic) www (softwar foundation)
>Environment:
	
System: NetBSD towers-of-dub.eterna.com.au 1.4Z NetBSD 1.4Z (_towers_) #43: Fri Jun 2 23:31:34 EST 2000 mrg@towers-of-dub.eterna.com.au:/orb/y/_towers_ alpha


>Description:

building today's -current on a sparc gives:

/orb/i/src/libexec/kadmind/../../crypto/dist/heimdal/kadmin/kadmind.c: In function `main':
/orb/i/src/libexec/kadmind/../../crypto/dist/heimdal/kadmin/kadmind.c:149: warning: passing arg 3 of `getsockname' from incompatible pointer type

>How-To-Repeat:

build current on a sparc.

>Fix:

the issue here is that getsockname() takes a socklen_t * as it's last (3rd)
argument under NetBSD (and, indeed, all recent UNIX systems), however the
heimdal code is using size_t * as the argument here.  this is (now) wrong but
there are older systems (eg, solaris 2.6) that define getsockname() to take
a size_t * as the last argument.  i don't know what the best portable fix for
this issue is so i leave it upto someone else.  if nothing else, netbsd should
include a local patch that uses socklen_t *.
>Release-Note:
>Audit-Trail:
>Unformatted: