On Fri, Jun 30, 2000 at 10:56:56PM -0400, Aidan Cully wrote:

 > On Fri, Jun 30, 2000 at 07:55:38AM -0700, Jason R Thorpe wrote:
 > > I think in the short-term (i.e. in time for 1.5), we should change
 > > Heimdal's behavior to match MIT's wrt. krb5_init_context().
 > 
 > That's a real easy patch...  If no one objects to it, the following
 > patch (untested) should do it, and I'll commit a tested one in two
 > weeks (to account for the 4th), and send a pullup afterwards.

Looks great.

 > 
 > --- context.c.orig	Fri Jun 30 22:49:11 2000
 > +++ context.c	Fri Jun 30 22:55:15 2000
 > @@ -179,11 +179,10 @@
 >  
 >      if (ret == 0)
 >  	p->cf = tmp_cf;
 > -#if 0
 > -    else
 > -	krb5_warnx (p, "Unable to parse config file %s.  Ignoring.",
 > -		    config_file); /* XXX */
 > -#endif
 > +    else {
 > +	krb5_free_context (p);
 > +	return ret;
 > +    }
 >  
 >      ret = init_context_from_config_file(p);
 >      if(ret)
 > 
 > --aidan

-- 
        -- Jason R. Thorpe <thorpej@zembu.com>