[24386] in Kerberos
Re: Problem building an API program against 1.4.1
daemon@ATHENA.MIT.EDU (Ken Hornstein)
Mon Aug 1 16:49:56 2005
Message-Id: <200508012048.j71KmxRr004202@ginger.cmf.nrl.navy.mil>
To: kerberos@mit.edu
In-Reply-To: <87zms1flxb.fsf@windlord.stanford.edu>
Date: Mon, 01 Aug 2005 16:49:00 -0400
From: Ken Hornstein <kenh@cmf.nrl.navy.mil>
Errors-To: kerberos-bounces@mit.edu
>> Hm. I think we have a problem here. Certainly MIT's krb5-config emits
>> -R options (on the appropriate platform), and the autoconf glue that
>> I've written assumes that it does. I was under the impression that gcc
>> (at least on Solaris) passes -R options to ld unmolested (I just tested
>> that, and that's true on Solaris).
>
>gcc attempts to do the right thing with -R on platforms that support it,
>but may not recognize the flag on platforms that don't use -R or have some
>other shared library path mechanism.
Presumably krb5-config won't output -R on platforms if the compiler
doesn't understand it, right? I'm just saying that I've never noticed
this to be a problem in practice (although judging by what I've seen
in people's krb5-config scripts, I appear to be the ONLY one who actually
uses them for their intended purpose :-/).
--Ken
________________________________________________
Kerberos mailing list Kerberos@mit.edu
https://mailman.mit.edu/mailman/listinfo/kerberos
