ssh coredump if gssapi-with-mic preferred, but no kinit done (yet)?

[Doug Rabson]

On 2 Jul 2007, at 17:13, David Wolfskill wrote:

> On Mon, Jul 02, 2007 at 04:01:56PM +0100, Doug Rabson wrote:
>> ...
>> It looks like I did break it. I should be able to repro and fix this
>> as soon as I put my FreeBSD box back together (probably tomorrow). In
>> the meantime, you can avoid this by rolling src/lib/libgssapi back to
>> about last Friday or so.
>
> OK; I tried precisely that, and am now able to use ssh to get from the
> desktop to the other machine without first needing to run kinit(1).
>
> Here's what I did:
>
> * Re-boot from the CURRENT slice.
>
> * cd /usr/src/lib/libgssapi
>
> * cvs update -D "last Friday"
>   The following files were updated:
>     catmint(6.2-S)[2] cvs update -D "last Friday"
>     cvs update: Updating .
>     U gss_accept_sec_context.c
>     U gss_display_status.c
>     U gss_export_name.c
>     U gss_indicate_mechs.c
>     U gss_init_sec_context.c
>     catmint(6.2-S)[3] ls -latr
>
> * make clean
>
> * make
>
> * make install
>
> * Login to the desktop (catmint) from a different window.
>
> * Verify that it works:
>
>     catmint(7.0-C)[1] ssh repo uname -a
>     FreeBSD repo.mail-abuse.org 6.2-STABLE FreeBSD 6.2-STABLE #61:  
> Sun Jul  1 06:28:55 PDT 2007     dhw at repo.mail-abuse.org:/usr/obj/ 
> usr/src/sys/SMP_PAE  i386
>     catmint(7.0-C)[2]
>
> Would additional poking around be of use?
>
> I did review the typescript from the "make buildworld" yesterday  
> for the
> libgssapi files, and did not see any warnings issued.
>
> I can, of course, make the typescripts (both from "make buildworld"  
> and
> the reversion) available to you or others; I don't see a pressing need
> to spam the list with them (and the "make buildworld" would be way too
> big anyhow).

I'm sure I will be able to reproduce it when I get my machine running  
again. Thanks for verifying that this commit caused the regression.