On Mon, Aug 09, 1999 at 06:36:40PM -0700, Jason Thorpe wrote:
> On Sun, 8 Aug 1999 03:03:37 -0400 
>  Aidan Cully <aidan@kublai.com> wrote:
>  > Then CRYPTOBASE could be replaced with a list of crypto trees, in the
>  > order you'd like them applied, with a bit of foo in the src/lib/Makefile
>  > &c., like this (untested):
> 
> Um, first of all I'd like to point out that your use of CRYPTOPATH inside
> of the Kerberos 5 was was *completely wrong* right from the very beginning.
> CRYPTOPATH was never meant to be used in that way, and you weren't even
> setting all of the variables necessary for it to function!
> 
> When I did a "make build" *even without* KERBEROS5 set, the make machinery
> you used didn't work *at all*.  You obviously didn't test any of it.

The problem is that the machine where I have CVS access was different than
my -current box, and I have very little control over the first one.
CRYPTOPATH didn't exist when I was finishing most of my work on kerberos5,
and I never got a very clear understanding 'til late in the process about
how it was used, so my compatibility environment was broken.

I'm very sorry to have screwed up in this way..  All that I can say is that
I was (and am still) under pressure to finish this up.

> So, maybe you can understand if I think this change is a little ... suspect.
> 
> So, as I understand your use of .PARSEDIR, it's competely unnecessary.  Looks
> like .CURDIR will do what you want.  It's worth noting that .CURDIR did what
> you wanted in the stuff I had to fix, too.

No, since I was after the $CRYPTOBASES bit you quote.  I also thought it
would be useful for (something like) libtelnet/Makefile.inc to find the
dist area of the crypto-us tree.

I know CRYPTOBASES could be put together in a different way..  This
seemed like the cleanest way to do it to me.

> I also note that you seem to have already committed it.  I guess you didn't
> even bother to look at all of the Kerberos5 include file machinery that
> I fixed (without the need for .PARSEDIR).

I did see what you fixed, and you'll note everything krb5 looks correct,
now, and I hope you see that I didn't intend to use .PARSEDIR where
.CURDIR could be used.

> I'll also note that you suggested this change on Sunday (when most of us,
> I would gather, are trying to get AWAY from their computer keyboards),
> and apparently committed it less than 24 hours later.
> 
> I'd suggest that you make it work without .PARSEDIR, and then back out the
> change to make(1).

If I haven't convinced you of its value with this message, I'll back it
out..  The worst it can be called is "useless", so I don't think there's
harm in it being in the tree until I hear from you.

--aidan