[3557] in SIPB bug reports
missing /usr/sipb link
daemon@ATHENA.MIT.EDU (yandros@Athena.MIT.EDU)
Tue Mar 9 21:18:42 1993
From: yandros@Athena.MIT.EDU
Date: Tue, 9 Mar 93 21:18:10 -0500
To: jik@GZA.COM
Cc: srz@Athena.MIT.EDU, bug-sipb@Athena.MIT.EDU, sipb-staff@Athena.MIT.EDU,
sipb@Athena.MIT.EDU
In-Reply-To: "Jonathan I. Kamens"'s message of Tue, 9 Mar 93 14:47:07 -0500 <9303091947.AA16971@pad-thai.aktis.com>
(removed aix32 and added sipb, since I think this is something all of
sipb should discuss)
Well, after not hearing from Jon for a reasonably long period of time,
I log in today in the middle of two `flamewars' (both actually valid
discussions, in truth).
The reason Derek sent this to sipb-staff is because when I told him to
be *sure* and send this in, and he said `sipb-staff?', I wasn't
thinking and said `sure'. The reason I told him to be sure to bring
this up is because I'd heard it discussed (and flamed about) before,
but I didn't think there'd ever been a real resolution. The issues,
as I see them, are:
o some Athena platforms have a link from /usr/sipb to
/mit/sipb. As I understand it, this dates from way back
when sipb programs went in /usr/unsupported/sipb.
o some sipb programs use this link to refer to files, rather
than using /mit/sipb of an AFS path.
o The 7.5a release for aix32 did not have this link. This
was, in fact, an accident, and Richard tells me that he
intended to add it in and just forgot. (hurray for SIPB
test machines! :-/ ) This does not imply that the link will
continue to exist, especially given the renewed push towards
layered athena.
o There are, apparently, non-Athena people who use software
from the sipb locker. This is apparently made much easier
by (and sometimes requires) the link. (sites not wanting to
create /mit - are there a lot of these? The impresion I get
is that there's mainly one or two.)
o changing the software to use /mit/sipb instead of /usr/sipb
involves a non-trivial amount of work.
Personally, my take is this: we should use /mit/sipb, since it's the
abstraction we're guaranteed to keep. Jon and I can do the work, and
in the process show other people how we normally do things w.r.t the
sipb locker... (good for prospectives :-) Some Non-MIT sites will be
at a disadvantage because of this, but if Layered Athena does actually
happen, this trouble should be minimized. Of couse, this is a good
time to make sure the code is set up to easily change these things...
:-)
Comments?
