vix@ubvax.UUCP (05/29/87)
This is a neat package, one I have used before on my Symmetric and on
various CFC machines at Berkeley. I am very reluctant to install it
here (Ungermann-Bass) though. I've applied every patch from ucb-fixes;
I think of these as "official", and I know that the patches are all
serialized -- so they always install correctly.
As much as I'd enjoy having a working control-T, I don't want to have to
back out Chris' patch whenever I need to install a bsd-fixes patch --
which I would have to do, some time down the road, when UCB patches a
section of code that Chris' patch has altered.
So, what can I do? Who do I see about getting UCB to officiate this
patch, so I can use it? Or, alternatively, what am I missing here?
--
Paul Vixie
Consultant (408) 562-7985
Ungermann-Bass {pyramid,cae780,amd,decvax,weitek,sco}!ubvax!vix
Santa Clara, CA vix@ubvax.ub.com (uucp) vix@Engr.UB.Com (CSNET)chris@mimsy.UUCP (Chris Torek) (05/30/87)
In article <1293@ubvax.UUCP> vix@ubvax.UUCP (Paul Vixie) writes: >... As much as I'd enjoy having a working control-T, I don't want to have to >back out Chris' patch whenever I need to install a bsd-fixes patch -- >which I would have to do, some time down the road, when UCB patches a >section of code that Chris' patch has altered. We---by which I mean the University of MD CS Department---keep all our kernel sources in RCS, so that we can retrieve Berkeley revisions so as to see just what was changed, and properly merge any conflicting changes. This is time-consuming but not difficult. (In other words, it is a royal pain, but when ya just gotta have a feature. . . .) No one at Berkeley really wants to uglify the tty driver yet more. Someone will probably rewrite the whole thing sometime soon. Maybe (if someone can jar certain AT&T stigmata loose from SysV licencing, or if someone reimplments them from existing descriptions) we can use streams and do it right instead. Ah well. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690) Domain: chris@mimsy.umd.edu Path: seismo!mimsy!chris
dpz@aramis.rutgers.edu (David P. Zimmerman) (05/31/87)
Paul, For a background, we have had the ^T code installed here for ages now, along with zillions of other hacks, additions, and bug fixes to 4.2, local and foreign. You seem to put a high importance on purity of the code and need of additions being officiated by UCB, but I don't see the necessity of this. "Official" and "serialized" don't always mean that they will install correctly. That only comes from professional work. "Serialized" just keeps them indexed, and "official" means nothing. Why would you have to back out of an external patch when you have to install a UCB correction to the same region? Just look at the code for each and figure out what the correct thing is. Yes, it might generate some conflict, but if you keep .ORIG files and a changes history around, that situation is, from experience, fixable without too much trouble. dpz -- David P. Zimmerman rutgers!dpz dpz@rutgers.edu