jpr@dasys1.UUCP (Jean-Pierre Radley) (06/24/89)
Now that I upgraded from SCO 2.2.4 to 2.3GT, a great unhappiness descended upon me because of SCO'c compilation of HoneyDanBer uucp. My nodename (which uucp looks at) or the first line of my /etc/systemid (which uuname -l looks at), is, in either case, 'jpradley'; I am so registered in the uucp maps. But this version of uucp truncates my nodename to seven bytes. Trying to communicate with a site that knows me as 'jpradley' when uucico emits only 'jpradle': it doesn't work. I called the SCO hot-line for a patch to quash the truncation, and was told that there was no workaround, I'd just have to live with it. Stubborn as I am, here's the workaround which I found: In my Permissions file, _every_ entry now contains MYNAME=jpradley; and I added the catch-all line: MACHINE=OTHER MYNAME=jpradley. Thank goodness this is evaluated _after_ uucico calls uname() and truncates what resides in utsname.nodename. -- Jean-Pierre Radley CIS: 72160,1341 jpr@jpradley.UUCP
ked@garnet.berkeley.edu (Earl H. Kinmonth) (06/25/89)
In article <10034@dasys1.UUCP> jpr@jpradley.UUCP (Jean-Pierre Radley) writes: > >But this version of uucp truncates my nodename to seven bytes. If this bothers you, just think, someone was probably paid real money for such a bit of fecal-headed programming. And, you, and many others will pay real money and spend real time to work around this. In many of the world's languages, seven bytes is barely enough to get you past the first syllable. Even in English you have to have a really generic wasp name (jones, smith) to be fully identified by seven bytes! Senator Bill Proxmire from Wisconsin used to give out a "Golden Fleece" award for the silliest projects to win federal funding. Computing in general, not just SCO, could use an "Itchy Hemorrhoid" award, targeted at programming decisions of the "no one could ever want a _____ of more than _____ characters" variety. (For those concerned about the choice of metaphors here, the logic is that there are a (too) large number of programming decisions that seem to be made by people who have their cranium up their fundament. Such contortions deserve recognition.) (If you want a less loaded justification for criticism, here it is. At various times, the godfathers of UNIX have indicated that unless there are compelling hardware limitations, strings should be either zero, one, or infinite in length! Too bad they can't follow their own precepts!)
nick@aimed.UUCP (Nick Pemberton) (06/27/89)
In article <25743@agate.BERKELEY.EDU>, ked@garnet.berkeley.edu (Earl H. Kinmonth) writes: > In article <10034@dasys1.UUCP> jpr@jpradley.UUCP (Jean-Pierre Radley) writes: > > > > >But this version of uucp truncates my nodename to seven bytes. [discussion of name length deleted] Where exactly does the truncation occur? I know that our site, 'aimed', and the upstream site 'ncrcan' (which is an NCR tower) both run hdb, and both tend to do truncations in things like the filenames being passed back and forth. But at the actual opening handshake of uucp (^PShere...), the full name is present. Where else is it fouling up? Much curious...
michi@anvil.oz (Michael Henning) (06/30/89)
In article <25743@agate.BERKELEY.EDU>, ked@garnet.berkeley.edu (Earl H. Kinmonth) writes: > In article <10034@dasys1.UUCP> jpr@jpradley.UUCP (Jean-Pierre Radley) writes: > > >But this version of uucp truncates my nodename to seven bytes. > > Computing in > general, not just SCO, could use an "Itchy Hemorrhoid" award, targeted > at programming decisions of the "no one could ever want a _____ of more > than _____ characters" variety. > Agreed. Ever had a look at what happens to the output of "rwho" if a device name has more than seven characters ? The output gets totally scrambled. A good many packages (e.g. Multiview) use device names longer than seven characters, in which case you just have to live with the fact that rwho messes up column alignments and reports an incorrect number of users. But, you *do* have to consider the fact that memory is expensive, and that a good programmer will not simply go and foolishly waste all of seven bytes. After all, we only have a couple of million of them available, plus several million more of virtual memory... Michi. -- | Michael Henning | Internet : michi@anvil.oz{.au} | | Anvil Designs | JANET : michi%anvil.oz@uk.ac.ukc | | P.O. Box 954 | ARPA,Bitnet: michi%anvil.oz.au@uunet.uu.net | | Toowong 4066, Australia | UUCP : ...!uunet!munnari!anvil.oz!michi |