ronald@robobar.Co.Uk (Ronald S H Khoo) (01/02/91)
[ I've crossposted to USENET because I'm hoping someone has an answer to a question regarding the SVR3 vi in general ] In article <9101012058.AA04469@jpr.com> on the SCO mailing list jpr@jpr.com wrote: > This is a mapping which I used daily in SCO Xenix to compose replies to > CompuServe messages. > > map #1 ?#: [0-9][0-9].* S[0-9][0-9]*/?Wyt N<<Ore^[pmao/post unf^[mbO > > It fails in SCO Unix. Does anyone know what they've changed in vi, besides > using terminfo instead of termcap? Regular expressions in SCO Unix vi seem to be severely broken, I *think* it's closure of a character class that does it. "ver" on SCO Unix says SVR3.1, does anyone know if the base port for that has such problems? Anyway, personally, I gave up with the SCO Unix vi and use the SCO Xenix vi instead. Doing that does pose problems, like % substitution doesn't work from the :! mode (top bit gets set), but that's a lot less hassle than broken regular expressions. The Xenix vi is directly descended from the BSD (no AT&T parentage other than the original ed code) "ver" says 3.7 6/10/83. Putting the date in is so much more informative than just "SVR3.1". $ cat /dev/icbm > AT&T. -- ronald@robobar.co.uk +44 81 991 1142 (O) +44 71 229 7741 (H)
jpr@jpradley.jpr.com (Jean-Pierre Radley) (01/02/91)
In article <9101020746.AA08202@robobar.Co.Uk> ronald@robobar.Co.Uk (Ronald S H Khoo) writes: >[ I've crossposted to USENET because I'm hoping someone has an answer to > a question regarding the SVR3 vi in general ] I also posted me original question here, it just didn't propagate to you as quickly as the mailing list. >In article <9101012058.AA04469@jpr.com> on the SCO mailing list >jpr@jpr.com wrote: >> This is a mapping which I used daily in SCO Xenix to compose replies to >> CompuServe messages. >> >> map #1 ?#: [0-9][0-9].* S[0-9][0-9]*/?Wyt N<<Ore^[pmao/post unf^[mbO >> >> It fails in SCO Unix. Does anyone know what they've changed in vi, besides >> using terminfo instead of termcap? >Regular expressions in SCO Unix vi seem to be severely broken, I *think* >it's closure of a character class that does it. "ver" on SCO Unix says >SVR3.1, does anyone know if the base port for that has such problems? The mailing list has this response: >From: Eli Liang <uupsi!sco.COM!elil> >Date: Wed, 2 Jan 91 2:00:15 PST >Message-Id: <9101020200.aa12864@scoville.sco.COM> >This is the result of a small bug that was introduced into vi when it was >internationalized. I fixed it for SCO UNIX 3.2v2. You may want to consider >upgrading your OS if this is a problem for you. Sounds good but it seems to me I already have that version. Running "uname -v -r" shows: 3.2 2 And running "what /usr/bin/vi" shows: /usr/bin/vi: printf.c:2.2 6/5/79 SCO UNIX 3.2V2 OS 09 Jun 90 Is there a later version of vi? >Anyway, personally, I gave up with the SCO Unix vi and use the SCO >Xenix vi instead. Doing that does pose problems, like % substitution >doesn't work from the :! mode (top bit gets set), but that's a lot >less hassle than broken regular expressions. That's what I'm doing too. But that phenomenon is not consistent. As user "jpr", it happens when I'm a given directory but not when "root" tries vi in that directory. It doesn't happen all the time. Jean-Pierre Radley NYC Public Unix jpr@jpr.com CIS: 72160,1341
allbery@NCoast.ORG (Brandon S. Allbery KB8JRR) (01/04/91)
As quoted from <9101020746.AA08202@robobar.Co.Uk> by ronald@robobar.Co.Uk (Ronald S H Khoo): +--------------- | > map #1 ?#: [0-9][0-9].* S[0-9][0-9]*/?Wyt N<<Ore^[pmao/post unf^[mbO | > | > It fails in SCO Unix. Does anyone know what they've changed in vi, besides | > using terminfo instead of termcap? | | Regular expressions in SCO Unix vi seem to be severely broken, I *think* | it's closure of a character class that does it. "ver" on SCO Unix says | SVR3.1, does anyone know if the base port for that has such problems? | | $ cat /dev/icbm > AT&T. +--------------- I beg your pardon? The SVR3.1 at work has "vi", and it does *not* have any regexp bugs --- I've had to use it often enough that I would most definitely have crashed into them by now. (I prefer Emacs, but I can't install it on every machine I work on --- client sites, for example.) In any case, don't be so quick to nail AT&T to the cross for something that showed up in an SCO product. It might have been AT&T, or SCO, or SCO might have gotten it from someone else (did Interactive have anything to do with the initial 386 SVR3.1 port?). I *do* have to wonder why a "3.2" Unix comes with a 3.1 vi, though.... ++Brandon -- Me: Brandon S. Allbery VHF/UHF: KB8JRR on 220, 2m, 440 Internet: allbery@NCoast.ORG Packet: KB8JRR @ WA8BXN America OnLine: KB8JRR AMPR: KB8JRR.AmPR.ORG [44.70.4.88] uunet!usenet.ins.cwru.edu!ncoast!allbery Delphi: ALLBERY