Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!cs.utexas.edu!uunet!auspex!hitz
From: hitz@auspex.auspex.com (Dave Hitz)
Newsgroups: comp.editors
Subject: Re: Vi Turing machine question
Keywords: vi
Message-ID: <2286@auspex.auspex.com>
Date: 25 Jul 89 18:14:07 GMT
References: <1570@agora.UUCP>
Reply-To: hitz@auspex.auspex.com (Dave Hitz)
Organization: Auspex Systems, Santa Clara
Lines: 40

In article <1570@agora.UUCP> rickc@agora.UUCP (Rick Coates) writes:
>
>
>First, I thought that the Turing machine simulation in vi is clever -
>a real tour de force.
>
>When I tried it on MKS vi (I would give you a version number if their
>documentation gave one), however, it bombed.  Upon investigation,
>it is the 
>
>map k ^V|D
>
>(where ^V is the control V character)
>line that is causing trouble.  Then, I tried it from inside unix vi (instead
>of :so).  No luck.  Even with 'remap' set, unix vi gives the message I
>got from MKS - 'missing rhs'.
>
>What gives here?
>...!tektronix!reed!percival!agora!rickc

What the "k" map does is clear the current line.  The '|' key with no
arguments moves to the beginning of the current line, and 'D' clears
the rest of the line.

However, in ex mode, the '|' character is a command seperator, so to
get it into the map itself, I quote it with ^V.  Since MKS vi is
complaining that there is no right hand side of the macro, it sounds
like it didn't handle the quoting correctly, and interpreted the '|' as
the end of the map.

The TM macros are very senstive to the exact behavior of vi.  In fact,
they rely on a number of bugs in vi.  I've been amazed that they run on
all UNIX derived versions of vi that I've used, but not surprised that
a new implementation of vi would break with them.

Why don't you send MKS the macros and tell them their vi just isn't up
to snuff if it can't run them?
-- 
Dave Hitz					home: 408-739-7116
UUCP: {uunet,mips,sun,bridge2}!auspex!hitz 	play: 408-970-8970