Re: Wildcard error message

[cld@xxxxxxxx (Carl Distefano)]

Reply to note from Harry Binswanger  Tue, 21 Jan
2003 17:46:41 -0500

> Aha. Just tested this on a very short file (400 bytes) and got
> the error [Error 567 - Wildcards must be in the same order on both
> sides of a change]
>
> ci /W,/W;/
> ...
> It's not an academic discussion. It would simplify my
> programming if I could get this solved.

Get what solved, Harry? You try a command; if it doesn't work, you
try a different command, until you find one that does work. That's
the way it is in programming. Error 567 is particularly delphic and
arbitrary. For example, as you point out,

XPLeNCODE v2.0
b-gin [UNTITLED]
ci / [wW],/ [wW];/ works;[cr|lf][cr|lf]but its prolix equival
ent[cr|lf][cr|lf]ci / [w8][w0][wX],/ [w8][w0][wX];/[cr|lf]
-nd
XPLeNCODE

using three wildcards "in the same order" on each side, triggers
567. Solution: use the one that works (usually the simpler one).

In the case of ci /W,/W;/ (which fails), given that the W wildcard
means "any string from 0 to 80 chars", how is that different from a
simple CI /,/;/? I don't see any difference.

In my experience, you can almost always recast an Error 567 CHange
command so as to avoid the error. Perhaps a real-world problem of
this kind -- if you have one handy -- could serve as a further
example.

--
Carl Distefano
cld@xxxxxxxx
http://users.datarealm.com/xywwweb/