Re: Wildcard behavior different in SE and CI

[Harry Binswanger hb@xxxxxxxx]

Carl,
I'm trying to change places in my XyBasic source code where I have things
like this:

'%--_XYENCode_start--
"'JM2.hide'.01-06'Q2"'^"'JM2.hb'Q2"'^"'JM2.unhide'Q2"'^
'%--_XYENCode___end--

Which is (in pseudo-code):

"[JM]2.hide:01-06[Q2]"
"[JM]2.hb[Q2]"
"[JM]2.unhide[Q2]
I want to totally remove the lines that call hide and unhide, leaving just
the JM 2.hb
The reason is that I'm automating, in XyBasic, surrounding any call to a U2
frame with:
[JM]2.HIDE:01-49,51-99[Q2]
.....
[JM]2.UNHIDE[Q2]

BTW, I wonder: don't some U2 frames uses s/g 51 also?

Thanks,
Harry
Reply to note from Harry Binswanger  Tue, 24 Mar
2009 18:07:36 -0400

> I found that this would find a "hit"
> ...
> Any thoughts?

Hard to tell without seeing the "hit" -- the string that the SEarch
command finds but the CI command doesn't. Why don't you post it here
(encode it)?

I suspect the problem is that your search string, "2.HIDE", is
always preceded by a 3-byte character (func JM), and 3-byters are
not reliably flagged by wildcard "[X]". It's a bug, a limitation --
call it what you will -- of the wildcard regime.

You might get better results if instead of:

'%--_XYENCode_start--
CI_/";3;X2.HIDE;W;~[/xxx;~[/
'%--_XYENCode___end--

you did something like:

'%--_XYENCode_start--
CI_/";2;X'JM2.HIDE;W;~[/xxx;~[/
'%--_XYENCode___end--

In other words, hard-code the func JM instead of using wildcard ;X
as a proxy for it.

Again, hard to say for sure without seeing the text you're searching
or trying to replace, but I bet there's a workaround.

--
Carl Distefano
cld@xxxxxxxx


Harry Binswanger
hb@xxxxxxxx