Change Invisible problems with BX

[Harry Binswanger hb@xxxxxxxx]

Here's a puzzler. I have a file on which I'm having difficulty using Change
Invisible to get rid of all the quote marks. But the difficulties arise
only when done in an XPL program and only when followed by a "sort" command
using BX instead of BC.
I'm running a little XPL program on a text file that is 27k long. (I also
tested it while slimming the file to 4.4k, see below).

First the one that works:

XPLeNCODE v2.0
b-gin [UNTITLED]
[TF_][BX_]ci /"//[Q2_][TF_][DF_][BF_][DF_][BC_]sort
[XC_][TF_]{<}EX{>}
-nd
XPLeNCODE


Now the one that eliminates only 38 of the files 40 quote marks:

XPLeNCODE v2.0
b-gin [UNTITLED]
[TF_][BX_]ci /"//[Q2_][TF_][DF_][BF_][DF_][BX_]sort
[Q2_][TF_]{<}EX{>}
-nd
XPLeNCODE

As you can see, they differ only in whether I use BX-Q2 or BC-XC.
The file to be sorted pairs names with email addresses, using tabs as
delimiters. It has the structure:

John Doe[tab][tab][cr-lf[
"Mary Roe"[tab][tab][cr-lf]

-and so on for about 700 lines.
I wanted to eliminate all quote marks from the file (currently there are 40
of them) so that the sort wouldn't put names in quote marks at the top of
the file. That's when I hit the problems.
I suspected memory problems with the long file, so I chopped it
progressively down until, at 4.4k, I got a different error using BX: the
Change Invisible changed:

 "Michael Wharton"

to:

Michael Whartonon"
I also tested it running two consecutive, identical Change Invisibles,
using BX-Q2 and that works fine, even on the 27k file.
I also tried using "BX (sort)." That worked fine on the 4.4k file, but on
the 27k file gave the same old "Michael Wharton" (with quote marks) after
the sort.

So I wonder if BX and BC have different memory requirements?

Harry Binswanger
hb@xxxxxxxx