Change Invisibles: migrating from Xy 3.55 to 4.017 (should be 4.0 18)

[Peter Brown pbrown@xxxxxxxx]

My earlier positive feedback on Carl's suggestions was a bit imprecise for
diagnostic purposes, because it was based on carrying out all of them:
1) changing BC to BX and XC (in fact, I used carriage return for readabilty,
too, Nathan :-) to Q2)
2) Adding BX es 1Q2 to the beginning of the routine
3) running the routine in expanded mode

In fact, as Carl suggested later on in the thread that it would be, (2) was
sufficient; turning on error suppression eliminated the delays in execution
of CI.

BTW, that was great good fortune for me. My routine also strips files of
their embedded footnotes, in which I send editorial notes to authors. (That,
by the way, has been a terrific way to communicate: I just load the footnote
command plus some text into a keyboard definition--BC fnXC EDITOR'S NOTE: .
With appropriate header in the file, that makes the files print with the
footnote queries at the bottoms of appropriate pages, leaving only the
footnote number in the text itself.)

My strip routine, which is probably about as inelegant as this list can
imagine, searches for "fn1" in expanded mode, then changes to normal mode
(CM), and finally does a backspace delete [BD] (it loops on those deletions
until all the footnotes are outta there). Trouble was, I wrote CM whenever I
wanted the program to switch modes (normal/expanded), and that meant that
running the routine in expanded mode would make my mode switching and
context sensitive deletions tricky to correct. The lesson, I guess, is to
avoid scroll lock (which embeds the CM function), and use XP and WG (for
expanded mode and normal mode, respectively).

Peter Brown
pbrown@xxxxxxxx

Carl Distefano writes:> Fortunately, the
solution is simple: insert
BX es 1Q2 at the top of the routine and the problem goes away. You don't
even
have to command BX es 0Q2 on EXit; Xy4 handles that automatically. But,
before I spin my wheels any further, let's wait for feedback from [Peter].

Nathan Savin writes:> By the way, when writing conversion and other problems
with a
series of CI searches, instead of XC I tend to use a newline.
That has the advantage of making the program more readable.