Re: XPL shortcuts etc. [long]

["..." adpf@xxxxxxxx]

≪	Now now Annie! No premature gloating! It's not just me; it's
Annie against EVERYBODY. ≫

So what else is new?

≪	But a lot of people tinkered with this problem, and somebody
always poked a hole in every solution. ≫

Right. It works in my system scores, maybe hundreds of times a day,
but what do *I* know.

≪	I can see at least one blatant situation where your solution
bombs (and this is just looking at the first third of it, i.e. S/G 101 in
text file FURTHER), namely on an open but empty text window: cursor ends
up on CMline no matter where it was originally located. ≫

Uh, you gotta be kidding, or I gotta be missing something. Like the
original location matters if the window's empty?

≪	(What *are* those 10 other files comprising 9951 bytes in
FURTHER.ZIP, anyway? No, wait, don't answer that ... )

Context.

≪4. The original leaves pfun XX on the CMline; my version leaves the arg
going in, with the symbol at the CMline cursor location--not necessarily
the last char in the arg. ≫

≪	So does mine. LDPM it. No change whatsoever to the pre-existing
CMline. Inserts the true func where the cursor was, within pre-existing
CMline string. Or puts the func in Text window, if that's where the
cursor was on entry. Am I misunderstanding you? ≫

A lot of these darlings are floating around--
	your terse v4 original,
	my literal retroport (makes all CMline conversions 255/128/nnn),
	my wordy revise,
	your true v3 version, and
	my florid version (the one I actually use).
I was comparing my wordy version to your original. Your true v3 version
does indeed do that, although neither of yours--not even your v4
original--restores the original TI state and mine both perform the v3
Big 3. (I didn't bother making the arg meaningful to the function in my
first version because you hadn't; both our second versions consider it).

As for @D, WL, etc., some combinations that involve xyWrite control
chars and trailing nulls require exception handling, but I don't consider
that a workaround. $A and WC are another matter, flat-out broken.
My florid version handles all control chars and in the interest of
comprehensiveness does a yucky workaround for $A and WC, a silly
conceit: $A is unusable and at best WC is a mystery func.

≪	Why do you have functions assigned to the alphanumeric keys in
your lower case TABLE? The accepted wisdom is that NI is only necessary
if the Alt shifting key is part of a TABLE= definition, and then
generally only on the number pad plus a few of the extended keyboard keys
>84 (because Alt otherwise bleeds through the Xywalls to DOS). ≫

Untrapped input is risky in public xpl; you never know what xyWriters
have done to their .kbds. Why doesn't much matter, but since you ask: Not
everybody uses a conventional physical keyboard. I don't. I have good
reason for using the key designated Pause/Break as 8; without NI it
misbehaves. NO and AR may or may not be superfluous to function character
input. You never know. In my case NI decidedly is not.

≪	Don't tell me you've upgraded to Warp v4 in order to control
XyWrite with OS/2 voice recognition? Wow! That's how I'm dictating this
note! Ain't it neat? ≫

RSI? 				--a

============================= adpFisher  nyc