Re: BC XC or BX Q2

[cld@xxxxxxxx (Carl Distefano)]

Eric Van Tassel writes:
≪ But some BX-Q2 routines do require [parens] ≫

What makes you say that? Examples, please!

I think I said pretty clearly yesterday that the only
command that requires parens is the esoteric SEarch for
func Q2 itself: BX (se Q2 ). If you write
BX se Q2 Q2
the XPL interpreter has no way of distinguishing between
the first Q2 as an inert string and the second as a
working func Q2. BX and Q2 are always interpreted as
matching pairs of "real" (working) functions. Once it
encounters func BX, the interpreter "waits" for the
next-occurring Q2 and executes the intervening string as
a command, just as if the command were typed on the
command line and executed with F9. Hence BX se Q2 Q2
will be intepreted as a SEarch command without a search
string, followed by a lone func Q2. The bare SEarch
command will produce an error.

Another way around this astronomically rare problem is to
use the Ascii-12 flag followed by the plain-text function
mnemonic "q2" as a proxy for a real func Q2 in the search
statement: BX se {12}q2Q2 . This avoids the problem of
the two Q2's, albeit less elegantly than do parens.

I repeat: Give me an example (other than the above) of a
command that fails with BX...Q2 but succeeds with
BX ().

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