Re: Sig+ xpl [was Re: Porting DLG funcs $0-$9|$A-$Z to U2]

[cld@xxxxxxxx (Carl Distefano)]

Reply to note from  Wed, 29 Aug 2001 12:44:27 -
0400 (EDT)

> Huh? Weren't we just discussing that AS BX ed Q2 fails?
> That's alone makes BX unreliable. You seemed to assume that
> anyone would know to throw in a {va$dr} with AS BX. Where is
> that written? What other undocumented work-arounds does BX
> require in what other situations? Or have BX work-arounds
> become so habitual that you don't think of them as such?

If you insist on a rigid congruence between BX and BC, the 
technique I offered the other day is a "work-around". If you see BX
and BC as complements, it becomes the preferred way to CAll a file
from a directory in v4+ XPL. (The technique itself is, of course,
straight-forward, not the arcanum you imply.) And if this
difference between BX and BC is undocumented -- so what?
Innumerable tips and techniques in XPL of any vintage are
undocumented except in the form of published code and discussions in
this forum. Undocumented doesn't equal unreliable.

Other differences between BX and BC -- notably in the way they
handle SEarch strings -- are alluded to in "Signature: Making the
Transition" and have been the subject of discussion here. Some of
these represent palpable advantages of BX over BC, e.g., the freedom
to put strings literally without having to grapple with the 5-to-3-
byte and 3-to-1-byte transformations that apply to the CMline.
Others are, simply, differences. Overall, the similarities far
outnumber the differences. And the differences in no way implicate
the reliability of BX.

The coding travails you describe -- I looked at your code -- are
wholly attributable to the for-Windows environments you chose to
work in. They have nothing to do with BX or v4+ XPL per se.

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