Re: uniqcomp.pm - Working for others?

["Robert Holmgren" holmgren@xxxxxxxx]

** Reply to message from "Patricia M. Godfrey"  on Fri, 11
Nov 2005 14:02:01 -0500

Patricia, Manuel:

Patricia, I decoded your possessive/genitive test, and it UNIQOMPed perfectly.
No endless loop, no problems.

You're all focusing on the test files that you're using, and the results. It
has nothing to do with the test files OR the results, except insofar as they're
wrong; the problem lies in errors in the decode of the program itself -- or,
possibly, something fundamental in most people's XyWrite setup (a default) that
is different than mine, and that I'm overlooking, and that affects how funcs
FD|FM work. For example, nobody has mentioned what view mode they're using
(eXPanded, Draft, what? I *always* start testing *everything* in eXPanded --
although in my tests of UNIQOMP, it makes zero difference). But I suspect that
the decode of the program that you're using is flawed, because there has been
concrete evidence of that, primarily from Manuel and Russ, who sent me
post-DECODE copies of the program that had errors & extra characters which came
from heaven-knows-where. Further, Manuel made a useful observation:
------------
If I copy and paste the code from XySearch, DECODE works OK but, after
decoding, "-ndXPLeNCODE" remains at the end of the program.
If I save the HTML XySearch page as a TXT file and I run DECODE, no
problem at all.
------------
That's provocative & interesting information, but I need analysis. I can't fix
this until I know *exactly* what the differences he mentions are -- not in the
DECODEd programs, but in the code *before* it is DECODEd, i.e. after it is
pasted from Email or imported as a file from HTML into XyWrite, but before
DECODEing. Manuel, please compare the unDECODEd code in the paste with the
unDECODEd code in the saved TXT file. Are there additional spaces characters
at the end of lines in the paste? Extra carriage returns? Do the line endings
use Ascii-10 (line feed, 0Ah) by itself, or Ascii-13 by itself (carriage
return, 0Dh), or the two-char 13/10 combo that is typical in XyWrite (0D 0Ah,
use LIST.COM to find out -- save as files, shell to DOS, LIST the files, Alt-H
to toggle into Hex mode)? Or just save the two versions and ZIP them to me --
please mark clearly which is the paste and which the file save. I suspect that
line endings have something to do with this, especially after reading your
observation that "-ndXPLeNCODE" remains at the end (to me, that suggests that
the last char in the unDECODEd file is probably a lone Ascii-10, which XyWrite
has a very tough time handling).

Mail on the Net generally moves through Unix servers, and Unix often handles
line endings with lone Ascii-10s. I suspect that different servers and
different Email clients are handling these characters differently. But what I
don't understand is why this issue hasn't cropped up persistently in the past.
I never have any problems DECODEing anything. Of course, the minute I get an
external DECODEable file into a XyWrite window, I run "T1310" against
it, to make sure line endings are regularized with real 13/10 carriage returns
-- then I DECODE. Maybe that's what's missing(?)... But that doesn't explain
why the ZIPped version that Russ used also bombed. I unZIP it, and it's
perfect. So that in turn makes you think: defaults.

Whew. Dazed & confused.

-----------------------------
Robert Holmgren
holmgren@xxxxxxxx
-----------------------------