Re: uniqcomp.pm - help requested

["Robert Holmgren" holmgren@xxxxxxxx]

** Reply to message from Russ Urquhart  on Tue, 15
Nov 2005 08:10:46 -0600


> under the Mac OS the line endings are
> either the CR or LF (can't remember which one exactly but i know it
> is not both.) With that said, in my case, whenever i know that i am
> going to take a file to VPC (which runs PC DOS 2000) i use either a
> unix command to change the line endings for that file to DOS or open
> the file in bbedit and have it save the file with DOS line endings.

Or, just CAll the file in XyWrite, and command T1310. That will
regularize the line endings as 13/10 (carriage return/line feed).

> I'm using the POST35.PRN and it has the [MODE TABLE and
> ATTRIBUTE DEFINITIONS] settings you described.

Good. One other requirement: DFL should have tables similar to this:

;Normal settings
MD BI=30
MD BO=15
MD BR=5
MD BU=9
MD FL=28
MD FN=18
MD FR=12
MD FU=129
MD IT=9
MD MM=0
MD NM=7
MD RV=4
MD SB=14
MD SD=2
MD SO=112
MD SU=20
MD UL=9
;
;B=Bold D=Subscript I=Italic L=BoldUnderline N=Normal
;O=BoldItalic R=Reverse S=Superscript U=Underline V=BoldReverse
;DELETE modes (Redlining)
MD DB=79
MD DD=76
MD DI=74
MD DL=75
MD DN=71
MD DO=78
MD DR=116
MD DS=77
MD DU=67
MD DV=66
;
;INSERT modes (Redlining)
MD IB=47
MD ID=44
MD II=42
MD IL=43
MD IN=39
MD IO=46
MD IR=114
MD IS=37
MD IU=41
MD IV=122
;
...

> Interestingly, when i have the Redline mode turned on for both of my
> documents, and run UNIQOMP, it does something different. It appears
> to go to the point of difference and then, instead of putting an
> endless string of NMIN, it appears to stay in the same place kind of
> vibrating, endlessly until stopped.

You can't run UNIQOMP with Redlining ON (I now turn it off automatically).
UNIQOMP goes into eXPanded mode for a few functions, but Redlining forbids
toggling into eXPanded mode.

Help Request!
==== =======
I realized that there is nothing wrong with the PEV command.
It's fine. What's wrong is Redlining (RED ON) itself:
it has a bug. While figuring out how it works (I have zero
prior experience with Redlining), I noticed that it never
makes use of any Insert MoDe except , whereas it makes
full use of all the Delete MoDes. How very odd! And I had
fully mimicked that (buggy) behavior in UNIQOMP, too.

Perhaps someone with more experience of Redlining could try
this experiment: Make a short dummy file -- just a couple
of lines. Put one or more MoDes other than NorMal at the
very beginning, e.g. , , , or anything
you wish -- let's say, for demonstration purposes, that you
use BOld . Start Redlining (RED ON). Insert some
text somewhere. Move to another spot and delete some text.
Stop Redlining (RED OFF). Toggle into eXPanded mode,
and look at the places where you added/deleted text.

You started the document with , therefore the whole
document should be BOld. On my machine, the deleted
sections are marked at the beginning with , and at the
end with  -- in other words, the deleted sections are
indicated as "mdDB" (DeleteBold) -- which means that they're
sensitized to the prevailing BOld MoDe. Now look at your
text *additions*: on my machine, they're marked at the
beginning with , and at the end with  -- in
other words, the deleted sections are indicated as "mdIN"
(InsertNormal) -- they're NOT sensitized to the prevailing
BOld MoDe! Instead, they're marked NorMal. In my book,
that's a bug.

So I rewrote UNIQOMP again to do this properly, and now
there are no longer any artifacts left in the text after you
PEV (i.e., additions are *no longer* stripped of the
prevailing MoDe, as I reported last night -- the prevailing
MoDe is maintained throughout -- PEV works!).

Assuming that I am using Redlining correctly (which is what
I'd like a more experienced user to confirm), what needs
doing here is to write a replacement Redlining routine (a
replacement for RED ON), that performs Redlining correctly.

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