NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/bit/listserv/ibm7171/117

Text File | 1992-11-17 | 3.5 KB | 73 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!utcsri!newsflash.concordia.ca!garrot.DMI.USherb.CA!uxa.ecn.bgu.edu!psuvax1!psuvm!auvm!SLACVM.SLAC.STANFORD.EDU!JXH
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.ibm7171
Date:         Mon, 16 Nov 1992 12:25:00 -0800
Sender:       Protocol Converter list <[e-mail]>
From:         John Halperin <[e-mail]>
Subject:      Re: 7171 and tabs
Comments: To: [e-mail],
          Herbert Schumacher <[e-mail]>
Lines: 60

> Hi,
> after 10 years we have a nice "problem" with the 7171: If you have
> a file containing hex '05' (Tabs) and you want to XEDIT it (for
> XEDIT we set REMOTE ON), you will see the error message: DMSXSC117S
> Error writing to display terminal.  If you now set REMOTE OFF, the
> error disappears.  Is it an error or does it work as designed??

This is an old problem.  With REMOTE ON, XEDIT uses the 3270 RA
(Repeat to Address) order to compress strings of 5 (?) or more
duplicate characters when sending the data to the 3270. XEDIT
also translates EBCDIC Tab characters (X'05') to the 3270
sort-of-character, FM (Field Mark, X'1E').  Strings of Tab chars
are thus written to the 3270 as an RA order with FM as its character
argument.

The question is:  Is it legal to use RA with sort-of-characters such
as FM?  The older 3270 publications were vague on this.  The Yale
Series/1 product didn't allow any EBCDIC control chars (less than
X'40') to be used with RA, and the IBM 7171 apparently copied that
restriction.  XEDIT went ahead and used RA with FM anyway.

I've appended an old posting by Howard Gilbert, one of the authors
of the Yale product, about this problem.  Anyone that wants a fix
to the Yale code in Series/1 source format can write me (don't
everyone ask at once :-).  Realistically, the "fix" is to run XEDIT
with REMOTE OFF.

-- John Halperin <[e-mail]>
   Stanford Linear Accelerator Center

---------------------- IBM7171 MAIL from IBM7171 at MARIST.
Date:         Wed, 11 Mar 87 17:01:13 EST
From:         Howard Gilbert <GILBERT@YALEVMX>
Subject:      Re: What is a character

The question arises about what you can "Repeat to Address". It has
been suggested that you can use "any character".
Everyone agrees that 'A' is a CHARACTER.
Everyone agrees that X'11'(SBA) is an ORDER.
On the other hand, SF followed by a byte can be
     1) An order if the byte is an attribute byte
     2) A character in the 3277 with APL.
The X'08' GE graphic escape is similarly used to build APL compounds
on a 3278.  It is certainly not a character, but it is not an order
either. On the other hand, Repeat to Address does not take a two
byte character (anyone know about Double Byte Asian support?)
There are four special "printer orders" CR, NL, EM, and FF.  They
may display on a tube, but they are orders which just look like characters.
FM and DUP probably have the best chance of being regarded as
characters.  They are even called characters in some but not all of the
documentation. However, they are technically called "control characters"
most of the time.
Therefore, the 3270 DS has characters, orders, printer format orders,
compound characters, DB characters, graphic escape, and control
characters.  Which can be used in a RA sequence?  I have asked the
question to IBM for many years and never found an answer to this or
any of dozens of other comparable issues.

Do not count on fixing it just because it works differently on a 3274.
It may work even differently on a 3271, 3174, 4361, 9370, ...