From:	MERC::"uunet!WKUVX1.BITNET!MacroMan" 28-FEB-1993 07:59:43.07
To:	MACRO32@WKUVX1.BITNET
CC:	
Subj:	Re: SMG Termtable user def's. (internals)

In article <01GV6GT585HE9BVFC9@MONMOUTH-etdl1.Army.MIL>,
 SCHENKENBERG@Monmouth-ETDL1.Army.MIL (Brian J. Schenkenberger, VAXman) writes:
> A client using SMG for terminal I/O has decided to add color to the
> application.  I have created two custom SMG termtable entries for a
> Wyse370 (pseudo-VTxxx + mfg. specific color esc. seq.) and a PC with
> an ANSI terminal emulation (ie. ProCOMM, CrossTalk) and IBM defined
> graphics codes.
>
> The terminals may now be specified (once TERM$TABLOC is defined) with
> $ SET TERMINAL/DEVICE=Wyse370  - or -  $ SET TERMINAL/DEVICE=ANSI_PC.
>
> GETDVI/DVI$_DEVTYPE now returns numeric type codes for the Wyse370 and
> ANSI_PC as 189 and 188, respectively.  Doing some investigation, the
> numeric assignment (order of) appears to be defined by the order in
> which each new terminal type was defined in the TERMTABLE.TXT source
> file.  However, the values 188 and 189 are always used for these new
> terminal def's.
>
> Question: Does anyone know if these values will be retained from version
> to version?  Will these #'s be retained if more terminals are defined?

        Yes "anybody" does: the answer is NO!

        In  addition  to  noticing  that  the  device  number  returned by
    F$GETDVI  depends  on  the  order  you've  entered  the  new  terminal
    definitions in your SMGTERMS.TXT  (or  separate  site-specific  file),
    note that the _value_ also depends on whether you've incorporated your
    definitions into SYS$SYSTEM:TERMTABLE.EXE (via TERMTABLE.TXT, etc.) or
    whether  you've  created an additional "termtable".EXE which you point
    to via TERM$TABLOC.

        Having worried about this (a lot!) for our site, which comprises a
    half  a   dozen   independently   managed   VAXclusters,  and  wanting
    _consistent_  DEVTYPE's across clusters (so that, e.g., SET HOST  gets
    the correct terminal type on  the  remote  end),  I  STRONGLY  suggest
    putting   your  additions  in  a  SEPARATE  "termtable".EXE,  and  use
    TERM$TABLOC to point to  it.   Put  ONLY  your  additions in it (don't
    override  DEC's  terminal definitions in SYS$SYSTEM by adding them  to
    yours: the system will look  at  your  _short_  list  first,  then  at
    SYS$SYSTEM:TERMTABLE.EXE  if it can't find what it's looking for).  If
    you do this, you won't  need  to  rebuild your site-specific termtable
    after every VMS upgrade (to get DEC's new terminal definitions), _and_
    the DEVTYPE numbers for your terminals won't change.

> I'm trying to find out IF these values can be used reliably from a DCL
> procedure using F$getdvi("TT","DEVTYPE").  I'd like to avoid a hack of
> calling SMG$INIT_TERM_TABLE_BY_TYPE and passing the terminal description
> back to a DCL symbol.

        It takes approximately two  (executable)  lines  of Fortran to use
    SMG$INIT_TERM_TABLE_BY_TYPE;  I use a 14-line (including comments  and
    blanks lines) Fortran program to retrieve  the  terminal  device  name
    string and set a symbol for further use in DCL.  I consider this safer
    than depending, e.g., on 188 and 189 because those values could change
    if  you  add more definitions to your termtable, or move things around
    in the file, but the  device  name  string  won't.   In other words, I
    don't think this solution should be sneered at as a "hack"...  :-)

> While I'm at it...  I'd like to hear from anyone that has defined SMG
> terminal tables other than the aforementioned.

        You just did ;-)

                -Ken
--
 Dr. Kenneth H. Fairfield    |  Internet: Fairfield@Slac.Stanford.Edu
 SLAC, P.O.Box 4349, MS 98   |  DECnet:   45537::FAIRFIELD (45537=SLACVX)
 Stanford, CA   94309        |  BITNET    Fairfield@Slacvx
 ----------------------------------------------------------------------------
 These opinions are mine, not SLAC's, Stanford's, nor the DOE's...