A new version of QCTerm is available for downloading. The new version is
0.90q, August 13, 2001, and can be gotten from:

     http://aics-research.com/qcterm/index.html

This new version corrects a few minor errors and a couple of major ones,
primarily associated with host-commanded printing and a CRLF problem that is
in MPE/iX's telnet client.


PRINTING

Although LOG BOTTOM and LOG TOP have worked as they should from the very
beginnings of QCTerm, the RECORD mode behavior was pretty ratty, primarily
due to our not understanding perfectly well how it should work. I believe
that that has all now been fixed in this version of QCTerm, and I do
appreciate Wyell's, Stan's and few other people's comments in that regard.

A good portion of getting this behavior to work correctly was archaeological.
Because of the changes that have occurred in the various models of real HP
terminals over the years, and because of the number of emulators that have
occurred, it took a little while to figure out how all of the various escape
sequences were supposed to work. RECORD mode has probably been the most
variable part of the terminal's behavior.

In the very earliest HP terminals, internal tape drives were present. The
commands such as "advance tape", "skip eof", etc. to drive those devices were
marked with a "u" (unit) designation. In contrast, transfer operations were
marked with an "s" (source) and "d" (destination) designations. Over the
years, with the disappearance of the terminal-based tape drives and the
number of HP terminal emulators that have been built, the "d" and "u"
designations have become confused. Although HP no longer supports the "u"
designation, Reflection does, and thus so do we. In effect, the "u" and "d"
designations have now become equivalents of one another and that's the way
they're now programmed in QCTerm.

The same is true of the "4d" and "6d" "to printer" designations. Originally,
the "4" value represented the terminal's external printer and the "6" the
later (on some models) internal printer. We now consider these two values to
be equivalents.

Also several people asked that we suppress QCTerm's page headers when
printing through RECORD mode so that they could print onto pre-printed forms
placed in a PC's printer. We've done that with this version. We've also added
a new QCTerm-only escape sequence so that the host program can select the
font size to be used and the number of copies to be printed. This sequence
will work with either PCL or PostScript printers.

We also prepared the groundwork with this release, even though you can't see
it, for van Gogh printing. In this mode, we expect not only to pass
PostScript (and other languages) through the emulator to the printer
transparently, we also expect to be able to print out exceptionally good
looking forms, short notes, bibliographies, and even small books, directly
from the host to the end-user's PC-based printer silently through QCTerm.


QEDIT and FORMSPEC

To date, neither Qedit nor Formspec has worked well on QCTerm, although
Robelle graciously put in a workaround into Qedit. However, the problem
hasn't lain specifically with either Qedit or QCTerm itself, but rather with
the telnet implementation on the HP3000. For some particular reason,
undoubtedly inherited from some ancient UNIX implementation, if MPE's telnet
sees a CRLF pair, it automatically (at no extra charge, and without asking)
strips off the trailing LF character, and that kills Qedit and Formspec,
which are both looking for the LF.

Because getting the MPE "advanced telnet" patch (which fixes this problem)
out to the world has taken so much longer than originally expected, I spent a
bit of time and deduced the absolute minimum condition that would satisfy
both Qedit and Formspec that we could correct in QCTerm. The "fix" that's now
in QCTerm is to transmit a CR NULL LF sequence. MPE doesn't recognize that
sequence as needing correction and thus it passes the LF's to both programs.

The addition of the NULL has no seeming effect on other programs, regardless
of whether the connection is either telnet or serial.

I very much appreciate Tom Brandt's help with Formspec.


DEBUG

Because of Bruce Toback's off-hand comments about having your HP3000
passwords on your PC somewhere, and because of someone else's security
concerns, we've now de-installed a debug escape sequence we put into QCTerm
several years ago.

The communications trace/debug routine is still there and more valuable then
ever. It's just that you can't get to it any longer using a host command. To
get into the routine, type ALT T+D or use QCTerm's "terminal" pull down menu.
When an ALT T+D is typed a second time, the debug mode is turned off and a
copy of the resulting file is automatically brought up into either Notepad or
Wordpad, depending on its file size.

Further, when you quit your current session, or bring up a new session of
QCTerm, the communications log/debug file is now automatically purged off of
your PC.


UNIX

There are a few other items people have requested. Most of these are not bugs
but merely changes in minor behavorial attributes. We will likely change all
of these and I do greatly appreciate the suggestions.

On the other hand, we are still getting a few bug reports from UNIX users,
and all of these too will be corrected. Because we don't use UNIX here and
have no access to a UNIX device, we would greatly appreciate if you could
provide us with access to a recreated problem and a script on how to get to
the problem you're seeing.

Unfortunately, we simply can't fix problems based on word descriptions. All
of the problems we're seeing now tend to be the results of very small, easily
correctable things, but they also now tend to be very rare.

Wirt Atmar

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *