Ada Terminal Emulator for Windows

Ada Terminal Emulator for Windows

[Ross Higson]

I've seen quite a few requests here recently for ANSI terminal support for
Windows (notably Windows NT/2000/etc that no longer include the
ANSI.SYS driver).

While that's not the main purpose of the Ada Terminal Emulator (see
summary description below), it happens to do it quite well.

You can either run existing ANSI applications (Ada or non-Ada) using the
"redirect" utility included in the package, or (Ada only) recompile them
after subsituting "Term_IO" for any references to Text_IO. Both methods
replace the default files (i.e. stdin, stdout, stderr) with ANSI compatible
terminal windows.

You can get the Ada Terminal Emulator from www.zip.com.au/~rossh

Note that the package also requires GNAT, and also GWindows (available
from www.adapower.com). For more details, see the documentation in the
Terminal Emulator package.

Ross Higson.


Ada Terminal Emulator - Summary
-------------------------------

The Ada Terminal Emulator package provides a set of terminal emulation
capabilities, implemented in Ada 95, to run under Windows 95/98/NT/2000.
All source code for the terminal emulator is provided under the GNU General
Public License. The package was developed using GNAT and GWindows.

In addition to providing simple "dumb" terminal emulation, the package
provides full emulation of DEC VT52/VT100/VT102 compatible terminals,
including double height and double width characters, smooth scrolling,
special graphics, display controls and national replacement character sets.
The emulator also implements a substantial subset of VT220/VT420 and ISO
6429 capabilities.

The package consists of three main components:

Terminal_Emulator:
------------------
An Ada package that provides facilities for creating, configuring and
interacting with terminal windows from an Ada program. Multiple terminal
windows can be created from the same program. Each window implements a
completely independent terminal emulator. Each can be separately configured
as a simple "dumb" terminal, or emulate full DEC VTxxx capabilities.

Major features:

Character attributes include italic, bold, underline, strikeout, reversed,
and flashing, as well as foreground and background color, using any windows
font (not just fixed pitch fonts).

Line attributes include double width and double height characters.

Support for Windows character sets; DEC special graphic character sets;
DEC multinational character sets; DEC national replacement character sets
and DEC display controls character sets (displays control codes as
characters).

Support for DEC VTxxx control sequences, ISO 6429 control sequences and
ANSI.SYS control sequences. Support for 7 or 8 bit control sequences.

Support for DEC user-definable function keys, numeric and application and
editing keypads, and language-specific keyboards.

Separately sizable virtual buffer, screen, view, and scrolling region:
- Virtual buffer sizes up to 32,767 rows (lines) by 1024 columns (characters).
- Screen sizes up to 32,727 rows (lines) by 1024 columns (characters).
- View sizes up to the lesser of the screen size and the display capacity.
- Rectangular scrolling regions of any size up to the screen size.

Support for smooth (soft) scrolling of the screen or scrolling region.

Support for separate cursors, character attributes and screen colors for
input and output operations

Mouse support - select text by character, word, line or rectangular region.
Copy and paste between terminal windows and other Windows applications.
Resize the screen, the view, or the font size using the mouse.

File support - load and save virtual buffer to/from text files.

Printer support - printer setup, page setup, print current selection, print
entire buffer.


Term_IO:
--------
An Ada package that can be used as a complete and transparent replacement
for the standard Ada text-handling package Text_IO. In addition to supporting
all the normal file handling capabilities of Text_IO, Term_IO replaces the
default files (i.e. standard input, standard output and standard error) with
one or more terminal windows. It also allows the creation and manipulation of
user-defined terminal windows using only Text_IO syntax. Each terminal window
can be configured to either strictly implement the semantics of Text_IO as
defined in the Ada 95 Language Reference Manual (LRM), or to implement relaxed
semantics that allow the user to take advantage of the terminal capabilities
in a more intuitive manner.

Redirect:
---------
A stand-alone program that allows existing text based application (including
non-Ada applications) to have their standard I/O (i.e. standard input, standard
output and standard error) redirected through a terminal window - without
recompiling or relinking. Redirect supports all the features of the
Terminal_Emulator, and adds line editing, command and filename completion and
command history. It can be used to create a sophisticated window-based command
line interpreter with only a few dozen lines of programming. A complete Ada 95
example is included.

Ada Terminal Emulator for Windows

[Ross Higson]

I've seen quite a few requests here recently for ANSI terminal support for
Windows (notablyWindows NT/2000/etc that no longer include the
ANSI.SYS driver).

While that's not the main purpose of the Ada Terminal Emulator (see
summary description below), it happens to do it quite well.

You can either run existing ANSI applications (Ada or non-Ada) using the
"redirect" utility included in the package, or (Ada only) recompile them
after subsituting "Term_IO" for any references to Text_IO. Both methods
replace the default files (i.e. stdin, stdout, stderr) with ANSI compatible
terminal windows.

You can get the Ada Terminal Emulator from www.zip.com.au/~rossh

Note that the package also requires GNAT, and also GWindows (available
from www.adapower.com). For more details, see the documentation in the
Terminal Emulator package.

Ross Higson.


Ada Terminal Emulator - Summary
-------------------------------

The Ada Terminal Emulator package provides a set of terminal emulation
capabilities, implemented in Ada 95, to run under Windows 95/98/NT/2000.
All source code for the terminal emulator is provided under the GNU General
Public License. The package was developed using GNAT and GWindows.

In addition to providing simple "dumb" terminal emulation, the package
provides full emulation of DEC VT52/VT100/VT102 compatible terminals,
including double height and double width characters, smooth scrolling,
special graphics, display controls and national replacement character sets.
The emulator also implements a substantial subset of VT220/VT420 and ISO
6429 capabilities.

The package consists of three main components:

Terminal_Emulator:
------------------
An Ada package that provides facilities for creating, configuring and
interacting with terminal windows from an Ada program. Multiple terminal
windows can be created from the same program. Each window implements a
completely independent terminal emulator. Each can be separately configured
as a simple "dumb" terminal, or emulate full DEC VTxxx capabilities.

Major features:

Character attributes include italic, bold, underline, strikeout, reversed,
and flashing, as well as foreground and background color, using any windows
font (not just fixed pitch fonts).

Line attributes include double width and double height characters.

Support for Windows character sets; DEC special graphic character sets;
DEC multinational character sets; DEC national replacement character sets
and DEC display controls character sets (displays control codes as
characters).

Support for DEC VTxxx control sequences, ISO 6429 control sequences and
ANSI.SYS control sequences. Support for 7 or 8 bit control sequences.

Support for DEC user-definable function keys, numeric and application and
editing keypads, and language-specific keyboards.

Separately sizable virtual buffer, screen, view, and scrolling region:
- Virtual buffer sizes up to 32,767 rows (lines) by 1024 columns 
(characters).
- Screen sizes up to 32,727 rows (lines) by 1024 columns (characters).
- View sizes up to the lesser of the screen size and the display capacity.
- Rectangular scrolling regions of any size up to the screen size.

Support for smooth (soft) scrolling of the screen or scrolling region.

Support for separate cursors, character attributes and screen colors for
input and output operations

Mouse support - select text by character, word, line or rectangular region.
Copy and paste between terminal windows and other Windows applications.
Resize the screen, the view, or the font size using the mouse.

File support - load and save virtual buffer to/from text files.

Printer support - printer setup, page setup, print current selection, print
entire buffer.


Term_IO:
--------
An Ada package that can be used as a complete and transparent replacement
for the standard Ada text-handling package Text_IO. In addition to 
supporting
all the normal file handling capabilities of Text_IO, Term_IO replaces the
default files (i.e. standard input, standard output and standard error) 
with
one or more terminal windows. It also allows the creation and 
manipulation of
user-defined terminal windows using only Text_IO syntax. Each terminal 
window
can be configured to either strictly implement the semantics of Text_IO as
defined in the Ada 95 Language Reference Manual (LRM), or to implement 
relaxed
semantics that allow the user to take advantage of the terminal 
capabilities
in a more intuitive manner.

Redirect:
---------
A stand-alone program that allows existing text based application 
(including
non-Ada applications) to have their standard I/O (i.e. standard input, 
standard
output and standard error) redirected through a terminal window - without
recompiling or relinking. Redirect supports all the features of the
Terminal_Emulator, and adds line editing, command and filename 
completion and
command history. It can be used to create a sophisticated window-based 
command
line interpreter with only a few dozen lines of programming. A complete 
Ada 95
example is included.

Re: Ada Terminal Emulator for Windows

[Larry Kilgallen]

In article <17vJa.7569$eE.70461@nasal.pacific.net.au>, Ross Higson <rossh@zip.com.au> writes:

> In addition to providing simple "dumb" terminal emulation, the package
> provides full emulation of DEC VT52/VT100/VT102 compatible terminals,
> including double height and double width characters, smooth scrolling,
> special graphics, display controls and national replacement character sets.
> The emulator also implements a substantial subset of VT220/VT420 and ISO
> 6429 capabilities.

The DEC official compatibility test is not available to the public,
but I would recommend someone who has a Windows machine should test
this terminal emulator against a DEC Notes session, which for me has
been the most severe test every time I have had to use a Windows machine.

Re: Ada Terminal Emulator for Windows

[Ross Higson]

It would be great if someone could take up Larry's suggestion, and email 
me any discrepencies they find.
I'll fix them in the next version.

I've been using Per Lindberg and Thomas E. Dickey's "vttest" program to 
test VT compatibility, but
there's a lot of  areas that vttest doesn't test (particularly for the 
VT420) - and also a lot of areas where
I cannot be absolutely sure my Windows version (actually a Cygwin 
version) of vttest behaves correctly.
Vttest relies partly on functionality provided by the Unix tty device 
(e.g. it uses stty to enable/disable host
echo, to perform inbound and outbound CR/LF translation etc).

I've compared the terminal emulator against xterm, and (apart from a few 
cases where xterm itself
fails the vttest) the results are identical. Does anyone know how xterm 
stacks up against the official
DEC compatibility test ?.

Ross.

Larry Kilgallen wrote:

>In article <17vJa.7569$eE.70461@nasal.pacific.net.au>, Ross Higson <rossh@zip.com.au> writes:
>
>  
>
>>In addition to providing simple "dumb" terminal emulation, the package
>>provides full emulation of DEC VT52/VT100/VT102 compatible terminals,
>>including double height and double width characters, smooth scrolling,
>>special graphics, display controls and national replacement character sets.
>>The emulator also implements a substantial subset of VT220/VT420 and ISO
>>6429 capabilities.
>>    
>>
>
>The DEC official compatibility test is not available to the public,
>but I would recommend someone who has a Windows machine should test
>this terminal emulator against a DEC Notes session, which for me has
>been the most severe test every time I have had to use a Windows machine.
>  
>

Re: Ada Terminal Emulator for Windows

[Thomas Dickey]

Ross Higson <rossh@zip.com.au> wrote:
> It would be great if someone could take up Larry's suggestion, and email 
> me any discrepencies they find.
> I'll fix them in the next version.

> I've been using Per Lindberg and Thomas E. Dickey's "vttest" program to 
> test VT compatibility, but
> there's a lot of  areas that vttest doesn't test (particularly for the 
> VT420) - and also a lot of areas where
> I cannot be absolutely sure my Windows version (actually a Cygwin 
> version) of vttest behaves correctly.
> Vttest relies partly on functionality provided by the Unix tty device 
> (e.g. it uses stty to enable/disable host
> echo, to perform inbound and outbound CR/LF translation etc).

> I've compared the terminal emulator against xterm, and (apart from a few 
> cases where xterm itself
> fails the vttest) the results are identical. Does anyone know how xterm 

I'm intrigued (other than the obvious - blinking characters not implemented -
are there any bugs that I may have overlooked?)

> stacks up against the official
> DEC compatibility test ?.

That would be nice to know.  (I ran vttest against dxterm a few years ago,
and it mostly-agreed with vttest, though iirc, its vt300 emulation was
not complete).

> Ross.

> Larry Kilgallen wrote:

>>In article <17vJa.7569$eE.70461@nasal.pacific.net.au>, Ross Higson <rossh@zip.com.au> writes:
>>
>>  
>>
>>>In addition to providing simple "dumb" terminal emulation, the package
>>>provides full emulation of DEC VT52/VT100/VT102 compatible terminals,
>>>including double height and double width characters, smooth scrolling,
>>>special graphics, display controls and national replacement character sets.
>>>The emulator also implements a substantial subset of VT220/VT420 and ISO
>>>6429 capabilities.
>>>    
>>>
>>
>>The DEC official compatibility test is not available to the public,

I tend to disregard the term "official" in this context.

>>but I would recommend someone who has a Windows machine should test
>>this terminal emulator against a DEC Notes session, which for me has
>>been the most severe test every time I have had to use a Windows machine.
>>  
>>


-- 
Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>
http://dickey.his.com
ftp://dickey.his.com

Re: Ada Terminal Emulator for Windows

[Ross Higson]

Thomas Dickey <dickey@saltmine.radix.net> wrote in message news:<bd9hsj$k8h$1@news1.radix.net>...

> > I've compared the terminal emulator against xterm, and (apart from a few 
> > cases where xterm itself
> > fails the vttest) the results are identical. Does anyone know how xterm 
> 
> I'm intrigued (other than the obvious - blinking characters not implemented -
> are there any bugs that I may have overlooked?)
> 

One instance is that when running vttest (version 2.7) under a Cygwin
xterm (version XFree864.2.0(165)), I see the following output halfway
through test 1:
-------------------------------------------------------------
Test of cursor-control characters inside ESC sequences.
Below should be four identical lines:

A B C D E F G H I
A B C D E F G H I
16CI
A 
  AB 
     AC 
        AD 
           AE 
              AF 
                 AG 
                    AH 
                       AI 
                          A

Push <RETURN>
--------------------------------------------------------------

Maybe this is a peculiarity of Cygwin. If I recall any others I'll
post them.

By the way, I've updated the terminal emulator package to include a
simple "comms" wrapper program that allows a PC to be more easily used
as a serial terminal. Hopefully, this will encourage someone to to do
a side-by-side comparison with a real terminal.

Ross.

Re: Ada Terminal Emulator for Windows

[Simon Clubley]

In article <3c1f2fe3.0306241554.2022d78e@posting.google.com>, rossh@zip.com.au (Ross Higson) writes:
> 
> By the way, I've updated the terminal emulator package to include a
> simple "comms" wrapper program that allows a PC to be more easily used
> as a serial terminal. Hopefully, this will encourage someone to to do
> a side-by-side comparison with a real terminal.
> 

Evaluating this emulator is on my list of things to look at over the next
couple of weeks or so if I have some spare time.

One of the things that is putting me off evaluating it is the several other
packages that I have to download and install. For example, do you really
need Gnat 3.15p to build your emulator/Gnatcom/GWindows, or can you use
Gnat 3.14p ? I seem to recall seeing reports of several problems with 3.15p.

Simon.

PS: Ross, you may like to be aware that VMS has a habit of breaking terminal
emulators; my fastest time for breaking an emulator when running it against
VMS is about 30 seconds... :-)

PPS: Larry's suggestion of using Notes is a good one; if an emulator is
going to break, I usually break it when trying to run it against TPU in EDT
keypad mode. If it survives that :-), I will give Notes a try.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       
VMS advocate: One who makes a Mac advocate look like a beginner at advocacy.

Re: Ada Terminal Emulator for Windows

[Larry Kilgallen]

In article <jfKQkTa1IhTj@eisner.encompasserve.org>, clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) writes:

> PS: Ross, you may like to be aware that VMS has a habit of breaking terminal
> emulators; my fastest time for breaking an emulator when running it against
> VMS is about 30 seconds... :-)
> 
> PPS: Larry's suggestion of using Notes is a good one; if an emulator is
> going to break, I usually break it when trying to run it against TPU in EDT
> keypad mode. If it survives that :-), I will give Notes a try.

Perhaps I should have qualified my statement.  I find Notes a more severe
test that Mail, but that is among the programs I am likely to use from a
strange machine.  Some other commands that might be a good test are MONITOR,
LICENSE LIST/FULL, SHOW CLUSTER and SHOW PROCESS/CONTINUOUS.  These all have
interactive capabilities which should be exercised rather than just
being satisfied with the default output text.

Re: Ada Terminal Emulator for Windows

[Simon Clubley]

In article <GDhxauyUg2GN@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:
> In article <jfKQkTa1IhTj@eisner.encompasserve.org>, clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) writes:
> 
>> PPS: Larry's suggestion of using Notes is a good one; if an emulator is
>> going to break, I usually break it when trying to run it against TPU in EDT
>> keypad mode. If it survives that :-), I will give Notes a try.
> 
> Perhaps I should have qualified my statement.  I find Notes a more severe
> test that Mail, but that is among the programs I am likely to use from a
> strange machine.  Some other commands that might be a good test are MONITOR,
> LICENSE LIST/FULL, SHOW CLUSTER and SHOW PROCESS/CONTINUOUS.  These all have
> interactive capabilities which should be exercised rather than just
> being satisfied with the default output text.

I never use Notes (except on Eisner) so it never occurred to me that it's a
good test of an emulator's capabilities; it's now been added to my list of
applications to try against a new emulator in the future.

I do use MONITOR as part of evaluating a new emulator however; it's a good
test of an emulator's line drawing capabilities.

BTW, Ross, what telnet negotiation options does your emulator support ?

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       
VMS advocate: One who makes a Mac advocate look like a beginner at advocacy.

Re: Ada Terminal Emulator for Windows

[Ross Higson]

clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote in message news:<+e8pEAcrjkqR@eisner.encompasserve.org>...
> 
> BTW, Ross, what telnet negotiation options does your emulator support ?
> 

None - feel free to add whatever you want !. Seriously though, If I get time 
I'll look into adding a telnet wrapper, similar to the comms wrapper. But I'm 
not a telnet expert, and it won't be top of my list as I have no need for it.
Anyway, I'll probably be too busy fixing the bugs you will no doubt find in the
emulator when you test it :)

To answer your previous question, I'll try compiling with GNAT 3.14. I'll update
the terminal emulator home page with the results.

Ross.

Re: Ada Terminal Emulator for Windows

[Ross Higson]

The terminal emulator now compiles with GNAT 3.14p as well as 3.15p,
but it required a small source change. The latest source distribution
(v0.5) is available from http://www.zip.com.au/~rossh/

I've also added a compiled version of the stand-alone Comms and
Redirect programs for anyone who wants to try out the emulator without
having to install GWindows first. But you will probably still need the
source distribution, since it contains the documentation.

Ross.

Re: Ada Terminal Emulator for Windows

[Larry Kilgallen]

As it happens, comp.os.vms the past few days has had a discussion
of some flaws in a commercial terminal emulator.  There are a few tough
cases mentioned that might be work considering as test scenarios.

Re: Ada Terminal Emulator for Windows

[Ross Higson]

Larry,

Thanks. If you are talking about the <ESC>[?2K (correct) versus 
<ESC>[2?K (incorrect) tests, my emulator fails that test as well. The 
MIT parser allows the flag character (in this case '?') to appear 
anywhere in the control sequence, so in situations like this both the 
control sequences are treated identically, whereas a real VT220 would 
ignorethe second case.  It's quite likely that the commercial emulator 
in question uses the same parser.

There's no easy fix to this problem without rewriting the MIT parser, 
which I currently have no plans to do. However, I guess there is a moral 
here - the parser is the only part of the emulator not written in Ada ! :)

Ross.

Larry Kilgallen wrote:

>As it happens, comp.os.vms the past few days has had a discussion
>of some flaws in a commercial terminal emulator.  There are a few tough
>cases mentioned that might be work considering as test scenarios.
>  
>

Re: Ada Terminal Emulator for Windows

[Larry Kilgallen]

In article <FZdLa.7948$eE.87173@nasal.pacific.net.au>, Ross Higson <rossh@zip.com.au> writes:

> There's no easy fix to this problem without rewriting the MIT parser, 
> which I currently have no plans to do. However, I guess there is a moral 
> here - the parser is the only part of the emulator not written in Ada ! :)

I think the moral might rather be the parser is the only part not
written by someone experienced with Ada.

Re: Ada Terminal Emulator for Windows

[Thomas Dickey]

Ross Higson <rossh@zip.com.au> wrote:
> Thomas Dickey <dickey@saltmine.radix.net> wrote in message news:<bd9hsj$k8h$1@news1.radix.net>...

>> > I've compared the terminal emulator against xterm, and (apart from a few 
>> > cases where xterm itself
>> > fails the vttest) the results are identical. Does anyone know how xterm 
>> 
>> I'm intrigued (other than the obvious - blinking characters not implemented -
>> are there any bugs that I may have overlooked?)
>> 

> One instance is that when running vttest (version 2.7) under a Cygwin

I should update the "2.7" part, but typically just use the patch date
for vttest.  That's currently 20030301.

> xterm (version XFree864.2.0(165)), I see the following output halfway
> through test 1:
> -------------------------------------------------------------
> Test of cursor-control characters inside ESC sequences.
> Below should be four identical lines:

> A B C D E F G H I
> A B C D E F G H I
> 16CI
> A 
>   AB 
>      AC 
>         AD 
>            AE 
>               AF 
>                  AG 
>                     AH 
>                        AI 
>                           A

This looks more like a problem with vttest running on Cygwin.
(There are scattered, but persistent bugs in its terminal support).

There is one gotcha to watch out for with test 1 (xterm's line-wrap feature
is set by some people, making a glitch in the box drawn in test 1).


> Push <RETURN>
> --------------------------------------------------------------

> Maybe this is a peculiarity of Cygwin. If I recall any others I'll
> post them.

> By the way, I've updated the terminal emulator package to include a
> simple "comms" wrapper program that allows a PC to be more easily used
> as a serial terminal. Hopefully, this will encourage someone to to do
> a side-by-side comparison with a real terminal.

> Ross.

-- 
Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>
http://dickey.his.com
ftp://dickey.his.com

Re: Ada Terminal Emulator for Windows

[Ross Higson]

Thomas Dickey <dickey@saltmine.radix.net> wrote in message news:<bdc4p4$59$1@news1.radix.net>...
>  
> I should update the "2.7" part, but typically just use the patch date
> for vttest.  That's currently 20030301.

That's the version I'm using.
 
> This looks more like a problem with vttest running on Cygwin.
> (There are scattered, but persistent bugs in its terminal support).

It appears to be unrelated to vttest, but it may well be a Cygwin problem.
The relevant output of vttest (CUF with embedded CR and CUU with embedded VT)
looks fine, and sending similar sequences to xterm without using vttest causes 
the same result.

If you need any more info, email me direct.

Ross.

Re: Ada Terminal Emulator for Windows

[Thomas Dickey]

Ross Higson <rossh@zip.com.au> wrote:
> Thomas Dickey <dickey@saltmine.radix.net> wrote in message news:<bdc4p4$59$1@news1.radix.net>...
>>  
>> I should update the "2.7" part, but typically just use the patch date
>> for vttest.  That's currently 20030301.

> That's the version I'm using.
>  
>> This looks more like a problem with vttest running on Cygwin.
>> (There are scattered, but persistent bugs in its terminal support).

> It appears to be unrelated to vttest, but it may well be a Cygwin problem.

I have had problems with cygwin doing staircasing - that might be related.
But I hadn't paid much attention to vttest on cygwin - I seem to recall that
it did not build (but see that the current version of cygwin does that much).
I'll have to compile xterm with traces to see whether it is receiving the
proper data.

> The relevant output of vttest (CUF with embedded CR and CUU with embedded VT)
> looks fine, and sending similar sequences to xterm without using vttest causes 
> the same result.

> If you need any more info, email me direct.

> Ross.

-- 
Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>
http://dickey.his.com
ftp://dickey.his.com