Borland Pascal 7.0 on 1.8Ghz plus computers. 
Author Message
 Borland Pascal 7.0 on 1.8Ghz plus computers.

It's looking grim for our legacy Borland Pascal 7.0 Code.  It still runs
great on Windows ME, XP, 95, 98, NT computers and processors up to
1.8ghz but a customer has recently installed it on a 2.4 ghz computer.

Every time you hit a key in the dos prompt screens of our program a have
to wiggle the mouse a bit to get your input recognised.

If you press the "G" 10 times nothing will happen until you move
the mouse a bit in the dos window and then 10 "G"'s will show up
all at once.

Using the CPU limiter "Tame" I was able to confirm that it is not
hardware or software but the the CPU speed itself that is causing this
problem.  Something like Tame is not really an option for us though.

Has any one run into this problem?  Borland has stopped supporting old
versions of Pascal for years and will not help.  Is this the end of
Dos Pascal?

--
Posted via http://www.*-*-*.com/



Wed, 27 Jul 2005 07:49:32 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:
> It's looking grim for our legacy Borland Pascal 7.0 Code.  It still runs
> ((..omitted..))
> Has any one run into this problem?  Borland has stopped supporting old
> versions of Pascal for years and will not help.  Is this the end of
> Dos Pascal?

Have you tried adjusting idle sensitivity?

And perhaps tinkering with the other  features (performance, fast paste,
etc.)?

I recall a case where a DOS program wouldn't run correctly under Win98 and I
fixed it by turning down idle sensitivity.  It turned out the customer had
hit the [?] on Idle Sensitivity and thought if a little was good, more would
be better.  He got away with it on all but one of his DOS apps.  I recall
another I fixed by turning on protection, but I can't really what the
problem was.

    - Bill



Wed, 27 Jul 2005 15:41:39 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:

> It's looking grim for our legacy Borland Pascal 7.0 Code.  It still runs
> great on Windows ME, XP, 95, 98, NT computers and processors up to
> 1.8ghz but a customer has recently installed it on a 2.4 ghz computer.

> Every time you hit a key in the dos prompt screens of our program a have
> to wiggle the mouse a bit to get your input recognised.

> If you press the "G" 10 times nothing will happen until you move
> the mouse a bit in the dos window and then 10 "G"'s will show up
> all at once.

> Using the CPU limiter "Tame" I was able to confirm that it is not
> hardware or software but the the CPU speed itself that is causing this
> problem.  Something like Tame is not really an option for us though.

> Has any one run into this problem?  Borland has stopped supporting old
> versions of Pascal for years and will not help.  Is this the end of
> Dos Pascal?

Of course not, there are more Dos Pascal compilers, that are actually still
maintained.


Wed, 27 Jul 2005 19:06:12 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.



Quote:

> It's looking grim for our legacy Borland Pascal 7.0 Code.  It still
runs
> great on Windows ME, XP, 95, 98, NT computers and processors up to
> 1.8ghz but a customer has recently installed it on a 2.4 ghz computer.

> Every time you hit a key in the dos prompt screens of our program a
have
> to wiggle the mouse a bit to get your input recognised.

> If you press the "G" 10 times nothing will happen until you move
> the mouse a bit in the dos window and then 10 "G"'s will show up
> all at once.

> Using the CPU limiter "Tame" I was able to confirm that it is not
> hardware or software but the the CPU speed itself that is causing this
> problem.  Something like Tame is not really an option for us though.

> Has any one run into this problem?  Borland has stopped supporting old
> versions of Pascal for years and will not help.  Is this the end of
> Dos Pascal?

> --
> Posted via http://dbforums.com

I had a similar problem. It had to do with the way the dos emulator
handles some of the mouse functions. If you check for a mouse function,
the program hangs in much the same way as like putting
readkey
instead of
if keypressed then readkey.

For a combined read I am using something like

function readkee:char;
var ch:char;
begin
  if mouseavailable then repeat until mousebutton=0;
  repeat until keypressed or ((mousebutton<>0) and mouseavailable);
  if keypressed then ch:=readkey else ch:=#13;
  readkee:=ch;
end;

For the mouse functions I use the following unit:
http://www.tank-cleaning.demon.nl/pasfiles/mouse.pas

--
Femme



Fri, 29 Jul 2005 07:27:46 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.
MattD:

Quote:
> Has any one run into this problem?  Borland has stopped supporting old
> versions of Pascal for years and will not help.  Is this the end of
> Dos Pascal?

Is there a reason not considering compiling your application with
a 32-bit compiler, targetted at Win32 ? The main problem is that
DOS was designed for a 4.77 Mhz machine and despite all the
compatibility efforts put in the Virtual DOS Machine in Windows it
becomes difficult to maintain with today's CPU clocks.
And certainly Microsoft has other priorities, too...
An alternative would be to have a real PC XT or AT emulator, if it exists.
I knew one on Atari ST, so you could install an Atari ST emulator,
and inside of it, a 16-bit PC emulator.
________________________________________________________
Gautier  --  http://www.mysunrise.ch/users/gdm/gsoft.htm

NB: For a direct answer, e-mail address on the Web site!



Fri, 29 Jul 2005 19:07:21 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.


Quote:

> It's looking grim for our legacy Borland Pascal 7.0 Code.  It still runs
> great on Windows ME, XP, 95, 98, NT computers and processors up to
> 1.8ghz but a customer has recently installed it on a 2.4 ghz computer.

> Every time you hit a key in the dos prompt screens of our program a have
> to wiggle the mouse a bit to get your input recognised.

Sounds like the DOS box is not being updated after the screen contents have
changed. Your DOS program is writing bytes to screen memory, either directly
or by going through the slower BIOS calls, but they won't appear on the
screen until the DOS box redraws the window.

Haven't used XP, but in other Windows versions you could toggle between
windowed and full-screen modes using alt-enter. In full-screen mode, when
data are written to screen memory the hardware makes them appear
immediately - no redrawing is required. This may be an adequate solution to
your problem, although on a big monitor it looks ugly.

Perhaps XP will not redraw the screen until the emulator is not busy.
Presumably Tame works by soaking up cycles so that the emulator is less
busy, which tricks the emulator into doing a redraw. I'm guessing this way
because I know the IDE will keep the CPU at 100%.  If this guess is anywhere
close to the mark you might benefit by googling for messages about reducing
CPU usage in the archives of this newsgroup.

FP



Fri, 29 Jul 2005 19:54:19 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.
Gautier schrieb:

Quote:
> MattD:

>>Has any one run into this problem?  Borland has stopped supporting old
>>versions of Pascal for years and will not help.  Is this the end of
>>Dos Pascal?

> Is there a reason not considering compiling your application with
> a 32-bit compiler, targetted at Win32 ? The main problem is that
> DOS was designed for a 4.77 Mhz machine and despite all the
> compatibility efforts put in the Virtual DOS Machine in Windows it
> becomes difficult to maintain with today's CPU clocks.

Why? Where is the problem? Windows already does catch all events
(keyboard etc.) and routes them to the DOS Box. But where's the
technical problem with to fast CPUs here? I can't see it. It's only a
matter of scheduling, isn't it?

Greetings

Markus



Fri, 29 Jul 2005 21:26:26 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:

> Gautier schrieb:
>> MattD:

>>>Has any one run into this problem?  Borland has stopped supporting old
>>>versions of Pascal for years and will not help.  Is this the end of
>>>Dos Pascal?

>> Is there a reason not considering compiling your application with
>> a 32-bit compiler, targetted at Win32 ? The main problem is that
>> DOS was designed for a 4.77 Mhz machine and despite all the
>> compatibility efforts put in the Virtual DOS Machine in Windows it
>> becomes difficult to maintain with today's CPU clocks.

> Why? Where is the problem? Windows already does catch all events
> (keyboard etc.) and routes them to the DOS Box.

Windows doesn't always do it good. E.g. NT systems have problems with DOS
mouse support.

Quote:
> But where's the technical problem with to fast CPUs here? I can't see it.
> It's only a matter of scheduling, isn't it?

In practice, the dos-subsystems of Windows are unmaintained. New hardware, and no
fixes. (iow, problems like RTE200).

I don't think the speed is the problem. Windows is.



Sat, 30 Jul 2005 02:35:12 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Thank you for all your replies.  I am in the process of trying a few of
them to see how they go.  The super-short term resolution for this is
that it DOES work in fullscreen mode (you can use alt+enter within a dos
window to do this).

Just to answer a few of your questions....

Quote:
>Bill Leary
>Have you tried adjusting idle sensitivity?
>And perhaps tinkering with the other features >(performance, fast
>paste, etc.)?

I will try the idle sensitivity as it has me curious but I have tried
all the other ms-dos options available in the propertes area.

Quote:
>Marco Van De Voort
>Of course not, there are more Dos Pascal compilers, that >are actually
>stillmaintained.
>Gautier
>Is there a reason not considering compiling your >application with a
>32-bit compiler, targetted at Win32 ?

Unless we have missed something, we have looked into this a year or two
ago.  It seems that our Borland pascal won't compile well at all on
other pascal compilers.  We looked into porting it to something like
Delphi but as this code in question is a result of 15 or more years of
legacy code it has been estimated at 1-2 years work for 2-3 people.  We
have to do it eventually, just not right now, and we need a solution in
the meantime.

Quote:
>Femme Verbeek
>For the mouse functions I use the following unit:
>http://www.tank-cleaning.demon.nl/pasfiles/mouse.pas

I will definately have a look at the code you have provided to see if
it can relate to what we are doing.  Unfortunately the link you
provided did not work.  Do you have the file handy you could possibly
e-mail me at:

Quote:
>Frank Peel

This idea worked, where did you get this theory from?  Any sources you
can point me towards?

Quote:
>Markus Humm
>But where's the technical problem with to fast CPUs >here? I can't see
>it. It's only a matter of scheduling, isn't >it?

I'm not sure what you mean by this.  The customer in question is using
NT 4.0 SP6.  We've been using this for years with no problems.  They
have a fresh install of the OS, install our software on a 2.4Ghz
computer and a slower 1Ghz machine and one has the problem and one does
not.  Testing here we slowed the CPU down a bit with a program like Tame
(I could not underclock it...stupid Dell's) and it started to work
correctly.  I am 99% sure this problem is directly related to the CPU
speed as we have ruled out everything else, although weather it is a dos
problem or a Borland compiler problem I am not sure.
Running dos by itself the screen updates as usual.

Again, thank you all for your time, I will let you know what I come up
with as more people might run into this problem as 2.4Ghz plus computers
become more commonplace.

--
Posted via http://dbforums.com



Sat, 30 Jul 2005 02:26:35 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:

> Thank you for all your replies.  I am in the process of trying a few of
> them to see how they go.  The super-short term resolution for this is
> that it DOES work in fullscreen mode (you can use alt+enter within a dos
> window to do this).

> Just to answer a few of your questions....

>>Bill Leary
>>Have you tried adjusting idle sensitivity?
>>And perhaps tinkering with the other features >(performance, fast
>>paste, etc.)?
> I will try the idle sensitivity as it has me curious but I have tried
> all the other ms-dos options available in the propertes area.

>>Marco Van De Voort
>>Of course not, there are more Dos Pascal compilers, that >are actually
>>stillmaintained.

Not the compilers are the problem, the Dos subsystem of Windows are effectively
not fixed. See e.g. the buglists on the DJGPP site, bugs are still here which
where detected already in pre-versions of Windows 2000!!!!

So there is no problem if you use dos compilers under dos (as long as they
maintained, or kind souls do a kind of 3rd party maintaining of unmaintained
compilers, like TP/BP).

However using Dos programs (specially DPMI ones) under Windows NT versions
is getting increasingly more difficult with each version. IOW it is _that_
what is unmaintained)

(see e.g. http://clio.rice.edu/djgpp/win2k/main.htm)



Sat, 30 Jul 2005 05:09:34 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:

> Thank you for all your replies.  I am in the process of trying a few of
> them to see how they go.  The super-short term resolution for this is
> that it DOES work in fullscreen mode (you can use alt+enter within a dos
> window to do this).

> Just to answer a few of your questions....
> >Is there a reason not considering compiling your >application with a
> >32-bit compiler, targetted at Win32 ?
> Unless we have missed something, we have looked into this a year or two
> ago.  It seems that our Borland pascal won't compile well at all on
> other pascal compilers.  We looked into porting it to something like
> Delphi but as this code in question is a result of 15 or more years of
> legacy code it has been estimated at 1-2 years work for 2-3 people.  We
> have to do it eventually, just not right now, and we need a solution in
> the meantime.

Perhaps you could compile a simple test program that replicates the
problem on the fast windows nt based system and then port that code to
something like freepascal (which is pretty good for porting old bp
code to) and seeing what happens.  Doing the same with other 'problem
snippets' will help guide you as to which compiler system is a good
bet for the least problematic porting project, for when you really do
have to face that hurdle.

Infact, if you posted the problem code you might even get some
solution right here!



Sun, 31 Jul 2005 01:01:24 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:
> >Gautier
> >Is there a reason not considering compiling your >application with a
> >32-bit compiler, targetted at Win32 ?

MattD:

Quote:
> Unless we have missed something, we have looked into this a year or two
> ago.  It seems that our Borland pascal won't compile well at all on
> other pascal compilers.  We looked into porting it to something like
> Delphi but as this code in question is a result of 15 or more years of
> legacy code it has been estimated at 1-2 years work for 2-3 people.  We
> have to do it eventually, just not right now, and we need a solution in
> the meantime.

A problem that will come in porting to another Pascal environment is
that your program will be trapped again into a dialect attached to
a certain compiler (commercial or not). A solution (but, it depends!)
can be to port it to Ada. A larger work, but better for the long-time
and the spectrum of compilers.
________________________________________________________
Gautier  --  http://www.mysunrise.ch/users/gdm/gsoft.htm

NB: For a direct answer, e-mail address on the Web site!



Sun, 31 Jul 2005 05:54:20 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:

>> >Gautier
>> >Is there a reason not considering compiling your >application with a
>> >32-bit compiler, targetted at Win32 ?

> MattD:
>> Unless we have missed something, we have looked into this a year or two
>> ago.  It seems that our Borland pascal won't compile well at all on
>> other pascal compilers.  We looked into porting it to something like
>> Delphi but as this code in question is a result of 15 or more years of
>> legacy code it has been estimated at 1-2 years work for 2-3 people.  We
>> have to do it eventually, just not right now, and we need a solution in
>> the meantime.

> A problem that will come in porting to another Pascal environment is
> that your program will be trapped again into a dialect attached to
> a certain compiler (commercial or not). A solution (but, it depends!)
> can be to port it to Ada. A larger work, but better for the long-time
> and the spectrum of compilers.

Could be. How many Open Source ADA compilers at a decent level are there?


Sun, 31 Jul 2005 05:57:38 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.



Quote:

> Thank you for all your replies.  I am in the process of trying a few
of
> them to see how they go.  The super-short term resolution for this is
> that it DOES work in fullscreen mode (you can use alt+enter within a
dos
> window to do this).

> Just to answer a few of your questions....

> >Bill Leary
> >Have you tried adjusting idle sensitivity?
> >And perhaps tinkering with the other features >(performance, fast
> >paste, etc.)?
> I will try the idle sensitivity as it has me curious but I have tried
> all the other ms-dos options available in the propertes area.

> >Marco Van De Voort
> >Of course not, there are more Dos Pascal compilers, that >are
actually
> >stillmaintained.
> >Gautier
> >Is there a reason not considering compiling your >application with a
> >32-bit compiler, targetted at Win32 ?
> Unless we have missed something, we have looked into this a year or
two
> ago.  It seems that our Borland pascal won't compile well at all on
> other pascal compilers.  We looked into porting it to something like
> Delphi but as this code in question is a result of 15 or more years of
> legacy code it has been estimated at 1-2 years work for 2-3 people.
We
> have to do it eventually, just not right now, and we need a solution
in
> the meantime.

> >Femme Verbeek
> >For the mouse functions I use the following unit:
> >http://www.tank-cleaning.demon.nl/pasfiles/mouse.pas
> I will definately have a look at the code you have provided to see if
> it can relate to what we are doing.  Unfortunately the link you
> provided did not work.  Do you have the file handy you could possibly
> e-mail me at:

> >Frank Peel
> This idea worked, where did you get this theory from?  Any sources you
> can point me towards?

> >Markus Humm
> >But where's the technical problem with to fast CPUs >here? I can't
see
> >it. It's only a matter of scheduling, isn't >it?
> I'm not sure what you mean by this.  The customer in question is using
> NT 4.0 SP6.  We've been using this for years with no problems.  They
> have a fresh install of the OS, install our software on a 2.4Ghz
> computer and a slower 1Ghz machine and one has the problem and one
does
> not.  Testing here we slowed the CPU down a bit with a program like
Tame
> (I could not underclock it...stupid Dell's) and it started to work
> correctly.  I am 99% sure this problem is directly related to the CPU
> speed as we have ruled out everything else, although weather it is a
dos
> problem or a Borland compiler problem I am not sure.
> Running dos by itself the screen updates as usual.

> Again, thank you all for your time, I will let you know what I come up
> with as more people might run into this problem as 2.4Ghz plus
computers
> become more commonplace.

Perhaps you can make a little test program that reproduces the error on
the fast machine, or just put the code on a webpage somewhere for us to
testrun it. Send us the program or a link to the programcode. I am
willing to test it on my 1.2 Mhz PIII machine. So far I never had any
problems with TP/BP programs. I fact that is what I am still using.

--
Femme



Sun, 31 Jul 2005 07:47:05 GMT  
 Borland Pascal 7.0 on 1.8Ghz plus computers.

Quote:
> > A problem that will come in porting to another Pascal environment is
> > that your program will be trapped again into a dialect attached to
> > a certain compiler (commercial or not). A solution (but, it depends!)
> > can be to port it to Ada. A larger work, but better for the long-time
> > and the spectrum of compilers.

Marco:

Quote:
> Could be. How many Open Source ADA compilers at a decent level are there?

* GNAT is open-source and more than decent: it is a solid compiler, covering
  the full Ada95 standard
* ObjectAda (Aonix) is commercial but at reasonable prices, also full-featured
________________________________________________________
Gautier  --  http://www.mysunrise.ch/users/gdm/gsoft.htm

NB: For a direct answer, e-mail address on the Web site!



Sun, 31 Jul 2005 17:46:46 GMT  
 
 [ 25 post ]  Go to page: [1] [2]

 Relevant Pages 

1. Turbo / Borland Pascal 7.0 with a to fast computer

2. I would like to know the difference between Borland Pascal 7.0 and Turbo Pascal 7.0

3. Turbo Pascal 7.0 vs Borland Pascal 7.0

4. PASCAL/C program that WORK ON SOME COMPUTER(486) but not all computers

5. Need replacement disks for Borland Pascal 7.0

6. WTB: Borland Turbo Pascal 7.0 for DOS and Windows

7. Borland Pascal 7.0

8. Borland Pascal 7.0 Bug

9. Metawindow 5.0 with Borland pascal 7.0 app crashes under win2K

 

 
Powered by phpBB® Forum Software