Delay for a fraction of a second in Borland C++ 4.52 IDE

It has been many years since I did much C programming but I Have Been Given the job of adding some capabilities to an existing program. I am using the “XP Mode” in Windows 7. The program was written in Borland C++ version 4.52 using the 16 bit 80186 instruction set. The source code is 404 pages long when pasted into word.

The existing program uses the void sleep(unsigned seconds); in many different places but I want to delay the program for a fraction of a second. I have found recommendations to use

Code:

---


#include <windows.h>
Sleep(500);

---

After adding this the program contains:

Code:

---

#include <windows.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <setjmp.h>
#include <dos.h>
#include <alloc.h>
#include <mem.h>
#include <time.h>
#include <tcp.h>
#include <conio.h>

 
Lots of code

 
Sleep(500);

---

When I do this I get:

Warning COMMS2.C 423: Call to function 'Sleep' with no prototype in function DownloadAFile
Linking cdacs.exe:
Linker Error: Undefined symbol _Sleep in module COMMS2.C

Another recommendation I found was to use

Code:

---

delay(500);

---

To get 0.5 second delay. When I did this I didn't get any warnings or errors but my program just stopped. (I didn't wait 500 seconds).

I will be grateful to anyone who can help me get a fractional second delay.

You might try either of these headers "mapiwin.h" or "winbase.h".
These both have Sleep() function in Borland C++ 5.5.

Note: The "winbase.h" is supposed to be included by windows.h; so, it is not likely to fix the problem.

The "mapiwin.h" says its a WIN16 vs. WIN32 compatibility header, it is more likely to be correct.

Tim S.

Are you still using the Borland compiler? Note the Borland delay() function delays in mili-seconds so delay(500) would be 500 miliseconds (1/2 second) not 500 seconds.


Jim

Thank you for the reply Tim.

When I use #include <mapiwin.h> I get the following error:

Compiling COMMS2.C:
Fatal ..\..\BC45\INCLUDE\PSHPACK8.H 38: Error directive: 8-byte packing is not supported!!

When I use #include <winbase.h> I get a bunch of errors including:

Compiling COMMS2.C:
Error ..\..\BC45\INCLUDE\WINBASE.H 168: Declaration missing ;
Error ..\..\BC45\INCLUDE\WINBASE.H 169: Multiple declaration for 'DWORD'

Also I see that Winbase.h contains the message:

winbase.h -- This module defines the 32-Bit Windows Base APIs

But, I am compiling a for 16-bit application. So this won't work.

Jerry

Thank you for the reply Jim.

Yes I am still using the Borland C++ 4.52 IDE. I understand that delay should be in milliseconds. I added the comment about 500 seconds just in case it was in seconds. Since then I tried

Code:

---

delay(1);

---

and my program still stopped when it reached that line.

Jerry

What do you mean by stopped? It should stop/delay for 1 milisecond. I suggest post the smallest possible program that illustrates your problem. If the program doesn't compile provide the complete error messages exactly as they appear in your development environment.


Jim

Quote:

---

Originally Posted by jimblumberg

What do you mean by stopped? It should stop/delay for 1 milisecond. I suggest post the smallest possible program that illustrates your problem. If the program doesn't compile provide the complete error messages exactly as they appear in your development environment.


Jim

---

My compiled program is running on a picoflash single board computer from JK microsystems imbedded in a critical process controller designed by the company I am supporting. The source code is 404 pages long when pasted into Word. Most of that code is needed to get the system up and running so the smallest program I can use is way beyond the scope of this forum. But, I have added print statements that print to the processor terminal so I can see step by step what is happening.
The section of code of interest is inside a loop. The code of interest is:

Code:

---

printf("#124 time = %02d:%02d:%02d\n", tm.ti_hour, tm.ti_min, tm.ti_sec);
 
SendResp(RspPtr);
 
sleep(1);

---

This works fine. I can watch the terminal and see each printout that includes the time as the program goes through the loop over and over again.

When I change the code to

Code:

---

printf("#124 time = %02d:%02d:%02d\n", tm.ti_hour, tm.ti_min, tm.ti_sec);
 
SendResp(RspPtr);
 
delay(500);

---

There are no errors and no warnings.

After downloading the .exe to the board, everything runs fine until I get to the loop. Then I see the first printout and then nothing more appears on the terminal. I have to shut off the power to get the processor to run again.

I have tried

Code:

---

printf("#124 time = %02d:%02d:%02d\n", tm.ti_hour, tm.ti_min, tm.ti_sec);
 
SendResp(RspPtr);
 
delay(1);

---

and that doesn’t work either.

Jerry

Can you write a program, and run it on your board, just to test delay()? Something simple like:

Code:

---

int main()
{

  while(1)
  {
      delay(5000);
      printf("Do you see the delay?\n");
  }

  return(0);
}

---

It is possible that your board doesn't have the proper support for the delay() function and you may have to resort to something like:

Code:

---

unsigned int i;
for(i = 0; i < 60000; ++i)
{ /* Blank Body */ }

---

Edit: Also unless this board is running Windows you probably shouldn't be including the <windows.h> header file.


Jim

OK Jim

I created a small modification to your program. I had to add #includes. I added print before delay

Code:

---

//#include <windows.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <setjmp.h>
#include <dos.h>
#include <alloc.h>
#include <mem.h>
#include <time.h>
#include <tcp.h>
#include <conio.h>
 
int main()
{
                while(1)
                {
                                printf("Before delay\n");
                                delay(5000);
                                printf("Do you see the delay?\n");
                }
                return(0);
}

---

The “Build” gave me

Compiling TEST.CPP:
Warning TEST.CPP 23: Unreachable code in function main()
Linking test.exe:

After downloading I got

C:\>test
Before delay

then nothing. So, from what you said the picoflash probably doesn’t support delay().

Following up on your suggestion I wound up with the following program

Code:

---


//#include <windows.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <setjmp.h>
#include <dos.h>
#include <alloc.h>
#include <mem.h>
#include <time.h>
#include <tcp.h>
#include <conio.h>
 
int main()
{
                unsigned int i;
                unsigned int j;
                printf("Before delay\n");
                for (j = 0; j< 1000; ++j)
                {
                                for(i = 0; i < 30000; ++i)
                                                { /* Blank Body */ }
 
                }
                printf("Do you see the delay?\n");
                return(0);
}

---

That allowed me to determine that the inner loop takes about 12 msec for this configuration.

That should allow me to get my job done. But some have warned that this could cause overheating of the CPU.

Jerry

Maybe:

Code:

---

#include <stdio.h>
#include <time.h>

int main(void) {
  clock_t start, stop;
  start = clock();
  stop = clock();
  while(stop < start + 1500) {  //1.5 seconds
      stop = clock();
  }
  return 0;
}

---

Quote:

---

Originally Posted by Adak

Maybe:

Code:

---

#include <stdio.h>
#include <time.h>

int main(void) {
  clock_t start, stop;
  start = clock();
  stop = clock();
  while(stop < start + 1500) {  //1.5 seconds
      stop = clock();
  }
  return 0;
}

---

---

Thank you Adak

I tried this and it works.

Any idea what the difference in heating might be between this approach and the tight “for” loop?


Jerry

Quote:

---

Originally Posted by Jerry900

Any idea what the difference in heating might be between this approach and the tight “for” loop?

---

None, unless clock() somehow halts or slows down the CPU while it executes (which would be highly unusual in my experience). You can probably call a system-specific routine to put the CPU to sleep for a short time inside the busy loop.

I'm guessing that whatever time source delay depends on is not running (maybe an interrupt is disabled) , so it will effectively halt the program forever.

Does the NOP instruction (if it exists) use less power?

NOP := No Operation Instruction on many embedded and early CPUs.

Tim S.

Quote:

---

Originally Posted by Jerry900

Thank you Adak

I tried this and it works.

Any idea what the difference in heating might be between this approach and the tight “for” loop?


Jerry

---

No, my cpu's all have high end aftermarket coolers since they "Fold" for medical research, so much. Can you test it with yours, by running a free utility like CoreTemp?