Announcement

**Daniel Raymer** · 20 May 2008, 10:50 AM

Still no Joy

I still haven't figured this one out. I've made a lot of improvements to the code, including adding a Contools pulldown menu and using the GFX line draw routine instead of the PB/CC version, but the basic problem is still there. Sometimes it just doesn't draw some of the lines.

Weird.

**Paul D. Elliott** · 20 May 2008, 12:27 PM

Daniel,

I ran it a few times and multiple flips within each run and it looked like
it had the same lines each time. Or at least I couldn't see anything missing.
Also ran it with the debugger and a trace file and nothing showed up as
an error. Haven't tried it with an "on error goto" to catch an error # yet
as I'm having a little trouble figuring out just where it's supposed to end
normally. It would have helped to have the program in between code &
/code brackets. Maybe you could post your revised code that way.

I'm running on Windows 2000 if it makes any difference.

**Guy Dombrowski** · 20 May 2008, 02:02 PM

Missing lines

Daniel,

Just add SLEEP 50 after GRAPHIC ATTACH and it should solve your problem.
There seem to be a timing problem with the graphic print and the sleep delay did it for me.

**Knuth Konrad** · 21 May 2008, 06:35 AM

When posting source code, especially as long as yours, please put the code in code tags (without the space after the leading "[")

[ code]
<Your source code here>
[ /code]

Example:

Code:

Function PbMain()

   ' Proper indented source code
   Local i As Long, s As String

End Function

This will preserve your source code formating and helps others to easier read your code.

**Michael Mattias** · 21 May 2008, 06:48 AM

Note that you can edit your own posts to insert those code tags... ie. no need to re-post in a new message.

**Daniel Raymer** · 21 May 2008, 12:32 PM

cannot edit - sorry

The EDIT button is gone from my first post - sorry. I missed the meeting where the [code] thingy was explained. I'll use it now.

**Michael Mattias** · 21 May 2008, 12:59 PM

IIRC, you can only 'edit' within 'x' days of post. (X=15 ??)

However, if you send a note to [email protected] and ask nicely if someone with Super-User board privileges would insert the code tags for you, I'll bet they will.

MCM

**Daniel Raymer** · 21 May 2008, 01:14 PM

Still no joy

Thanks for the suggestions.

Adding Sleep seems to help but the problem still occurs. Maybe I have a memory or CPU hardware problem? Maybe I have evil gremlins under my desk?

Such suspicions are the last resort of the incompetent.

**Paul D. Elliott** · 21 May 2008, 02:04 PM

Just curious but are you using the latest version of the compiler?
Or maybe it's something to do with your computer hardware or other
software.

Have you tried running ( and maybe even compiling ) on a totally different
computer & OS ?

**Rodney Hicks** · 21 May 2008, 05:16 PM

I copied to PBCC 4.04.0042 and ran it as is and did not see any missing lines after several instances and samplings. Seems like there is nothing wrong with it. Running Vista Home Premium.
I ran it from the IDE and from the EXE. Seems to work like a charm.

Rod

**Fred Buffington** · 22 May 2008, 01:07 AM

Running WinME with a disk that needs replacing, cant run scandisk or defrag.

That said, the program seems to work fine here after several flip options.

**Gary Barnes** · 22 May 2008, 07:28 AM

Hi Dan,
I hope that you don't mind, but I did a slight rewrite of your code.
Truth to tell, I had to do this in order to understand it properly.

Code:

#COMPILE EXE
#DIM ALL
#INCLUDE "win32api.inc"

FUNCTION PBMAIN () AS LONG
LOCAL filename AS STRING
LOCAL filehandle, norecs, gwin, penup, firstplot AS LONG
LOCAL databuffer() AS STRING
LOCAL x,y AS SINGLE

filename = CURDIR$+"\hlp.plt"
filehandle = FREEFILE
OPEN filename FOR INPUT AS #filehandle
FILESCAN filehandle, RECORDS TO norecs
REDIM databuffer(1 TO norecs)
LINE INPUT #filehandle, databuffer() TO norecs
CLOSE

GRAPHIC WINDOW "DANPLOT",0,0,800,800 TO gwin
GRAPHIC ATTACH gwin,0,REDRAW
GRAPHIC SCALE (-50,50)-(50,-50)
GRAPHIC COLOR %BLACK, %WHITE
GRAPHIC CLEAR


firstplot = %true
FOR norecs = 1 TO UBOUND(databuffer)
    penup = %true
    IF RTRIM$(databuffer(norecs)) <> "" THEN
       penup = %false
       x = VAL(LEFT$(databuffer(norecs),15))
       y = VAL(MID$(databuffer(norecs),16))

       IF ISTRUE firstplot THEN
          GRAPHIC SET POS (x,y)
          firstplot = %false
       END IF

    END IF

    IF ISTRUE penup THEN
       GRAPHIC SET POS (x,y)
    ELSE
       GRAPHIC LINE -(x,y)
    END IF
NEXT norecs

GRAPHIC REDRAW

WHILE iswindow(gwin):SLEEP 100:WEND

END FUNCTION

It isn't as fancy as your program, but I hope that it will help.
Best Regards

**Robert DeBolt** · 22 May 2008, 07:54 PM

Thanks, Gary!

Just what I've been looking for.

**Can Le** · 20 Jul 2008, 03:16 AM

Hi,

Please help me how to define Data Type for Num to be used from both 4 digits to a 12 digits long ?

Thank you for your help

Can Le

**Rodney Hicks** · 20 Jul 2008, 03:53 AM

Data Type Integer
Size 16 bits (2 bytes), signed
Decimal Range -32,768 to 32,767
Binary Range -2^15 to 2^15-1

Data Type Long-integer
Size 32 bits (4 bytes), signed
Decimal Range -2,147,483,648 to 2,147,483,647
Binary Range -2^31 to 2^31-1

Check your help file under Data Handling. There is a complete chart there.

The size of the number depends on the number of bytes used for storing.
4 bytes for a signed Long integer results in a number from -2,147,483,648 to 2,147,483,647, as shown above.

To store a 12 digit integer use the QUAD integer.

Rod

**Dale Yarker** · 20 Jul 2008, 06:17 AM

Can,

SINGLE is the least precise floating-point type. DOUBLE or EXTENDED floating-point types would have been okay for the numbers you gave.

But, if you are working with integers, it is much better if you use an integer type. Unsigned types are byte (8 bits), word (16 bits), and dword (32 bits). Signed (plus or minus) types are integer (15 bits and sign bit), long (31 bits and sign bit), and quad (63 bits and sign bit).

On 32 bit computers it is almost always best to use long or dword. Byte, word, and integer make for slower running code.

Once you select the correct type for job, and understand the differences between the type, you will have no problems with PB.

Cheers,

**John Gleason** · 20 Jul 2008, 07:19 AM

Originally posted by Can Le View Post

MY problem is:
My calculation will yield an integer number Num from 4 to 12 digits,
My question:
How to define Data Type for Num to be used from both 4 digits to a 12 digits long ?

Will PB Windows version has problem with under flow and over flow too ? If yes, I may say good bye to PB Windows and go back to PB DOS as Data Type for Perl, I think Perl is having the best Data Type!

No need to go back to DOS.

What you need here is the QUAD data type and you'll get no overflow or underflow. Example:

Code:

#COMPILE EXE
#DIM ALL

FUNCTION PBMAIN () AS LONG

    LOCAL CanLeData AS QUAD
    
    CanLeData = 1234
    PRINT CanLeData
    CanLeData = 123456789012
    PRINT CanLeData
    WAITKEY$

END FUNCTION

**Can Le** · 11 Sep 2008, 09:36 PM

Thank you all for your helps!

I used QUAD and EXT and both worked for my ranges on integer and float, however, I encountered problem when formula yields 13 digits which gave me zero on value! I came back this thread and and used DOUBLE, it worls for me with 1.234567891131517.25 for 1.234567891131517 E17 !

Thank you

Announcement

Why I hate Windows - even with PBCC

Why I hate Windows - even with PBCC

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment