Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.clipper.visual-objects > #953 > unrolled thread

What's wrong with this cast?

Started byD.J.W. van Kooten <public@ic2remove.this.com>
First post2011-04-26 21:33 +0200
Last post2011-04-28 22:26 +0000
Articles 7 — 4 participants

Back to article view | Back to comp.lang.clipper.visual-objects


Contents

  What's wrong with this cast? D.J.W. van Kooten <public@ic2remove.this.com> - 2011-04-26 21:33 +0200
    Re: What's wrong with this cast? "Jamal" <vodotnet_nospam@yahoo.com> - 2011-04-26 16:38 -0400
      Re: What's wrong with this cast? D.J.W. van Kooten <public@ic2remove.this.com> - 2011-04-26 22:46 +0200
        Re: What's wrong with this cast? "Geoff Schaller" <geoffx@softwarexobjectives.com.au> - 2011-04-26 22:17 +0000
    Re: What's wrong with this cast? Robert van der Hulst <E-55525A53-4141-E@heliks.nl> - 2011-04-27 11:42 +0200
      Re: What's wrong with this cast? D.J.W. van Kooten <public@ic2remove.this.com> - 2011-04-27 17:41 +0200
        Re: What's wrong with this cast? "Geoff Schaller" <geoffx@softwarexobjectives.com.au> - 2011-04-28 22:26 +0000

#953 — What's wrong with this cast?

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2011-04-26 21:33 +0200
SubjectWhat's wrong with this cast?
Message-ID<s87er6d3aomad9n4sjrh8ucetm5usn9aji@4ax.com>
In a method where a subtraction is supposed never to yield <0 it did
nevertheless as the variable was a dword it gave a run time error.

I decided to work with floats inside (realized later that int was
better, but that is not the issue now ) and after checking for non
zero values, I assigned :

nStockDW:=DWORD(_CAST,nStock)

where nStock=float 
nStockDW = dword

However: the first value of nStock (1,00) did not yeild dword 1 but a
huge number like 43423454554645, and a run time error after use of
course.

What is wrong with this conversion?

Dick

[toc] | [next] | [standalone]


#954

From"Jamal" <vodotnet_nospam@yahoo.com>
Date2011-04-26 16:38 -0400
Message-ID<ip7ag4$a27$1@dont-email.me>
In reply to#953
Try:

nStockDW := DWORD(nStock)

Jamal

"D.J.W. van Kooten"  wrote in message 
news:s87er6d3aomad9n4sjrh8ucetm5usn9aji@4ax.com...

In a method where a subtraction is supposed never to yield <0 it did
nevertheless as the variable was a dword it gave a run time error.

I decided to work with floats inside (realized later that int was
better, but that is not the issue now ) and after checking for non
zero values, I assigned :

nStockDW:=DWORD(_CAST,nStock)

where nStock=float
nStockDW = dword

However: the first value of nStock (1,00) did not yeild dword 1 but a
huge number like 43423454554645, and a run time error after use of
course.

What is wrong with this conversion?

Dick 

[toc] | [prev] | [next] | [standalone]


#955

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2011-04-26 22:46 +0200
Message-ID<kober6tt0kkkb1ee0sa6ui93f5d19hme1n@4ax.com>
In reply to#954
On Tue, 26 Apr 2011 16:38:25 -0400, "Jamal"
<vodotnet_nospam@yahoo.com> wrote:

Hello Jamal,

>Try:
>
>nStockDW := DWORD(nStock)

Ah! I always thought I needed to add Cast. Indeed this works, thanks.

Dick

[toc] | [prev] | [next] | [standalone]


#957

From"Geoff Schaller" <geoffx@softwarexobjectives.com.au>
Date2011-04-26 22:17 +0000
Message-ID<4db7445e$0$77723$c30e37c6@exi-reader.telstra.net>
In reply to#955
The difference is that CAST assumes you know what you're doing and that 
you know the value is correct for the new data type. DWORD() performs a 
type conversion and therefore a correction if necessary.

To get a huge number, it must have been negative or a fraction.
For example 0.999999999

Geoff



"D.J.W. van Kooten" <public@ic2remove.this.com> wrote in message 
news:kober6tt0kkkb1ee0sa6ui93f5d19hme1n@4ax.com:

> On Tue, 26 Apr 2011 16:38:25 -0400, "Jamal"
> <vodotnet_nospam@yahoo.com> wrote:
>
> Hello Jamal,
>
>
> >Try:
> >
> >nStockDW := DWORD(nStock)
>
>
> Ah! I always thought I needed to add Cast. Indeed this works, thanks.
>
> Dick

[toc] | [prev] | [next] | [standalone]


#963

FromRobert van der Hulst <E-55525A53-4141-E@heliks.nl>
Date2011-04-27 11:42 +0200
Message-ID<19010142267.20110427114237@heliks.nl>
In reply to#953
Hi D.J.W.,
 On Tue, 26 Apr 2011, at 21:33:11 [GMT +0200] (which was 21:33 where I live) 
you wrote about: 'What's wrong with this cast?'

> In a method where a subtraction is supposed never to yield <0 it did
> nevertheless as the variable was a dword it gave a run time error.

> I decided to work with floats inside (realized later that int was
> better, but that is not the issue now ) and after checking for non
> zero values, I assigned :

> nStockDW:=DWORD(_CAST,nStock)

> where nStock=float 
> nStockDW = dword

> However: the first value of nStock (1,00) did not yeild dword 1 but a
> huge number like 43423454554645, and a run time error after use of
> course.

> What is wrong with this conversion?

if nStock is a float, then you are casting the address of the float to a DWORD.
Floats are a reference type in VO. You can easily see that by doing the cast
twice  and running CollectForced() in between:

nStockDW:=DWORD(_CAST,nStock)
CollectForced()
nStockDW:=DWORD(_CAST,nStock)

You will see that the second cast returns a different value.

Like Jamal said: nStockDW := DWORD(nStock) will do the trick.

However when nStock is negative then this could throw a runtime error (depending
on the overflow compiler setting)



-- 
Robert van der Hulst
AKA Mr. Data
www.heliks.nl

[toc] | [prev] | [next] | [standalone]


#966

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2011-04-27 17:41 +0200
Message-ID<c0egr6db2rfmogg30b095d9rl0eh8qf3kg@4ax.com>
In reply to#963
On Wed, 27 Apr 2011 11:42:37 +0200, Robert van der Hulst
<E-55525A53-4141-E@heliks.nl> wrote:

Hello Robert,

>if nStock is a float, then you are casting the address of the float to a DWORD.
>Floats are a reference type in VO. You can easily see that by doing the cast
>twice  and running CollectForced() in between:
>
>nStockDW:=DWORD(_CAST,nStock)
>CollectForced()
>nStockDW:=DWORD(_CAST,nStock)
>
>You will see that the second cast returns a different value.

Thanks for the explanation. It is no longer your responsibility, but
do you agree this information should  have been in the help, subject:

"Type Casting and Conversion Operators"

It says:

Their general format is:
<idType>(_CAST, <Value>)
where <idType> is one of the data types listed above.

DWORD and FLOAT are in that list without mentioning that it doesn't
work. 

Even better would have been to add DWORD(_CAST,nFloat) under the
option to create random numbers. Random is not working either and this
cast would do a great job I now understand <g>.

>However when nStock is negative then this could throw a runtime error (depending
>on the overflow compiler setting)

That's why I did this in the first place. Any negative float is set to
0 if it's <0 before the dword conversion.

Dick

[toc] | [prev] | [next] | [standalone]


#982

From"Geoff Schaller" <geoffx@softwarexobjectives.com.au>
Date2011-04-28 22:26 +0000
Message-ID<4db9e990$0$77723$c30e37c6@exi-reader.telstra.net>
In reply to#966
<rofl>

Dick, the problem is that dword(_cast, fFloat) DOES work, provided you 
understand the context. It doesn't convert the contents of the float, 
just its pointer, and that could be quite valid. So the help 
instructions are correct when read in the context they are written. Over 
the years (and at conferences) there have been numerous warnings about 
casting vs conversion functions so I am a little surprised you got 
caught - you were there. Robert covered this very point in his VO Voodoo 
sessions.

Unfortunately you can't make help instructions idiot proof because the 
world can always make a better class of idiot <g>.

Geoff



"D.J.W. van Kooten" <public@ic2remove.this.com> wrote in message 
news:c0egr6db2rfmogg30b095d9rl0eh8qf3kg@4ax.com:

> On Wed, 27 Apr 2011 11:42:37 +0200, Robert van der Hulst
> <E-55525A53-4141-E@heliks.nl> wrote:
>
> Hello Robert,
>
>
> >if nStock is a float, then you are casting the address of the float to a DWORD.
> >Floats are a reference type in VO. You can easily see that by doing the cast
> >twice  and running CollectForced() in between:
> >
> >nStockDW:=DWORD(_CAST,nStock)
> >CollectForced()
> >nStockDW:=DWORD(_CAST,nStock)
> >
> >You will see that the second cast returns a different value.
>
>
> Thanks for the explanation. It is no longer your responsibility, but
> do you agree this information should  have been in the help, subject:
>
> "Type Casting and Conversion Operators"
>
> It says:
>
> Their general format is:
> <idType>(_CAST, <Value>)
> where <idType> is one of the data types listed above.
>
> DWORD and FLOAT are in that list without mentioning that it doesn't
> work.
>
> Even better would have been to add DWORD(_CAST,nFloat) under the
> option to create random numbers. Random is not working either and this
> cast would do a great job I now understand <g>.
>
>
> >However when nStock is negative then this could throw a runtime error (depending
> >on the overflow compiler setting)
>
>
> That's why I did this in the first place. Any negative float is set to
> 0 if it's <0 before the dword conversion.
>
> Dick

[toc] | [prev] | [standalone]


Back to top | Article view | comp.lang.clipper.visual-objects


csiph-web