21 replies to this topic

[Richard Bassett]

Well, my Friday's been spoiled by ProIV and not for the first time.
So I thought I'd share this "feature" with you..

#v1 = 0
#v2 = - #v1
IF #v1 = #v2 THEN UMSG("Yes, #v1 equals #v2", -1) ;
IF #v1 <> - #v1 THEN UMSG("No, #v1 does not equal -#v1", -1)

You have to try it to believe it..

[Joseph Bove]

Richard,

I tried it. I was told that they are equal.

By any chance, are you doing this in a function where you have mapped an output parameter from a global function to 0?

If so, you can end up changing the value of 0....


Regards,

Joseph

[Richard Bassett]

Joseph,

No, that's the entire (test) function.

The point is not that #v1 = #v2, it does, as you would expect.
The point is that #v1 <> -#v1 is also true.. Which is WRONG, WRONG, WRONG..

I hit this on V4.6 on Linux but it's so fundamental I'd bet it's universal unless it's been subsequently fixed.

You did see the second message too? Or didn't you?!

On a historical note..

Edited by Richard Bassett, 27 August 2005 - 10:59 AM.

[Rob Donovan]

Hi,

I only see the first UMSG....

Rob D.

[Bob Filipiak]

Guys,

FWIW, I have always been under the assumption (and I do know how spell ass u me) that zero is an UNSIGNED value, i.e. neither positive nor negative.

But then again, I can be wrong.

Bob Filipiak

[Richard Bassett]

Hi,  I only see the first UMSG....

VERRY interesting Rob..
I know you'd have been careful but would you mind just sanity checking that you did the code exactly as shown.
And maybe just try the following more obvious one-liner that I should have posted to start with..

IF 0 = -0 THEN UMSG("Equal",-1) ELSE UMSG("Not equal",-1) ;

I see "not equal". Let me know your mileage and what platform you're on.

zero is .. neither positive nor negative .. but then again, I can be wrong.

Not in this universe you can't Bob.

[Rob Donovan]

I Cut & Pasted it from your post, into ProIV.

I get the 'Equal' UMSG from the 2nd example.

I'm using Windows XP (SP2) and ProIV 5.5r333

Rob D.

[Bob Filipiak]

Gents,

Maybe it is due to some defective math processors??

The computer is the only entity that thinks 2 + 2 = 5

Then, perhaps we're stuck in the



TWILIGHT ZONE


What do I know??

Bob Filipiak

PS> I spent the last 3 hours trying to find a silly syntax error. Doesn't pay to work when you are tired. I need a vacation

[Richard Bassett]

Rob,

Thanks for that. I'll try it on a a couple of other systems and see what I get.
Hopefully your result indicates that it's been fixed during the last 5 years

Bob,

There are certainly defective math processors but FWIW that is unlikely to be relevant to ProIV because ProIV's "numeric" data type is a decimal representation implemented in software. ProIV doesn't have the integer or binary-floating-point data types that are typically implemented by hardware.

I was definitely in the Twilight Zone Friday afternoon though.

[Mike Schoen]

This appears to be a 4.6 feature.
I tried the same code under 4.6 and 5.5, and 5.5 got it right.

[Richard Bassett]

Thanks for all the cross-checks folks - and apologies for not checking it on a 5.5 system on Friday (dial-up access issue).
I was in a bad frame of mind Friday PM though and I admit I wasn't really expecting V5.5 would make any difference

[Rob Donovan]

.... so it was a bug....

They still havent fixed the other one though

[Donald Miller]

Shouldn't that be

#v2 = -1 * #v1

Or does Pro IV see that as exactly the same as

#v2 = -#v1