[RESOLVED] Zero fill right shift in VB6

[kavian]

Hi,
in Javascript we can convert a big number to 32 bit integer with Bitwise Operators, like this:

4066098149795 >>> 0; // result is: 3059087779

How i can do this in VB6?

[Elroy]

You're probably looking for this page or this page, or maybe even this page.

Or, to just explore what Donald as assembled, you can start here.

Enjoy,
Elroy

[LaVolpe]

This can be done a different way also, but is limited to positive whole numbers in the Double range.

First we need to get the value within the range 0 to 2^32. This part is a workaround of not being able to use MOD or integer division with doubles above the 2^31-1 range.

Dim d32 As Double, dNr As Double, shiftedValue As Long
dNr= 4066098149795
d32 = Int(dNr / 4294967296#)
d32 = dNr - (d32 * 4294967296#)

Now we can use long-standing unsigned to signed (Long) conversion

    If d32 <= 2147483647# Then  ' compare to max positive long value: 2^31-1
        shiftedValue = d32 
    Else
        shiftedValue= d32 - 4294967296# ' result will be a negative Long or 0
    End If

Note that shiftedValue can be a negative number due to the high bit being set.
The value 3059087779 as a long is -1235879517, regardless the hex value is: B655F9A3

Welcome to the forums

[dilettante]

DLL for Unsigned Long & Integer Arithmetic (see the second post for the version with shift and rotate operations) might do what you need.

Supports both Integer and Long.

However if this is about "big number" operations you may want another library that handles those.

[kavian]

This can be done a different way also, but is limited to positive whole numbers in the Double range.

First we need to get the value within the range 0 to 2^32. This part is a workaround of not being able to use MOD with doubles above the 2^31-1 range.

Dim d32 As Double, dNr As Double, shiftedValue As Long
dNr= 4066098149795
d32 = Int(dNr / 4294967296#)
d32 = dNr - (d32 * 4294967296#)

Now we can use long-standing unsigned to signed (Long) conversion

    If d32 <= 2147483647# Then  ' compare to max positive long value: 2^31-1
        shiftedValue = d32 
    Else
        shiftedValue= d32 - 4294967296# ' result will be a negative Long or 0
    End If

Note that shiftedValue can be a negative number due to the high bit being set.
The value 3059087779 as a long is -1235879517, regardless the hex value is: B655F9A3

Welcome to the forums

Thanks LaVolpe, your advice was always very useful.

[jpbro]

DLL for Unsigned Long & Integer Arithmetic (see the second post for the version with shift and rotate operations) might do what you need.

Supports both Integer and Long.

However if this is about "big number" operations you may want another library that handles those.

Yikes a compiled binary! Run for the hills!

[dilettante]

They are allowed in UtilityBank submissions.

[jpbro]

Ah, good to know Kind of weird tough based on the admin feedback that binaries are just too dangerous to host....I guess everywhere but there for some reason. I'm sure it's a good reason too

[dilettante]

Some things are impractical to post without the ability. One obvious case is an ActiveX DLL or OCX, where you really need to post a precompiled interface compatibility library (at least one with all members stubbed out) along with the source.

Of course you can't just post binaries there, matching source is required and submissions are scrutinized more to ensure compliance.