This will be used to construct more complex gates like XOR, and to reset binary bits.

Example:
; Set bit 7 in $iFlag
$iFlag=$iFlag | &0040
; Reset bit 7 in &iFlag
$iFlag=$iFlag & ~&0040

Where "~" is a bitwise NOT.

EDIT Changed "boolean" to "bitwise" in the text and title to make things clearer

[ 07 December 2001: Message edited by: Richard Howarth ]

break on

if testbit(9,8) ; is bit 4 (from the right) of 1001 set ? (it is)
 ? setbit(9,4) ; set bit 3 on (answer should be 1101 or 13
endif

exit 1

function setbit($a,$b)
 $setbit = $a | $b
endfunction

function testbit($a,$b)
 $testbit = $a & $b
endfunction

function clearbit($a,$b)
 ; To be done
endfunction

Like some c programmers (k&r man myself), I carry around a header file of usefull macro's. And especially I use these three all the time ... wouldn't be caught dead without them :

#define SETBIT(x, b)   ((x) |= (b))
#define CLEARBIT(x, b) ((x) &= ~(b))
#define TESTBIT(x, b)  ((x) & (b))

So just need to work-out the clearbit math and we're SET (sort-of-speak) :

-Shawn

[ 05 December 2001: Message edited by: Shawn ]

function clearbit($a,$b)
 $clearbit = $a & (-1-$b)
endfunction

-Shawn

Yeah, I spotted the "options" thread and thought "no one has posted a bit reset function" - and then realised it couldn't be done elegantly.

BTW I sat down with pen and paper for half and hour trying to construct a NOT from AND and OR gates before some smart-arse in the office found a proof on the 'net that proved it was impossible

Ooh - just thought of another way to do it. Watch this space.

Nice one. My solution was a lot more complicated.

On an aesthetic note I might change (-1) to &FFFFFFFF. Not sure.

Of course having just re-invented basic addition and subtration, I guess we could just use this:

? 255 - (32+16)

lol

-Shawn

I've suggested these before and I would imagine Ruud has seen/thought of these as well. But at the risk of turning KiXtart into a mish-mash of C and VB I'll throw these out again anyway.

As a C programmer yourself, I know you already appreciate the "handyness" of autoincrement. And in a very small way, from a performance standpoint, are key targets for any kind of script engine optimizing

$count++ ; automatically increment by 1
$count-- ; automatically decrement by 1
$count += 10 ; same as $count = $count + 10
$count -= 10 ; same as $count = $count - 10

I can't count how many times i've koded this:

$index = $index + 1

all the while grinding my teeth and tensing my shoulders and biting my tounge ... but - Ruud obviously has a master plan for KiXtart and he certainly hasn't steered us wrong yet.

But I would love to get yours and the boards thoughts in regards to these ideas (borrowing/mixing syntax from other languages) or if KiXtart should stay true to it's BASIC roots ?

-Shawn

cj

Another way of doing the same job: ClearBit() Or Not_or()
I had hidden an other way of clearing bits, as function Not_Or() in UsrMustChgPwd()

Function Not_Or($Num,$NotBin)
; Binary Not Or
; Clears the bits set in $Num, that are also set in $NotBin
$Num = 0 + $Num
$NotBin = 0 + $NotBin
$Not_Or = ($Num | $NotBin) - $NotBin
EndFunction

My reason for calling the function Not_Or() was:
You use binary Or to add a bit to a Number
Not_Or() will do the opposite (Remove/clear a bit)

You rely on -1 is transformed into largest possible numer, because negative numbers are not alowed in KiXtart,
so i believe your way is just as safe as Richards: &FFFFFFFF instead of -1, as long as KiXtart dosn't allow
negative numbers.

I'm not shure if it would be a good idea to implement your C++ statements, or not:

$count++ ; automatically increment by 1
$count-- ; automatically decrement by 1
$count += 10 ; same as $count = $count + 10
$count -= 10 ; same as $count = $count - 10

What about:
++$count
--$count
In for-next loops, do something, increace/decrease after

Other KiX-users might want to implement the syntax of their favorite programming language.
I would prefere the pascal syntax:
Inc($Count) for increasing $count by one
Inc($Count, $x) for increasing $count by $x
Dec($Count) for decreasing $count by one
Dec($Count, $x) for decreasing $count by $x

I believe that the pascal-syntax would be easy to implement for Ruud. But i won't make it a
first priority wish to get these functions included (I could make an UDF to do it)

What about performance:
Of course it would be better for performance to have the funtions: Inc() and Dec() and operators (++, --, +=, -=)
implemeted in KiXtart, than executing UDF's.
I can make my pascal UDf's, but you can´t make C-UDF function ++(), it would be illegal.

-Erik

ps.
I think that Ruud has made a great effort to make the syntax for string- and object handling
very similar to the visual basic syntax, so if the the KiX syntax gets to be like VB, at least
we don't have to 'rewind' when going from VB or Java to KiXtart!!

cj-
You can't use ~ it is opted by Richard as boolean not, to keep this VB like ! is the not-operator.
Is ~ a boolean not in VB ?

On the subject of "borrowing" elements from other languages I'd say, it depends. Adding the pre- and post-fix operators like "++" because they execute faster is a great idea, but is no good if the downside is that KiXtart is made more complicated and the interpreter is slowed down for the rest of the script. The rumored facility for tokenisation or byte-code precompiling will remove a lot of that overhead of course.

Sometimes you have to sit back and say "does this really fit in" with KiXtarts purpose? For example I'd love to see associative arrays taken from AWK or PERL. These are arrays which are keyed by a variant subscript rather than a numeric subscript, but this is a very complex feature and can be achieved in other ways.

One thing we really need are static variables for UDFs. Until we have these UDFs will have to expose global variables to maintain state information.

[ 07 December 2001: Message edited by: Richard Howarth ]

Shawn,
You're getting dirty here. Wanting to beat Quake 3 or UT with Kixtart ?

[ 07 December 2001: Message edited by: Popovk ]

"Welcome to the KiXtart Coding Coven, wipe your feet and leave your sacrificial dagger by the door."

The problem is that KiXtart 4.00 already has a "NOT", which is a logical operator, i.e. it works like C "!". As KiXtart supports "|" for bitwise OR, and "&" for bitwise AND which also match the C convention it seemed sensible to continue that way with the bitwise NOT.

How does VB handle logical NOTs? You can't tell from the context if you want a logical or bitwise operation.

Here is a very simple "C" program that demonstrates the difference:

code:

---

#include 
main(){
	unsigned char c=170;
	printf("%d Logical NOT = %d, converted back=%d\n",c,!c,!(!c));
	printf("%d Bitwise NOT = %d, converted back=%d\n",c,~c,~(~c));
}

---

code:

---

170 Logical NOT = 0, converted back=1
170 Bitwise NOT = -171, converted back=170

---

Now, if the code read "NOT c", how would you know which NOT was required?

code:

---

Sub test()
  Dim btest As Boolean
  Dim ntest As Long
  btest = Not 170
  ntest = Not 170
  MsgBox "btest : " & btest & Chr(10) & "ntest : " & ntest
End Sub

---

You'll get this :

code:

---

btest : True
ntest : -171

---

VB don't know screen output, that's why the msgbox is here
(well, until you add a complete module of API command to support dos box output)

I know this is a bit off-topic but I'm intrigued now.

In your example "btest = NOT 170" is "True". Does this mean that 0=True, anything else=False?

To really have the 'NOT 170' working as logical, it should be of type boolean. After rereading doc about it, VB switch to logical when one of the values is of boolean type. As it's not the case in our example, VB use OR as bitwise.
Usually, i always use the Cbool() function when doing logical operations
For your purpose, it should be like this :

code:

---

btest=not cbool(170)

---