[meTaLiCroSS]

PHP Code:

                    static Float:flBlend, bool:bSum
                    server_print("prepare sum=%d blend=%f", _:bSum, flBlend)
                    
                    if(bSum)
                        flBlend += 0.1
                    else
                        flBlend -= 0.1
                    
                    if(flBlend >= 1.0)
                        bSum = false
                    if(flBlend <= 0.0)
                        bSum = true
                        
                    server_print("modified sum=%d blend=%f", _:bSum, flBlend) 


*no need to show more code, so stop those "this is an xy problem" comments and that shit*



so, you can guess that on the green block, the "if(flBlend >= 1.0)" statement works and sum is set to 0 (false)

but, what happens on the red block?

flBlend is 0.1, and bSum is false, so it should decrease it in 0.1 resulting 0.0, but...

PHP Code:

                              if(bSum)
                        flBlend += 0.1
                    else
                        flBlend -= 0.1 /* = 0.0 */
                    
                    if(flBlend >= 1.0)
                        bSum = false
                    if(flBlend /* 0.0 */ <= 0.0) //lolwat
                        bSum = true // this isn't called 


the "if(flBlend <= 0.0)" meaning it's "if(0.0<= 0.0)" isn't validated (because it said that sum is 0 and blend is also 0.0), so wtf amxx, is this "correct" ?

[Xunfop]

IEEE 754

[claudiuhks]

PHP Code:

new Float:Value = 0.0, Float:Other = 0.0;

if (Value <= 0.0)
{
  server_print("0");
}

if (Value <= Other)
{
  server_print("1");
} 


Output:

PHP Code:

0
1 


I don't really understand what you are trying to explain into the first message, but I am sure stock Float: operator <= (Float: fA, Float: fB) and native floatcmp(Float: fA, Float: fB) works properly.

Try to use fl1 = fl1 + fl2 instead of fl1 += fl2.

[hornet]

Not sure why that is happening. But can be fixed by changing <= 0.0 to < 0.1

[meTaLiCroSS]

Originally Posted by hornet Not sure why that is, but can be fixed by changing <= 0.0 to < 0.1

Yeah, just did it after posting this, but I'm still thinking why that code failed

Originally Posted by claudiuhks I don't really understand what you are trying to explain into the first message, but I am sure stock Float: operator <= (Float: fA, Float: fB) and native floatcmp(Float: fA, Float: fB) works properly.

Buddy, i know that... can you please see again the post?

[claudiuhks]

Originally Posted by hornet Not sure why that is, but can be fixed by changing <= 0.0 to < 0.1

Real numbers usually work with more decimals. 0.0 is lower than 0.1 but lower than 0.0005 aswell.
If a number is real and you define it positive, you can't say it's higher than -1, cause it can be -0.002. You need to say that it's higher or equal to 0.

Originally Posted by meTaLiCroSS Buddy, i know that... can you please see again the post?

My message has been updated. OK, will take a look again.

[meTaLiCroSS]

Originally Posted by claudiuhks Real numbers usually work with more decimals. 0.0 is lower than 0.1 but lower than 0.0005 aswell. If a number is real and you define it positive, you can't say it's higher than -1, cause it can be -0.002. You need to say that it's higher or equal to 0. My message has been updated. OK, will take a look again.

You're right. But, strange thing is that no more decimals are added to the num.

The code that I gave above shows everything, numbers calculated are 0.0, 0.1, 0.2, 0.3, ... 0.9, 1.0, there's no other minor number like 0.05, or anything else, it's just the same number plus 0.1 or -0.1, so I'm still not understanding, you can see the printed screen above, flBlend = 0.0, and bSum = false, so the "flBlend <= 0.0" check was avoided

[jimaway]

0.0-0.1 = -0.099999?

[Black Rose]

Could this have an effect?
I did set it to 65536 and I still got the same -0.099999 though.
core.ini

; Plugin optimization flags - add these up to get what you want
; lowering these may stop crashes on very old CPUs
; set 65536 to disable optimizer, NOT 0!
;-------------
; 1 - float arithmetic
; 2 - float comparisons
; 4 - float rounding
optimizer	7

[.Dare Devil.]

I find that it is useless to use == and =something operator for floats check.

When you do something like this
a = 0.0
if(a) // called

But when you use +-/* then there will be no accurate float.
=(something) check will be useless.