Rounding

Is there anyway to round a real to the tenth decimal place? Or a way to simulate this?

Example: Turn 2.763 into 2.8? I need to turn it into a string afterwards but that is easily accomplished with R2S. I'd convert it to an Integer but I'd lose the tenth place.

Don't know what precision parameter means, but I'm sure this is the right way

EDIT: try with those parameters, if it doesn't work I have another solution

you could try this: I2R(R2I(X*10+0.5))/10 ...dunno but this seems pretty messy..

EDIT: fixed minor error and ..it will return 2.800 and not 2.8 :(

Presicion is what decimal place... I have no idea how I'd fill that in though.

And width... the hell...

Time to go play with it a bit. Be back in a few sir.


Edit: I have 0 idea what width does. Precision is how many decimal places out it rounds to. Strangely enough, it will ONLY round the tenths spot. It doesn't round anything farther.

Why would you want to round to an Nths place? That doesnt' make much since!

Anyways; wc3 has a precision of 3 decimal places with reals so I would suggest multiplying by 1000 then splitting the 3 decimals into strings using substr, cast them to integers check if the lowest decimal is > 5 and if it is increase Ndecimal+1 by 1 until you get to the rounding point.

Needlessly complicated operation though, why would you want it?

I am working on a Casting function that can take any spell casted, singe/aoe/instant, and all in 1 function calculate damage, create a castbar, and call a function that corresponds to the spell. I simply needed this as a "tracker" to show the player how time the cast had left. Etc.

Cure (3.0) > Cure (2.9) > Cure (2.8)


Still unsure if I want to add knockback to casting upon being dealt damage, but this is EXACTLY what I needed. Thank you for that R2SW function sir.

Oh. Count down using two integers and make them look like a psuedo real.

[R2SW] takes awhile. Not sure if mine's any faster though.

For example:


I haven't tested that; but the logic is there. The problem becomes casting a real to an integer; personally I'd ignore rounding and set Tick2 = Tenths digit, ignoring the rest. (Using string casting.) If that's even neccesary.

Well considering this system will be implemented for all casting, it will probably be called around 30-40 times a second so yes, speed it pretty important. I'll look into your method.

You guys are overcomplicating this. Multiply the number by 10, round it to an integer, then divide by 10.

((int(x*10))/10)

Unless wc3 truncates instead of rounding, which would be pretty lame.

[edit]

Nope; it was bad coding truncating it. That method works, just very strange casting:


Returns 1.9

Someone should benchmark; this should certainly be faster then my method; but who knows. Also; R2SW might actually be faster?

<Deaod> just tested, R2SW is about as fast as I2S()

Take it as you will, but as to whether the method that was just suggested is faster or not, I do not know.

If it's as fast as I2S() then it's certainly faster then my method.

in that case...

if [(x*10 - int(x*10)) > .5] then [(x*100 + 1)/100]
else use the truncating method ((int(x*10))/10)

Yes but is'nt calling multiple I2R and R2I and I2S going to be slower than R2SW anyway?

Yes. R2SW sounds like the fastest method.