PureBasic Forums - English

Are you looking for CompilerError "my error"?

I know this doesn't have the syntactic sugar, but you may be able to use this: Structure Car name.s speed.i EndStructure Global DefaultCar.Car DefaultCar\name = "DefaultCar" DefaultCar\speed = 100 Procedure DoCarStuff() Protected car.Car car = DefaultCar Debug car\name Debug car\speed EndP...

Alpha returns the value of the alpha channel only, so it's always between 0 and 255.

There really isn't any excuse anymore for not using HTTPS on a site
taking user credentials. So +1.

Use Delay().

skinkairewalker wrote:why it uses a lot of cpu ?

Because you're spending all your time in an endless loop.
Delay() passes on execution time to other processes.

That might work:

No C code?

chris319 wrote: Here is my attempt at converting it to C, but it isn't working correctly. I suspect PB is doing some type conversions which I'm not getting right in C.

Code: Select all

#define Kr 0.3333
unsigned char V, Y, Rd;
float Vtemp;
Vtemp = (float)(V) - 128;
Rd= Y+((unsigned char)(Vtemp/Kr));

Your C example without the casting clutter is:

Well, (Y + Vt) / #Kr is something different than Y + (Vt / #Kr),
which one is it?

Ask MS for more info about this wierd system :) Actually it's all documented. :) https://msdn.microsoft.com/en-us/library/windows/desktop/bb761389(v=vs.85).aspx TL;DR: An image list is not actually a list but a "sprite map" (i.e. one large image containing small images). So every time an ...

I can't run PB right now so I'm going on a long shot here and say that since you create different images, there is a lot more overhead with copying that data into the gadgets and keeping track of updates (though the increase seems a lot). Maybe you could try to see if it makes any difference to disa...

Should be 2^53 (9'007'199'254'740'992)

You could use double (.d) instead of float (.f) to get more precision. Remember that floating point types are not about accuracy, they are just approximations that are supposed to be fast and accurate *enough* for most computations. If you need accurate integers, use integer types (e.g. .q) or use a...

Single precision (32 bit) floats use 23 bits for the mantissa so the largest integer number
they can accurately represent is 2^(23+1) = 16777216.

Well, that was only 307 days too late. Still counts!