Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Nadaka]

Thank Ogd for GUI builders

Yes... Though I have come to loathe visual studios gui builder. It may be one of the best in the business, but it is still terribly buggy.

If there was a decent gui builder for eclipse this would be a java project instead of a c# project.

[GlyphGryph]

What's so silly about programming a roguelike in Ruby?

I guess you could use the c++ visual libraries without too much trouble, but its not exactly super fun and last I checked ruby visual libraries were crap. Plus there's the whole hassle of getting it to run on other people's computers, though that might just be personal experience - there's probably a way to package it up to make it easy, but hell if I know how to do it.

I guess it's an option, at least.

[Siquo]

I have no idea as to what you're trying to do...

Gatling gun inherits armpart's values, and armpart inherits frameparts values, but never will armpart inherit gatlingguns' values. You'll also need to call "set" on the plarm, btw.

[dreadmullet]

I would like to instantiate an arbitrary class in C++. I know how to do this in C#:

Code: [Select]

T Instantiate<T>() where T : new(){
     
     T instance = new T();
     
     return instance;
     
}

The above code creates an instance of the type passed to it. If you wanted to make a Ball, it would work like this:

Code: [Select]

Ball b = Instantiate<Ball>();


How do I do this with C++? I am aware of templates, but there has to be a better way.

[Siquo]

How do I do this with C++? I am aware of templates, but there has to be a better way.

Not that I (intermediate C++ experience) know of. And what is so bad about templating that function? Unless you're in some kind of 64k demo contest, the multiple copies won't be that much of a strain on your resources...

[dreadmullet]

So apparently I can just do "new T", like this:

Code: (C++) [Select]

template<typename T>
T Instantiate(){
     
     return *(new T);
     
}

That's useful, but it's just part of what I'm doing. What if I want to compare if a variable is of a certain type? In C# it looks like this:

Code: (C#) [Select]

bool Test<T>(){
     
     Ball ball = new Ball();
     
     return ball is T;
     
}


[Siquo]

So apparently I can just do "new T", like this:

Yep, that's what I meant. Now you can cast the new T into Ball to check the class, OR overload&specialise the template, but there's a lot of "buts" and "only-ifs" connected to that, and I know I don't know all of them, function template specialisation is advanced. 

[dreadmullet]

Your first link got me going in the right direction. Thanks!

I think these two lines are basically the same:

Code: (C#) [Select]

if(Ball is T)


Code: (C++) [Select]

if(typeid(Ball) == typeid(T))


[Siquo]

I do not know what "is" does, but keep in mind that typeid does an exact match, and that derived classes such as class Football : public Ball{} will not match. For that you'd need the second example in that first link with dynamic_cast.

[Nadaka]

Does anyone know of any good OCR libraries for c#?

I found this: http://www.pixel-technology.com/freeware/tessnet2/

but I don't know if will be able to pull data from diablo 3 screenshots yet.

[kaijyuu]

Venting about c++ ahead! 'Cause I feel like it.

Two things I loathe about it:

1) No easy way to maintain local scope and eliminate redundancy. If you've a function that does the same identical block of code multiple times throughout, the normal way to eliminate redundancy is to turn that block of code into a function call. This means you lose local scope. The ways around this are 1) Pass your local variables as arguments (sloppy), 2) Turn your locals into globals (retch), or 3) use macros (potentially long compile times, ahoy!). I've started gravitating toward globals and I feel dirty.

2) Switch statements are horrible. Note that state machines are my bread and butter when it comes to coding object behavior in games. The normal response to people using switch statements is "use polymorphism!" but that's not an option when I want my enemy to choose from 100 different possible behaviors. I'm not gonna have 100 identical classes save their behavior subroutine. Oh how I wish for a switch statement without a default case, so it didn't have to be an if/else chain on the assembly level (which gets much slower the more cases you have. Not good when you want hundreds if not thousands of objects active.).



I guess I was spoiled by assembly where neither of those were problems.

[Levi]

Yeah, I get that.  c++ needs closures badly.  Then you wouldn't need to worry about losing local scope.

[Nadaka]

Kaijyuu, I really would pass local variables to the function way way before using globals.
Also C++11 has anonymouse functions you should be able to define in scope, but I have not really looked at them.

You can also create an alternative to the switch statement that can match an arbitrarily large number of cases much faster than O(N).

Step 1: create a dictionary/vector using your favorite collections library. The keys are cases to be switched, the values are function pointers/objects to the method to execute in each case.

Step 2: instead of doing a switch, pull the function from the dictionary/vector using the value and then execute that function.

There you have a switch with execution time that scales with the efficiency of your dictionary, most likely O(log(2)N).


Edit: its a bit of a pet peeve of mine when people call anonymous functions closures. They are not really the same thing even though an anonymous function can be a closure and a closure can be an anonymous function. This pet peeve is further annoyed because always assume that closure refers to syntactic closure.

http://en.wikipedia.org/wiki/Syntactic_closure

[Levi]

Edit: its a bit of a pet peeve of mine when people call anonymous functions closures. They are not really the same thing even though an anonymous function can be a closure and a closure can be an anonymous function. This pet peeve is further annoyed because always assume that closure refers to syntactic closure.

http://en.wikipedia.org/wiki/Syntactic_closure

Heh, sorry.  I've never been good with terminology.   

[Nadaka]

Like I said, its a pet peeve. I know the amount of annoy I get is at least a bit irrational.