Print

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

[SethCreiyd]

If the intent was declaring a new class inside a method of an existing class, you can declare it the normal way.

Code: [Select]

>>> class C(object):
    def __init__(self):
        class D(object):
            def __init__(self):
                self.dVar = 5
                print self.__class__, self.dVar
        self.newClass = D()

       
>>> c=C()
<class '__main__.D'> 5


The syntax is no different, just the scope.  Not sure what you were trying to do with 'self', but if you want a neat way to make dynamic classes, you can use the type() built-in.

Code: [Select]

>>> class C(object):
    def __init__(self):
        D = type("D", (object,), dict(dVar=5))
        self.newClass = D()
        print self.newClass, self.newClass.dVar

       
>>> c=C()
<__main__.D object at 0x02664090> 5



[Dutchling]

Wait. I actually DID to it right, I just misspelled something.

I guess I shouldn't have a DATE and a DATA class which I both use a lot Dx.

Thanks for the help though

[Rose]

Well, for the last few days I've been working on making some functions that will automatically generate 3d models of roads and intersections with just a few parameters, namely intersection locations, which intersections are connected to which, and the road widths.
I've finally managed to get something working:

Spoiler (click to show/hide)

As you can see, it's not perfect, but at least it's not doing anything obviously wrong, so this makes me happy with it so far.

[alway]

Here's a little C++ tidbit that's come up in my past 2 interviews:
If I have 3 functions,
void Foo1(int, int)
int Foo2()
int Foo3()
and I call them like so:
Foo1(Foo2(), Foo3());
which order do Foo2 and Foo3 run in?

Spoiler: Answer (click to show/hide)

The behavior is undefined. In Visual Studio, it runs left to right, in GCC it runs right to left. This caused their respective companies all sorts of Fun, hard to find bugs when doing multi-platform code which was being compiled with both Visual Studio and GCC. Any time Foo2 and Foo3 have order-specific side-effects, you should avoid this form as it will break things if someone else compiles your code differently.

[Levi]

Ha, interesting. 

Spoiler (click to show/hide)

I think for sake of sanity it SHOULD go left to right.  Funny that it isn't guaranteed.

[Mego]

Ha, interesting. 

Spoiler (click to show/hide)

I think for sake of sanity it SHOULD go left to right.  Funny that it isn't guaranteed.

That is interesting.

Spoiler (click to show/hide)

I agree, expressions with operators of equivalent precedence evaluate left-to-right, so function parameters should as well.

[MorleyDev]

In the interview I'd make a point to go on a rant that it shouldn't matter, and that in the code the order of them shouldn't matter like that because it means the functions Foo2 and Foo3 have non-obvious and unsafe side-effects, which is very unclean. And also point out that of I encountered code like that, I'd see it as better to feel the pain as soon as I realise it and take the time to refactor the code to not do that, than to leave unclean code in the system and let it cause hidden problems later down the line.

Only time it could be unavoidable is in a third-party library, but still it's unclean and should be dealt with and prevented. Passing function calls inside function calls like that is always asking for trouble and personally I try to never nest function calls like that. It just decreases readability with no benefit.

[MorleyDev]

Double post, oops...to give this some meaning, I'll leave you with a thought:
All code is guilty until proven innocent, unworthy until tested, and pointless without singular and well-defined purpose.

[alway]

In the interview I'd make a point to go on a rant that it shouldn't matter, and that in the code the order of them shouldn't matter like that because it means the functions Foo2 and Foo3 have non-obvious and unsafe side-effects, which is very unclean. And also point out that of I encountered code like that, I'd see it as better to feel the pain as soon as I realise it and take the time to refactor the code to not do that, than to leave unclean code in the system and let it cause hidden problems later down the line.

Only time it could be unavoidable is in a third-party library, but still it's unclean and should be dealt with and prevented. Passing function calls inside function calls like that is always asking for trouble and personally I try to never nest function calls like that. It just decreases readability with no benefit.

Eh, it's the video game industry, so most of the code was written by some variety of third parties, be it CryEngine, Havok, DirectX, Unity, UnrealEngine, ect.

[MadocComadrin]

Here's a little C++ tidbit that's come up in my past 2 interviews:
If I have 3 functions,
void Foo1(int, int)
int Foo2()
int Foo3()
and I call them like so:
Foo1(Foo2(), Foo3());
which order do Foo2 and Foo3 run in?

Spoiler: Answer (click to show/hide)

The behavior is undefined. In Visual Studio, it runs left to right, in GCC it runs right to left. This caused their respective companies all sorts of Fun, hard to find bugs when doing multi-platform code which was being compiled with both Visual Studio and GCC. Any time Foo2 and Foo3 have order-specific side-effects, you should avoid this form as it will break things if someone else compiles your code differently.

I would have answered right-to-left, generalizing that since many C implementations expect arguments on the stack in reverse order (ie, calling printf from assembly code would require pushing the arguments onto the stack from "right-to-left."), and that C++ probably doesn't deviate from C too much in that regard (and can often be compiled to C before being compiled to object code) it would be right-to-left. However; that answer makes sense, as I had picked up the original information from using GCC.

[Siquo]

Ha, I knew that one!

However, I'm tempted to give up on C++. Moving to a 64bit OS just made my linker choke himself to death on my project and I'm honestly sick and tired of linking-that-is-not-working. I'm going to try me some Unity next.

[alway]

Yeah, Unity is great for prototyping 3D games as well as making 3D games which could fall under the category of 'typical.' Games which are edge-cases in terms of code use are its main weakness; anything particularly weird like expansive voxel terrain or such would be problematic, but for typical game systems, it's great. It's also extremely easy to pull in assets like models, animations, textures, and so on. You can have a simple game up and running within a day.

[Rose]

I've started using Unity, and it's pretty decent so far, and what I'm doing is far from a 'typical' game.

Anything extra besides the basics that you'd need to do in Unity, you'd need to do in your own engine anyway.

Like this stuff:

Spoiler (click to show/hide)

The difference is in Unity it's much quicker to see the result.

[Nadaka]

So Japa, you trying to do a sim city like engine?

[GlyphGryph]

I hate Unity, personally, but that's mostly because the web player is not cross platform and it seems to be most popular for web stuff. Which fucking sucks.

* GlyphGryph shakes fist.