Jump to content

  • Log In with Google      Sign In   
  • Create Account

Interested in a FREE copy of HTML5 game maker Construct 2?

We'll be giving away three Personal Edition licences in next Tuesday's GDNet Direct email newsletter!

Sign up from the right-hand sidebar on our homepage and read Tuesday's newsletter for details!


We're also offering banner ads on our site from just $5! 1. Details HERE. 2. GDNet+ Subscriptions HERE. 3. Ad upload HERE.


randomguy12345

Member Since 09 Oct 2011
Offline Last Active Sep 17 2012 04:07 AM

Posts I've Made

In Topic: any reason that heirarchies can not be registered for value types

06 September 2012 - 07:52 PM

ha..ha.. Actually my bad. It was a typing mistake.
I do not want the "ZFraction" and "ZIrrational" to be a subclass of "ZInteger".
I want the "ZFraction" and "ZIrrational" to be a subclass of "Real". I have edited and corrected my original question.
We keep ZFraction , because we want to store say 1/3 instead of 0.3333333333.....
but many Irrational do not have fraction representations, they may be quadratic surds etc.


Actually, I want to develop a object oriented scripted framework for high speed " mathematical computation", as it needs to be very high speed hence i am not doing it in java. I want to do it ALL in C++ classes, but only the interfaces will be exposed to angelscript. The reason i chose angelscript has nothing to do with static typing or feature x or feature y . i chose angelscript because it lets me define "my own" types(ie, value types) which will exist on "stack", not on heap, this will give me less cost to speed performance compared to heap approach. This (custom-classes on stack) is not possible in almost any of the C like syntax based scripting languages for C++, it seems to be possible only in angelscript.

But on the above said url in my question , when i was preliminarily reading . It says that if i make them value types then i can not make object oriented heriarchies in them.

my heirarchy (ZFraction and ZIrrational are subclasses of ZReal . There are other classes in this heirarchy as well ZComplex, ZNatural ) etc. All will be implemented as C++ classes as proper parent and child classes.

Now, you said in your reply that
"It's currently not possible to do reference casts for value types, but that doesn't mean you can't register both ZInteger and ZFraction even if they are base and subclass in C++."

Now, from your reply i could not understand much, that by any workaround whether you are saying its possible: yes or no .

If i expose ZFraction and ZIrrational and ZReal (originally implemtned in C++) to angelscript as VALUE TYPE, then are you saying that i can still expose child-parent relationship of ZFraction and ZReal to angelscript, even if they are value types, using some workaround? Any relevant url for this?
sorry for the trouble, but its extremely important for me.

In Topic: implicit data type conversion functions from Class A to class B

06 September 2012 - 07:30 PM

Thanks a lot. :-)

In Topic: C#-style property getters/setters

23 October 2011 - 06:25 AM

thumbs up to "InvalidPointer" and "WitchLord" for incorporating this into angelscript. This brings angelscript a step closer to powerful languages.

my theme for embedded language is to have all the object oriented, functional and other-developer friendly "features of the core language itself", BUT without the baggage of a large runtime library.
properties are very important part of that. Posted Image


In Topic: exceptions in agelscript

12 October 2011 - 11:58 AM

i am also going to consider dart language ( www.dartlang.org ) by GOOGLE
considering that it will also be executable in a VM too (apart from javascript mode)
not much details are out yet.

main worry is what is the size of their VM? that i will have to look out.

In Topic: support for closure

10 October 2011 - 09:45 PM

may i request you to keep a roadmap weboage on angelscript homepage(www.angelcode.com/angelscript )
the goals link (http://www.angelcode.com/angelscript/goals.asp ) there is different from upcoming/wishlist features , which will be a list of user's approved requests by you.
may be version number and feature can be given, but **no timeline** of the version number, if you can't promise so.
for example: http://groovy.codehaus.org/Roadmap

This will atleast keep users interested by having an opinion that "though this language has no feature X, but may be developer is interested, so i can start integration of my classes today, so that whenever that feature comes i will be ready." OR he may have an alternative opinion that "Though this language has no feature X, and this is essential for me, hence i should look for some other scripting language".
As in this thread in 2009 http://www.gamedev.net/topic/533245-1bind-static-member-func-2namespace-in-angelscript/ , you had said that namespace is on you todo list.


MOST IMPORTANTLY: by having this roadmap web page , If somebody needs a feature X which is already is in your approved roadmap(todo list), then he may financially contribute you/angelscript so that you can implement that feature 'early;.

Here is a sample roadmap like:
in roadmap These features may be in core or as an addon
namespace (version 3.1)
exception (Version 2.9)
autoboxing (version 2.9)
closure (version 2.7)

Not in roadmap:
mixin
multiple inheritance

PARTNERS