Comments on: A Nearly-Automatic C++ Run-Time Type System http://www.realityprime.com/articles/the-audience Advanced Technology Research Mon, 23 Aug 2010 18:10:19 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: avi http://www.realityprime.com/articles/the-audience/comment-page-1#comment-23115 avi Mon, 30 Nov 2009 04:20:45 +0000 http://www.realityprime.com/articles/the-audience#comment-23115 Updated here: http://www.realityprime.com/articles/c-runtime-type-system Updated here:
http://www.realityprime.com/articles/c-runtime-type-system

]]> By: avi http://www.realityprime.com/articles/the-audience/comment-page-1#comment-22227 avi Fri, 16 Oct 2009 18:15:50 +0000 http://www.realityprime.com/articles/the-audience#comment-22227 I'm hoping to get this released, but I used the original version for a company (I open sourced it to them) and I need to do some work to extricate it from any changes I made for them. Long story short, this just needs me to spend 1-2 hours cleaning it up, and I haven't had the time. Sorry for the delay. I’m hoping to get this released, but I used the original version for a company (I open sourced it to them) and I need to do some work to extricate it from any changes I made for them.

Long story short, this just needs me to spend 1-2 hours cleaning it up, and I haven’t had the time. Sorry for the delay.

]]> By: Alex http://www.realityprime.com/articles/the-audience/comment-page-1#comment-22106 Alex Mon, 12 Oct 2009 13:51:39 +0000 http://www.realityprime.com/articles/the-audience#comment-22106 Hi Avi, Your system looks quite interesting, and I'd love to take a look at it. Have you released the source yet? Cheers Hi Avi,

Your system looks quite interesting, and I’d love to take a look at it. Have you released the source yet?

Cheers

]]> By: Avi http://www.realityprime.com/articles/the-audience/comment-page-1#comment-13372 Avi Fri, 16 Jan 2009 20:38:51 +0000 http://www.realityprime.com/articles/the-audience#comment-13372 I'll try to get it posted soon. Sorry for the delay. I’ll try to get it posted soon. Sorry for the delay.

]]> By: Niels http://www.realityprime.com/articles/the-audience/comment-page-1#comment-13357 Niels Thu, 15 Jan 2009 14:47:14 +0000 http://www.realityprime.com/articles/the-audience#comment-13357 Did you ever decide to publish this? It sounds like just the thing for a little project of mine where I have to serialize certain member variable of a large number of classes to a database. If you've published it, I'd be very interested in looking at it and providing feedback and improvements. Did you ever decide to publish this? It sounds like just the thing for a little project of mine where I have to serialize certain member variable of a large number of classes to a database.

If you’ve published it, I’d be very interested in looking at it and providing feedback and improvements.

]]> By: Alex Toone http://www.realityprime.com/articles/the-audience/comment-page-1#comment-12922 Alex Toone Wed, 17 Dec 2008 07:29:28 +0000 http://www.realityprime.com/articles/the-audience#comment-12922 This sounds like a very nice system avi. I have a (sort of) related Type custom allocation/management system for a project I've been working on and this would be very interesting to me if you decide to GPL and share your code. I'd be interested in taking a shot at some of the improvements you mentioned too. Give me an email avi if you are interested. What I'd really like to do is wrap up a project like yours with a custom performance Type allocation and management system similar to the one I've had to develop, GPL it and throw it up on sourceforge. This sounds like a very nice system avi. I have a (sort of) related Type custom allocation/management system for a project I’ve been working on and this would be very interesting to me if you decide to GPL and share your code.

I’d be interested in taking a shot at some of the improvements you mentioned too.

Give me an email avi if you are interested. What I’d really like to do is wrap up a project like yours with a custom performance Type allocation and management system similar to the one I’ve had to develop, GPL it and throw it up on sourceforge.

]]> By: Brandon http://www.realityprime.com/articles/the-audience/comment-page-1#comment-12075 Brandon Wed, 05 Nov 2008 22:44:41 +0000 http://www.realityprime.com/articles/the-audience#comment-12075 I came across this article via a Google search. I'm a student working on a semester long project with the goal of implementing something similar to WinFS. I was wondering if there was any way I could get a copy of the code, or even a more technical description of how you implemented this. It's something I've been trying to accomplish for almost a month now, but I'm not a very experienced C++ programmer (I only recently decided to make the transition from C to C++). I don't intend on using your code (part of the project is that I must implement anything non-trivial on my own), I just need more than the description given to be able to "get it" and figure out how to implement my own system. Thanks. I came across this article via a Google search. I’m a student working on a semester long project with the goal of implementing something similar to WinFS. I was wondering if there was any way I could get a copy of the code, or even a more technical description of how you implemented this. It’s something I’ve been trying to accomplish for almost a month now, but I’m not a very experienced C++ programmer (I only recently decided to make the transition from C to C++). I don’t intend on using your code (part of the project is that I must implement anything non-trivial on my own), I just need more than the description given to be able to “get it” and figure out how to implement my own system. Thanks.

]]> By: Nick http://www.realityprime.com/articles/the-audience/comment-page-1#comment-4952 Nick Wed, 10 Oct 2007 06:03:04 +0000 http://www.realityprime.com/articles/the-audience#comment-4952 BTW, my C++ got all mangled in the reply. It was supposed to say Attrib ANGLE float ANGLE bar; (Similarly, template angle brackets are missing everywhere in the post) also, in the constructor, I meant to mention that the code initializing the hash table is hidden in a boring INITREFLECT(bar); macro but I manually expanded it for explanation.. BTW, my C++ got all mangled in the reply.

It was supposed to say

Attrib ANGLE float ANGLE bar;

(Similarly, template angle brackets are missing everywhere in the post)

also, in the constructor, I meant to mention that the code initializing the hash table is hidden in a boring

INITREFLECT(bar);

macro but I manually expanded it for explanation..

]]> By: Nick http://www.realityprime.com/articles/the-audience/comment-page-1#comment-4951 Nick Wed, 10 Oct 2007 05:59:15 +0000 http://www.realityprime.com/articles/the-audience#comment-4951 I needed something similar, and wound up with something along the lines of: class Foo { REFLECT // hide hash table public: Foo(); Attrib bar; }; Foo::Foo() { attribs[bar.Name()] = &bar; } where attribs is a hash table hiding inside the reflectable class. I ended up needing to wrap the attribute in a template in order to be able to control access (so that if something external to the class modified the variable, the change wouldn't be unnoticed). I did a similar thing as you for type identification, and compare pointers to strings containing a typename. Coupled with a bit of partial specialization template std::string ConvertValueToString(const T&); template T StringAsValue(const std::string&); I got nice automatic reflection (and script language bindings) due to the attribs map sitting in every serialized class. The thing I don't like about my system is that it has a portion that must be dynamically intialized. The thing I do like about it is that since the attributes themselves are templated, I've been able to swap in a lot of different algorithms and approaches to reflection without modifying the code that uses the reflection at all. Thanks for a though provoking post. I needed something similar, and wound up with something along the lines of:

class Foo
{
REFLECT // hide hash table
public:
Foo();
Attrib bar;
};

Foo::Foo()
{
attribs[bar.Name()] = &bar;
}

where attribs is a hash table hiding inside the reflectable class.

I ended up needing to wrap the attribute in a template in order to be able to control access (so that if something external to the class modified the variable, the change wouldn’t be unnoticed).

I did a similar thing as you for type identification, and compare pointers to strings containing a typename.

Coupled with a bit of partial specialization

template
std::string ConvertValueToString(const T&);
template
T StringAsValue(const std::string&);

I got nice automatic reflection (and script language bindings) due to the attribs map sitting in every serialized class.

The thing I don’t like about my system is that it has a portion that must be dynamically intialized. The thing I do like about it is that since the attributes themselves are templated, I’ve been able to swap in a lot of different algorithms and approaches to reflection without modifying the code that uses the reflection at all.

Thanks for a though provoking post.

]]>