Developers Corner

by Shockie on 10/19/2002 16:36, 279 messages, last message: 02/04/2006 13:50, 158677 views, last view: 05/05/2024 13:36

i would like to have this thread for development problems/ideas, and what people have done. i believe it could be quite useful!

   Board Index   

#260: Building with Visual C++ 5

by D.plomat on 12/22/2003 12:21

Finally made it!

The only C++ compiler i have on a windos laptop is this old VC++ 5
First tried with MinGW but didn't found why it always complained about missing 'socket.h' (or more exactly why wasn't it here after some reinstallations) and finally gave up.

After opening the C++ 7 project file in notepad to figure out the options/libs etc, i still had 4 errors like this one in tools.h, about block and dynent:
error C2300: 'dynent' : class does not have a destructor called '~T'

Couldn't find a patch / option in VC++5 to solve this (probably this should be ignored with non-classes types/classes without destructor). So i finally commented-out the offending statement:

void setsize(int i) { for(; ulen>i; ulen--){} /*buf[ulen-1].~T();*/ }; // UGLY Visual C++ 5 workaround

I suppose this won't hurt, as vector is used in Cube with structs only?

Then i've a working binary, except that it uses windowed software renderer, and the first image is ok, but when i move the mouse i got a lot of HOM. That was probably temp objects files in the project, as i've recompiled 2 days later without changing anything and now it's OK (still in windowed mode, but at least no HOM's and accelerated OpenGL :)

BTW, anyone else tryed it on this vintage VC++ 5?

reply to this message

#261: vc6

by |apskaus on 01/02/2004 12:41

Since i'm a newbie at this, simply trying to open the new sources with vc6 is giving me a headache :P I used GCC with the previous version. Seems like i need a new workspace file or something, but i have no idea how to make one. Has anyone else compiled this with vc6?

reply to this message

#262: Re: vc6

by D.plomat on 01/03/2004 22:58, refers to #261

There is a workspace file in the vcpp dir of the source archive. It's for VC++7, i don't know if it can be opened with VC++6.
If you want, tomorrow i can send you the file i've done for VC++5, it's probably compatible with VC++6
Probably with VC++6 you won't have the incompatibility with templates that i've made an ugly workaround for.

reply to this message

#263: Re: vc6

by |apskaus on 01/03/2004 23:21, refers to #262

Cool :) I can't open that file with vc6, but one for vc5 should work just fine. my email is kurtbjarne2k at the hotmail.

reply to this message

#264: Re: vc6

by D.plomat on 01/04/2004 00:37, refers to #263

I've just sent it.

You've got to unzip the files on a directory named vc5proj, at the same level as the vcpp dir.

It's for cube_2002_10_20 but it can probably be easily adapted to the new Cube version.

Let me know how it worked with VC++6, if it worked then when i'll be back home i'll upload it to my webpage.

reply to this message

#265: Re: vc6

by |apskaus on 01/04/2004 02:59, refers to #264

oh, i was hoping for a 2003 one. I have no idea how to adopt it.

But i'll test it out with 2002 later anyway. then maybe i can figure out how to make it work with the new version.

reply to this message

#266: Re: vc6

by D.plomat on 01/04/2004 10:15, refers to #265

Well, i think you can use it the same way with the 2003 version, as AFAIK the compile options and filenames are the same, this will probably work. It's just that i still didn't tested it with the 2003 version.

reply to this message

#267: Re: vc6

by |apskaus on 01/04/2004 17:57, refers to #266

Well, i know there's source files that wasn't there in the 2k2 release. But that shouldn't be a big problem.

reply to this message

#268: Re: vc6

by D.plomat on 01/04/2004 18:34, refers to #267

Normally the #include will make them used automatically. Or maybe you'll have to add them manually to the file list, don't remember as i don't use very much VC++

reply to this message

#269: Dev Q

by pushplay on 01/15/2004 04:51

Could someone (Aard or Eihrul probably) explain the difference between enet_time_get_sec() and time(NULL) so that I can make use of the new enet? Thanks.

reply to this message

#270: Re: Dev Q

by eihrul on 01/15/2004 22:21, refers to #269

No diff, just use time(NULL) instead.

reply to this message

#271: Re: Dev Q

by pushplay on 01/16/2004 00:11, refers to #270

Ok, thanks.

That leaves me with a problem in that when I use the new enet and time(NULL), the client crashes when it connects to the server. I'll have to track that problem down.

reply to this message

#272: vc++

by Gron on 01/17/2004 04:12

Hi I vc++6 and i can't find the project file can someone helpme

reply to this message

#273: Re: vc++

by D.plomat on 01/17/2004 19:54, refers to #272

The included project file is for VC++7.

I've sent you the project files i've made for VC++5/6

reply to this message

#274: Re: vc++

by Gron on 01/18/2004 13:38, refers to #273

Thanks! this should be very helpful :)

reply to this message

#275: Re: vc++

by D.plomat on 01/18/2004 18:12, refers to #274

I've just seen i've made an error in the 2002_10_20 version, but i suppose you want to compile 2003_12_23.
Anyway, in some time i'll put those on my web page.

reply to this message

#276: lib.obj

by beckers on 01/14/2006 19:28

I did the instructions in #17, but MSVC++ says:

fatal error LNK1104: cannot open file 'c:\w\cube\lib.obj'

can someone help me out of this

reply to this message

#277: alternative rendering tehniques

by sonyafterdark on 01/21/2006 20:21

Ok, a maths question.. . :)

What is the overlap area between whatever 2 triangles? All tip coordinates are known.

I have been told there is no such expression in terms of rational functions and most other standard functions like sin or cos and that it cannot possibly be a polynomial.

I would love not having to resort to an algorithm for a solution, as a last ditch option.

I'm planning to use it for a 3D software renderer I'm working on. Instead of sampling at the pixel centres or in several specific points of a pixel (as all the current 3D renderers do, hardware support or no) this one will sample 2 triangles per pixel (2 pixel halfs).

Calculating the overlap area between a given projected triangle and each half of the current screen pixel. This would extremely accurately determine what input that particular projected triangle should have on that particular screen pixel as far as determining it's final colour, via a colour components averaging between all the projected triangles visible at that particular pixel weighed by each triangle's total overlap area with the 2 pixel halfs. The result for each colour component would naturally be rounded to the nearest integer and added to the other colour components to give off the final pixel colour. This, for now, excludes texture mapping which is an even more touchy issue with this "sampling" method.

Best regards,

Vishy.

reply to this message

#278: why software rendering?

by asdff on 02/04/2006 02:59

this is not a sarcastic flame question.

just curious why you want to render in software.

reply to this message

#279: Re: why software rendering?

by sonyafterdark on 02/04/2006 13:50, refers to #278

I want to render in software because I want to code my very own entropy renderer for perfect sub-pixel accuracy. If I want to make it run optimally, so that it can hopefully be run in real time, I need to know the answer to this. There is no way to achieve this using either OGL or DX. And besides. I'm not that good a coder. I barely know a bit of the windows api. I'm more interested in developing better math and 3D algorithm thus improve the theory, leaving the actual implementation and coding to those who are knowledgeable in this field. If I manage to develop optimally fast math for an entropy renderer, I'm sure hardware implementation would eventually follow. I also want to render in software because it's the only way I am able to code an engine, with my level of coding knowledge. I can only use fpcx.dll's directdrawblit function to flip the software rendered frame to the display.

reply to this message

   Board Index