• Advertisement

BinaryStylus

Member
  • Content count

    50
  • Joined

  • Last visited

Community Reputation

133 Neutral

About BinaryStylus

  • Rank
    Member
  1. ToolBars With VC++ [ FIXED ]

    Yeah - I'm really sorry - This could serve as a lesson :) Basically - I have a big data structure - that is where all the HWNDs for my windows. When I CreateWindowEx the result goes into this repository. Well I was trying to create the Toolbar with it's parent selected from the repository - Except at this point the HWND hasn't been saved there - 'Cause WM_CREATE happens before CreateWindowEx returns - Hence it happens before the window's handle gets saved. My Toolbar was being built with a NULL parent Handle - And I can't belive its taken my 48 Hours to realise that :'( Thanks To everyone
  2. ToolBars With VC++ [ FIXED ]

    It just doesn't seem to work - I am just getting this error: 1406 - Cannot create a top-level child window. No Matter what I try - I can't force this thing to co-operate. I've even tried to create this tool bar with the assumption that it has to sit under main window. I don't know! LOL - Common controls seem to be - well . . . Does anyone use em?
  3. Hi There - Using VS2003 and VC++ .NET - I've not done too much window controls programming - mainly just edit boxes and Dialogs etc. Decided to use a Toolbar to jazz things up a bit - Primarily I needed a way to show the user if the application was in a particular "State" or not - A toolbar, I thought, would let me do this - Because I could leave the button depressed to show the user that the state was Active. OK - Well I can't get tool bars to work at all - Best I've gotten is the system to report error: 1406 - Cannot create a top-level child window. Which is interesting because MSDN states "A top-level window is any window that is not a child window" (http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnwui/html/msdn_styles32.asp) Has anyone got any reference material for this - MSDN keeps talking about using CreateToolbarEx but then states that this is obsolete. I've gotten the commctrl.h header included and comctl32.lib in the linker - I call InitCommonControlsEx and that returns OK - But then when I do something like hwndTB = CreateWindowEx(0, TOOLBARCLASSNAME, (LPSTR) NULL, WS_CHILD | TOOLTIPS | CCS_ADJUSTABLE, 0, 0, 0, 0, hwndParent, (HMENU) ID_TOOLBAR, hInst, NULL); It gives me Error 1406 again. I gather that the Resource tools for making Toolbars relate to MFC - 'cause they don't seem to help at all if you're just using the API. I've wasted a whole weekend on this - and am just going to implement my own toolbar class with button controls - Which I can probably do in a short time - I just can't believe that Toolbars (Which are pretty significant for MS Windows) are this un-intuitive! Anyone got any demo code or anything that puts a tool bar up on screen? Cheers everyone! [Edited by - BinaryStylus on August 27, 2006 3:29:06 PM]
  4. Hey Dude! It looks like your code takes an XY and Z and makes a translation Matrix 1 0 0 0 0 1 0 0 0 0 1 0 X Y Z 1 and then Multiplies that with your Matrix (A). Now I guess (A) is an orthogonal rotation matrix with possibly a translation Component and will be of the order. a b c 0 d e f 0 g h i 0 X Y Z 1 If the matrix (A) does not contain a translation before you multiply it - then all you need to do is place X in the X position in your matrix (A), Y in the Y pos, etc. Otherwise, (If your Matrix (A) could contain translation data before the multiplication), all you need to do is Add your X to the existing X in your matrix (A). There are HUGE optomisations that can be made if you know that your matrix will contain only Orthogonal rotation + Translation components. because basically you are only then dealing with a 3 X 3 matrix with a vector (Translation component) tagged onto it. But Certainly - because you build an identity matrix with just a translation in your function - there will be no problem with going straight to a solution with pure addition.
  5. Question about LOD's

    Hi - I've got that book and liked it a lot. The Roam seemed quite good if you had computer (psudo random and on-the-fly) generated terrain. But I conceed your point that the tiles are probably a better solution - being tiles they lend themselves to generated terrain also. Here's the deal though. You've got your terrain. It's nice and it has a LOD running and you're even running physics between the terrain and the other Models. Its just that if the other models are running their own LOD - the two might/could look wrong together. I'm taking about those slightly naff moments when any illusion is broken - like where the models look to be in one graphical style and the terrain is another. And I suppose Roam breaks down there too. At least with the tiled solution it will be easier to link the terrain tile's LOD change distances with that of the models. Ahhh - what do I know. Besides the scale is different. And I guess popping would start to become an issue if you slaved the terrain LOD to the models. One thing I was playing with was like a terrain "detail map" to add extra vertex detail - (but nothing significant enough to be considered by the physics) - to the terrain at close distances - to make the stages in detail of the Models match up with those of the terrain. OK - I'll be quiet now :(
  6. Question about LOD's

    Hi, I'm comming up to the point where I am interested in doing LOD for my models. And I am considering having a value with each vertex that is a weighting - 0 is always shown - 1,2,3,4 etc are only shown at closer and closer distances. Using an index buffer makes things easier - as it lets me keep all the level 0 Vertices at the front of the vertex buffer and all level 1 after, etc - so that if the distance is such that I only need up to level 3 data - i only send the first few tens of data - as opposed to the whole lot. Additionally I don't need to actually save the Weighting value in the vertex itself - because if the vertex is in a certain range inside the vertex buffer - it is implicitly a certain value. And the same thing can be done with the index buffer - although it is much quiker and efficent for the model editor to make a new index buffer for each level of detail - than work them out on the fly. Hey just some Ideas I'm playing with for what I'm working on. Tim!
  7. Does anyone find it hard to keep going?

    Yeah - Its good to know other people get like this too. I was stressing that I lacked dedication :P Anyway - I done some stuff on paper, I've thought about it, I have a testing strategy, and I've got a long list of small jobs - So I can just start crossing things off my list. Thanks all!
  8. Does anyone find it hard to keep going?

    Yeah - I'm getting back into it again now - Don't know about playing Video games tho :) Last one I did was WOW and I even started booking time off work to coincide with Guild raids. I should have had therapy (Sooner). I want to start playing some indie games - thats the best idea someone's given me in ages. Cheers
  9. Laptop suggestions

    Dude - Try This Hope it helps a little
  10. Have a major burnout problem. I work like stink - obsesivly for 30+ hours a week on my little projects for a couple of months - Really start to learn stuff and then make progress - and then all of a sudden its really hard to focus anymore. I've now sort of wasted 3 weeks - havn't lost any enthusiasm - but just can't seem to get in deep again. Hmmm. . . Or maybe its the Summer :(
  11. Scariest game?

    Action Adventure / Survival Horror - are my favourites - Most scary IMHO is System Shock 2 and Silent Hill 3 - I also thought that Doom 3 was pretty scary. Love this thread
  12. New To Programming

    If you want a really cool book to get you going with C++ FAST then please try Mark Walmsley's book "Programming in C++" its only £6 from Amazon and it is really nice. It'll get you up to speed very quickly. Well its the book I learn't from, Hope this helps
  13. New To Programming

    Yeah - I agree with Stecal849 - You could really put yourself off programming if you make your learning curve too steep. You can do cool stuff with almost anything - VBA and Access is all you need to make a really rubbish space invaders game - :) - I know I got bored at work. Keep it up
  14. Access Violation Reading 0xffffff [Fixed]

    Just found this in the MSDN library: To create an array whose base is properly aligned, use _aligned_malloc, or write your own allocator. Note that normal allocators, such as malloc, C++ operator new, and the Win32 allocators return memory that will most likely not be sufficiently aligned for __declspec(align(#)) structures or arrays of structures. So I suppose that this is it :( // Edit // ======== My solution and it works thus far is to define a macro: #define __Align(iAlignment) void * operator new(size_t s){return(_aligned_malloc(s,iAlignment));} void operator delete(void *p){_aligned_free(p);} void * operator new[](size_t s){return(_aligned_malloc(s,iAlignment));} void operator delete[](void *p){_aligned_free(p);} And include it in any class/structure that uses __m128 data - Typically I use 16 in the iAlignment parameter. This is cool because the constructors for the classes still get called - and operator new is automatically declared static. In complicated classes I set up a structure/class for thier data and have all of the __m128 stuff moved into that struct/class - and just give the main class a pointer to the data class. The data class has __Align(16) in it - and the main class' constructor and destructor have calls to new and delete (respectivly) to make the data pointers live. Additionally in the data class I make sure that any __m128 stuff is declared at the top and leave any int/double/float etc at the bottom - so the __m128 stuff is shoved against the 16 byte boundary. // ------------------ Its not really a "Fix" or a one-size-fits-all solution but it does seem to be a way for dealing with data alignment. Microsoft state that thier C++ new and Malloc aren't strict enough for __declspec(align(16)) - and they suggest using _aligned_malloc - Stroustrup also says you're asking for trouble using malloc instead of new for dynamic class creation - So this is what comes out. Additionally I placed my __Align macro in a non-trivial class that had __m128 data - and the problem still occured - Even after I moved the __m128 data to the top of the class. That class had several non-trivial data types - D3DPRESENT_PARAMETERS for example declared in it - and I think that maybe that is just too much for new when it is trying to work out (with alignment) how big the size_t value should be and what the offsets for the data are. [Edited by - BinaryStylus on April 15, 2006 8:54:59 AM]
  15. Access Violation Reading 0xffffff [Fixed]

    TY - Thats nice. I'm working on a memory manager now - The intel article suggests that the problem occurs when I use New - which is a fair point - as the temporary variables work fine. It seems that I will just need to write an overloaded New operator for every class that contains an __m128 variable. Apparently _aligned_malloc is the way to go. Cheers
  • Advertisement