Re-designed the cave theme. It's now much cleaner with new cavern carving routine mixed with occasional rectangular rooms and small rooms that are attached to free places after the main generation is done. Here is a result:
The nice thing about rooms is that I can later add room themes randomly, something like treasure rooms or rooms full of monsters.
Thursday 12 November 2009
Wednesday 11 November 2009
Bugs
I've found five bugs from Teemu since release of v1.0. Four of them are fixed. Here is the one I have to fix:
Some of the areas are not connected and some items and monsters are created in them. I made sure that stairs are always generated in the origin point of room so it's pretty much the center of it and those points are always connected in other points with corridors so stairs never get created in detached pieces, but I should still fix them. It should be easy.
Some of the areas are not connected and some items and monsters are created in them. I made sure that stairs are always generated in the origin point of room so it's pretty much the center of it and those points are always connected in other points with corridors so stairs never get created in detached pieces, but I should still fix them. It should be easy.
Monday 9 November 2009
Tweaking the engine
I've got the plan for Teemu v1.1 pretty much figured out. There may be some extra things that I want to add, but they are not important for the main "plot" of the game. I'm actually quite optimistic this time, because I have done some changes to the engine that makes it more data-driven. It's very similar to the way I'm refactoring Kaduria.
Besides now I have a testing tool that re-generates the level when I press [r]. I didn't have a working one in v1.0 which made the testing of level generation quite challenging. I had to exit the game and restart it to try it again. Now I can make quick changes and test it really good to see if there is any problems.
It's obvious that even a small and relatively simple game is easier to maintain with data-driven approach, even if it's just some parts of the source code that actually are data-driven.
Besides now I have a testing tool that re-generates the level when I press [r]. I didn't have a working one in v1.0 which made the testing of level generation quite challenging. I had to exit the game and restart it to try it again. Now I can make quick changes and test it really good to see if there is any problems.
It's obvious that even a small and relatively simple game is easier to maintain with data-driven approach, even if it's just some parts of the source code that actually are data-driven.
Friday 21 August 2009
Thursday 20 August 2009
File dialog trouble
What happened was that Windows file dialog collided with SDL_syswm.h which defines WIN32_LEAN_AND_MEAN. I think the best solution for now is create another file for file dialog only, but I really need to get rid of windows file dialog. It should not be used in SDL program, but I can't think of any alternative solutions right now. I really hate gui libraries that are gigantic and most of gui libraries are just like that. Besides I like to work with Stile's gui, because it's making everything possible and not making the program too slow.
Fixed name clash with 'pixel' and also made working RGB handles. It required almost total rewriting of Color_Handles class, but I'm glad I did it. The only problem is HSV bug that is fetching the base color value from HSV selector area, when I think it should take it from current tool color.
---
The solution for collision was easier than I thought. Just switch the places of windows. h and SDL_syswm.h.
Fixed name clash with 'pixel' and also made working RGB handles. It required almost total rewriting of Color_Handles class, but I'm glad I did it. The only problem is HSV bug that is fetching the base color value from HSV selector area, when I think it should take it from current tool color.
---
The solution for collision was easier than I thought. Just switch the places of windows. h and SDL_syswm.h.
Friday 22 May 2009
Trouble with Teemu
After 1.0 release I really haven't been that interested to code Teemu. I'm busy with Agduria (and Kaduria) and it may well be possible that Agduria will be playable game sooner than Teemu 1.1. Maybe I added too many features in Teemu 1.1 and going through the to do list isn't fun. Also I feel that I said what I wanted to say with Teemu: that I'm capable of programming a complete game, no matter how simple it is. Not that it's really that important, but I began to get annoyed about jokes related to Kaduria.
Besides it was true what many people said about a smaller side project. It's a great way to improve programming and game design skills and also create solutions that can be used in other projects. As games "7DRL" or small roguelikes can't compete with major ones, that is certain, but they provide good practice for programmers, if there was no previous experience of that kind of stuff.
Besides it was true what many people said about a smaller side project. It's a great way to improve programming and game design skills and also create solutions that can be used in other projects. As games "7DRL" or small roguelikes can't compete with major ones, that is certain, but they provide good practice for programmers, if there was no previous experience of that kind of stuff.
Friday 17 April 2009
Teemu update
It's too slow to program Teemu @ work so I'm taking it back to home. Besides DevC++ is worse than VC and when the project is getting bigger it's clumsy to browse it with DevC++. Plus you don't always have time to program at work, because some people keep telling you should do some things, you know, work stuff. That's breaking my concentration always when that happens.
I could probably concentrate on Teemu anyway, hack the 1.1 ready, release the source code and let other people take care of possible new versions.
I could probably concentrate on Teemu anyway, hack the 1.1 ready, release the source code and let other people take care of possible new versions.
Subscribe to:
Posts (Atom)