TWiki > Team1 > WorkLogs > KrislinLee	TWiki webs: Main | TWiki | Know? | Sandbox?
	Team1 . { Changes | Index | Search | Go }

Worked on Tank and Bullet class. I created the Tank class to be a Bullet factory so that launch returns a pointer to the Bullet being launched. It is very important to catch these pointers. Otherwise, there will be memory leaks.

In response to the bug of the bullet hitting the terrain instead of the tank when it is moving fast enough (that is, is within a timestep of both the tank and the terrain), I worked on the collision detection code. I rearranged the code so that it will first check for the earliest hitting time for the terrain, and then it will check for the earliest hitting time for a tank. If it finds that there was a collision, it chooses the earliest of the two, preferring tank collisions to terrain collisions, and performs the desired collision. I also added code that keeps track of what tank or triangle is being collided with. I also added many many comments to this code and renamed a few of the variables to try to make this very important part of the code be more readable. By the time I needed to leave because of appointments made previously, the code compiled but there was an error with the ball colliding mid-air. I didn't know what to do at the time, and one of my team members thought that it may have been a result of the sky, so I had to leave it that way.

I continued to work on the ball colliding mid-air problem, finding out that there was no sky limit. Jacob and I tried some ways to debug the code, which proved very inconsequential. I worked at it and got no where until the point where I needed to leave due to appointments made previously.

The code was fixed while I was away, though I had to find this out second-hand. It seems that the problem was a result of some test cases that were backwards in our planar face detection code. These were removed rather than fixed because they were no longer relevant because of some manipulation of the axes and the types of triangles being generated. So now I'm working on sounds. I listened to sounds and decided what sounds are appropriate for what events. Then I made a first attempt to add sounds using Microsoft's PlaySound? function. This try was unsuccessful, and the project wouldn't even save because of some memory problem. However, since the code that was made didn't work, it was no big loss. I will work on this more later.

Jacob and I worked on the presentation for class. Then I worked on sounds for the rest of the night. I was successful in getting sounds compiled and running. I tweaked whether the sounds were asynchronous or synchronous as I thought appropriate. The successful hit messages are randomized. The end of the game messages are randomized as well, though I chose bird sounds for some reason. I was thinking: since the war is over you can now hear the sounds of the birds. But whether this end sound is appropriate or not is debatable. I personally think it works. I attempted to reconstruct the solution -- project naming conventions to reflect the change in name from Doom Tanks to Moon Tanks. I ran into some problems with the project file not being able to be opened again, so I went to a version that did seem to work and left it at that.

The sounds were tweaked a bit more to make the movement of the turret a clicking sound since it used to be a higher sound for phi movements. I created the some sequence diagrams. The presentation is over. However, when it came time to submit the files, we learned that the project file wasn't able to be opened after being closed, similar to the problem that I encountered last night (it still seems pretty strange that I was able to open the file today to tweak it some more even).

I worked on trying to fix the project not opening problem by starting a new solution and adding the files in again. After many attempts, I determined that that process will only give a solution that can compile and run but when closing it will not allow you to open the project file again. Jacob later looked at this problem with me, and we determined that it was something that happened when you add the sound resources. The problem is still needing to be resolved. We decided that if we couldn't fix the code then we would just send the source and resource files over and give instructions on how to build the solution.

I wrote a formal Risk Analysis. I then looked back to the project-not-opening problem. With further checking to see at what point the project would not open, I determined that Visual Studio has a problem with the '#' character being in the name of a resource file, which some of the sounds that we used had. Therefore, I changed the names of these files, and the project opened just fine. I made a clean project folder, made sure it worked fine, made a release version, and zipped up the release version and the project folder. I submitted these to cs121submit. The Risk Analysis is still waiting to be turned in. All these files can be found at http://www.cs.hmc.edu/~krlee/MoonTanks/ . I would've put it on our CVS, but there is currently no more room on our CVS. Let me know if there is help required for the rest of the documentation/web page stuff.

So the User's Guide and the Web page were not completed yet, so I spent today making a User's Guide and created and posted a web page. I couldn't put it in our team's directory because it does not have enough space. This may be due to the amount of files in our CVS directory. So the website can be found at http://www.cs.hmc.edu/~krlee/MoonTanks/ . That whole folder should be public readable in case the files need to be copied over. I made a couple of minor revisions to the SRS in the process. I also added glut32.dll to the Resource folder to make the game playable on computers without glut.

-- KrislinLee - 13 Dec 2003