4:45: I have to write a program implementing a simulated printer queue. I spent about 4 hours on it yesterday, and I'm on about hour 14 today. This program has been a bastard from start to finish, and I need to have it done in 3.5 hours. I have to do it in MSVC, which sucks. My program keeps b0rking when I switch from debug compilation to -O2.
Fiona Apple - Paper bagdebugging...
4:57: fixed that one, had a long overflow, switched to unsigned long, that seemed to work. Now, however, it seems that my queue never empties.
Pink Floyd - Money
debugging...
5:08: alright, the queue is sometimes empty, and I'm getting reasonable average times, urgent prints go faster than normal ones. The queue never seems to be full, so I've kicked it up to some higher numbers, which leaves me more time to type while I wait for the days to simulate.
Black Star - B Boys Will B Boys
debugging...
5:26: Fucked if I know. Behavior seems random. When I start my search at a high number of jobs per day, the queue is never empty, even as the number decreases. When I start it at a low number, search goes much faster, but the queue is never full, even when they reach the same number. arrggggggh. It's cold outside.
Rage Against The Machine - Guerilla Radiodebugging...
5:43: hands shaking. tired. Considered replacing my queue implementation, decided against it. Not even sure what's messed up. Giving up isn't an option anymore, two hours of sleep would just be worse than not sleeping until I'm done with my classes. Definitely skipping work in the math help room this afternoon.
Lovage - Everyone has a Summer
debugging...
6:05: Wow, breakthrough. I'm still in bad physical shape, my head seems to drift off from my body every once in a while, I'm actually surprised I was coherent enough to make a breakthrough. I wasn't reinitializing the probability that a job will be printed (jobs per day / cycles per workday) when I started a new simulation. This means I've effectively been running and rerunning the same godawful simulation for the past hour, and why the simulation was different if I started it at 4000 jobs per day vs. 10000. Still no sunlight, I'm surprised.
Muddy Waters - Sad Sad day
debugging...
6:50: Sunlight, code improving, documentation being done. I also improved a tricky algorithm by a lot, I think. I have to print this out in 40 minutes if I go to the class, 1:40 if I don't. work.
Beethoven - 6th Symphony (unknown orchestra)
debugging...
7:29: OK, it works. my new algorithm works about 100 times faster than the previous one. Once morning came, I felt a remarkable clarity. Unfortunately, I don't think I have enough time to finish documenting the code and go to class, so I'll be working on his code while he's giving class. My professor doesn't really care, fortunately. I can't wait to get real food.
People Under the Stairs - Acid Raindrops