Monday, February 6, 2017

Andrew DiNunzio PPJ 1/31/17 - 2/7/17

This week I spent a lot of time on the repo itself.

Our repository was approaching the 2GB limit on BitBucket, so I created a new repository. This time, I made use of Git LFS, which has a 5GB capacity instead of 2GB. I set it up to track a majority of the large binary files. Git LFS has the benefit of not storing every revision to binary files (which is often actually just storing the entire file multiple times) in the repository itself. This makes (theoretically) makes cloning, pushing, and pulling faster. I haven't seen this speed improvement though.

We had some issues with the new repo, so I spent some time getting people on the team updated with it. We also had a broken build for about 3 days this week. We're not sure what caused the issue, but it seems Unreal Engine doesn't work well while transferring files between two versions (even if it's just a difference of build version), there was an issue while transferring it from an external hard drive, or Unreal crashed at some point while saving, causing the files to be corrupt. It could also be an issue with Git LFS, but I don't know why that would be the case.

There were 2 corrupt files that needed to be reverted, and unfortunately that caused Cancelosi a good deal of work.

I also spent some time working on the AI with Cancelosi and DiLucca. We got it to a point where it swings the tool around after grabbing it, and it even scored once. It still has a very long way to go, but it's much better than what we had up until now.

I created a "calibration room" to help with the development of the AI. It's basically just a room that throws balls within a certain radius of you from random locations along a plane in the distance. This will help us gauge how we throw the ball, and it will help us test the AI when faced with balls coming from different angles and speeds.

Time spent: Total: ~18 hours

  • 5 hours - Setting up the new repo using Git LFS
  • 5 hours - Fix broken build while trying to salvage some work
  • 5 hours - Worked on AI to properly hold the tool and swing once it has the ball
  • 3 hours - Calibration room

  • AI actually returns the ball sometimes now
  • The new repo makes use of Git LFS to keep the local repos small (and was also an amazing learning experience).
  • Git LFS complications and broken builds hindered progress a bit this week

No comments:

Post a Comment