Add Gradle Build Caching for Status Checks

[Advay17]

While working on our team library, which is based off of the vendor template, I realized that the build checks take up a lot of actions minutes. Of course, there are some modifications that could easily be made to reduce the number of minutes used, such as restructuring the checks so that it builds on one platform first before moving on to other platforms to ensure that the code is written correctly. However, one change that isn't as obvious is storing build artifacts using Github's cache feature. This speeds up builds by about 30 seconds per artifact. While actions minutes are cheap enough, it still is nice to reduce costs when possible by adding caching.

[ThadHouse]

Github's caching isn't free either. So if you have a private repo, you'll have to pay for caching too. And from our experience, that caching actually costs a lot more than minutes.

[sciencewhiz]

Does your team library use jni? If not, this template is overkill.

[Advay17]

Yes, our team has some JNIs that we are working on implementing and using. Plus, with the amount of repositories we are maintaining for our robots, mirroring changes for common files across different bots is becoming somewhat tedious, which is why we have moved away from just copying and pasting files. Sure, we could use submodules/subtrees, but these solutions are messy in their own way.

From my testing, it seems that caching on all file systems takes up about 5-6 GB for our relatively small size repository(a little over code 20 files so far). Since caches only store the latest version of a branches, and since most libraries aren't very large on their own storage wise, I'd imagine this would be reflective of the size most vendor dependencies will be at consistently. This falls under the 10 GB limit that Github imposes for free caching, so I'd imagine that most teams would benefit from it. At the very least, IMO this should be a part of the CI workflow but commented out, so that experienced users can use its benefits if necessary.

[Advay17]

Note, I haven't factored in Windows' cache size yet, only estimated it. Still, even if it were to push beyond the max size cap, it's simple enough to remove it from the estimate