Use maven artifacts

[auscompgeek]

There isn't a particular need to build a fork of allwpilib, so we should just use the upstream maven artifacts.

[Lytigas]

I agree. How often are those released? If they're only pushed up a few days before kickoff I might be more reluctant.

[auscompgeek]

Same cadence as allwpilib itself. In addition to the releases, development artifacts from master also get built.

[Lytigas]

I followed where MavenArtifcats.md pointed, but at the moment I can't seem to find the development packages; they don't seem to exist recently enough. I also can't find up-to-date NI libraries on there. I'd look through the build system to find out where it fetches them from, but gradle has always been complete gibberish to me. Hopefully you can point out the right place to look.

[auscompgeek]

ni-libraries is at https://first.wpi.edu/FRC/roborio/maven/release/edu/wpi/first/ni-libraries/

From my understanding, for 2020 they've moved Maven artifacts to a different server because Java didn't agree with the TLS deployment on first.wpi.edu... my browser history seems to suggest it is at https://frcmaven.wpi.edu

[auscompgeek]

Filed wpilibsuite/allwpilib#2068.

[Lytigas]

This would also be a good chance to update other things, like the compiler and HAL to 2020. But, I'm not in the beta. How do you handle this for pyfrc?

[auscompgeek]

Short answer is we don't very well. At the very least we could get simulation working for everyone and try linking everything for athena, but testing on a real robot would indeed require the beta roboRIO image.

[Lytigas]

I do know some people in the beta, I might be able to get access. Anyway, 2019 development stopped with 2019.4.1, so it seems like we should stop 2019 development with 0.4.0 and at least get things compiling and react to breaking changes made in the HAL.