Mergatron 2020

[rupertnash]

"Make HemeLB Approach Adequacy Again"
As of today, there are 24 branches and 171 open issues. This is not cool. Let's drain the swamp.

Branch wrangling:

• Rename master -> main (@rupertnash)

  • Create main branch and set as default
  • close all open PRs against master, delete master

• NBC stuff - this removes various long term annoyances with HemeLB (vis, steering, net) (@rupertnash)

  • Rebase nbc_rebase onto current main
  • Test for correctness
  • Test for performance
  • Merge and delete nbc* branches

• Parallel in time (may in fact not require merges - but need to check @rupertnash)

• Octree setuptool stuff (octree, oct-huang, oct-cgal) @rupertnash

• RBC: @mobernabeu and @timmkrueger to set goals - pls edit this message!

  • Update to new testing framework
  • Fix minor unit test regressions
  • Update to new CMake @mobernabeu / @rupertnash
  • Write some user level documentation

• betterVinlets: I believe this can just go @djgroen pls confirm!

• hierarchical-decomp: @djgroen?

• colloids: @schmie?

• geoStrip5000: @schmie?

• velocity_iolet_setup_tool @mobernabeu?

• guoForce: @schmie?

• regularisation: @uschille?

[uschille]

regularisation is probably not worth merging. I looked at the impact of cubic terms in the equilibrium distribution a while back, but only for D3Q19, and I think this only makes sense when also regularizing the collision operator. If no one objects, I'm going to git push :regularisation. We can open a PR at a later time when we have made more progress towards stabilizing intermediate Re flows.

[mobernabeu]

Added some tasks to RBC stream. @rupertnash: could I book a couple of hours of pair-programming with you to plan migration to new CMake and Catch2 please? W/c 12 Oct, ideally.

[rupertnash]

I agree about the docs - I've already started a new branch for this.