Topology refactor #154
Topology refactor #154
Conversation
Do we need to compile with |
I guess, but maybe make that part optional? |
|
Looks great to me. Down the line it may be useful to have an ability to mock an arbitrary topology so that scheduler developers can test behaviors without actually having to have actual machines. |
The topology.rs crate is insufficiently generic, and reflects implementation details of scx_rusty more than it provides generic use cases for modeling a host's topology. This adds a new topology2.rs crate that will replace topology.rs. We have this as an intermediate commit so that we don't bundle updating scx_rusty with adding this crate. Signed-off-by: David Vernet <void@manifault.com>
Now that we have this new Topology crate, let's update Rusty to use it instead of using the old one. Signed-off-by: David Vernet <void@manifault.com>
Byte-Lab
force-pushed
the
topology_refactor
branch
from
ff0ab7e
to
43624a8
Compare
Definitely. It shouldn't be terribly difficult to do with a builder pattern. As you said, we can add it down the line when needed. |
The topology.rs crate is insufficiently generic, and reflects
implementation details of scx_rusty more than it provides generic use
cases for modeling a host's topology. This adds a new topology crate
that replaces topology.rs, and also updates
scx_rustyto use the newcreate.