Advecting a swarm that was never populated fails with 'negative dimensions are not allowed' (DMSwarm local size -1)

[lmoresi]

A uw.swarm.Swarm that has had no particles added on any rank reports swarm.local_size == -1 (the DMSwarm size before any population), and swarm.advection(V_fn, dt) then dies inside numpy with ValueError: negative dimensions are not allowed. In serial as in parallel. An empty rank of a populated swarm (size 0) is handled since #680; the never-populated swarm is not the same case and deserves a clear error naming the cause. Found while resolving #693 (a driver advected its tracer swarm before the first release).

Reproduction:

mesh = uw.meshing.UnstructuredSimplexBox(minCoords=(0,0), maxCoords=(1,1), cellSize=0.25)
x, y = mesh.X
swarm = uw.swarm.Swarm(mesh)
swarm.local_size            # -1
swarm.advection(sympy.Matrix([[-y, x]]), 0.01)   # ValueError: negative dimensions are not allowed

🤖 Generated with Claude Code

https://claude.ai/code/session_018T2VHUGaZiQVJ95qQ4DiSL

[lmoresi]

Fixed in 15da74b on #688: Swarm.advection raises RuntimeError('This swarm has never been populated ...') when the DMSwarm size is still -1; tests/test_0116_swarm_never_populated.py covers it with the populated swarm as the control. Reopen if #688 does not land.

🤖 Generated with Claude Code

https://claude.ai/code/session_018T2VHUGaZiQVJ95qQ4DiSL