Overset intergrid boundary and hole search

[jbakosi]

Overset intergrid boundary and hole search

---

This change is 

[jbakosi]

@adityakpandare I think is ready to merge - after a final look. Thanks for the help with this.

[jbakosi]

The asynclogic problems (partially) should be taken care of by https://github.com/quinoacomputing/quinoa/pull/597/commits/7bdcd1e54fb31e6612a0f1260815e0dca9e33d21. Those that migrate still deadlock, you could disable migration for those by

diff --git a/tests/regression/inciter/transport/SlotCyl/asynclogic/CMakeLists.txt b/tests/regression/inciter/transport/SlotCyl/asynclogic/CMakeLists.txt
index 9ded28ef3..aba6eff55 100644
--- a/tests/regression/inciter/transport/SlotCyl/asynclogic/CMakeLists.txt
+++ b/tests/regression/inciter/transport/SlotCyl/asynclogic/CMakeLists.txt
@@ -27,6 +27,7 @@ function(add_async_test scheme virt npes ppn migr)
     set(ARG "-c slot_cyl_${scheme}.q -v -u ${virt} -b ${migr_cmd}")
     set(TEXT_RES "diag.0;diag.1")
     set(TEXT_BAS "noop.ndiff.cfg;noop.ndiff.cfg")
+    set(migr_cmd "")
   else()
     set(ARG "-c slot_cyl_${scheme}.q -i unitcube_1k.exo -v -u ${virt} -b ${migr_cmd}")
     set(TEXT_RES diag)

[jbakosi]

I did find a couple of data structures in Discretization that I forgot to pup, fixed in https://github.com/quinoacomputing/quinoa/pull/597/commits/3d5a0300261d11d30f6d24dd47fc4524e8fb6632, but that still does not fix the migrating overset asynclogic tests. At this point I'm not sure how to fix those. Unless we come up with something obvious, I vote for disabling those tests, and putting migration to the overset todo list.