With the early-2012 simplification of the Unicode update tools, we do not need to rebuild ICU & tools any more between generating property names data and the rest.
However, after a UCD update and running genprops, we still need to rebuild ICU and the Unicode tools before running genuca so that it picks up the new case mapping properties and NFC normalization and probably further properties.
It would be nice to merge genuca into genprops as well, like merging gencase, genpname, etc. This would require that the newly-built UCD data be made available (injected) via some temporary APIs that are consistently called by genuca.
The hard part is probably making sure that genuca does not call – directly or indirectly! – any ICU-library properties APIs but rather those new internal-to-genprops APIs.