I ran into a PG backend crash when running UPGRADE FUNCTIONS on a
PostgreSQL 8.1.1 instance.

I upgraded to something newer (8.1.9) and that seemed to alleviate things.

Looking at release notes, I think I can locate a seeming cause for this:

http://www.postgresql.org/docs/8.2/static/release-8-1-4.html

"Fix crash from using and modifying a plpgsql function in the same
transaction"

UPGRADE FUNCTIONS obviously does this; it, within the same transaction,
loads (modifies) ALL the Slony-I stored functions, including the one for
doing function upgrades, and runs the one function, near the end.

I suppose I'll add a "gotcha" to the FAQ on this one.

I had been hoping to get the 1.2.10 release out today; I still have some
more testing I want to do, so I think it'll have to wait until Monday. 
FYI, I haven't heard any test reports, positive or negative, from anyone
other than Jan Wieck about 1.2.10...