selftests failing

[Robey Pointer]

As of this morning's bzr.dev @1765, there are 20 failures and 1 error  
in the selftest (on osx).  The 'tee' log is huge (350K -- so I won't  
attach it unless someone asks for it) but at least the last few  
failures seem to be encoding-related.  Is this known?  Bad timing on  
my part?

The only reason I care is that I use the selftests to verify that I  
haven't broken anything when I'm hacking on my own branches.

robey
(I'm baaaaack...) :)