I may be missing something, but what is wrong with:

try:
  if dictionary[key] == value:
    # do stuff
  else:
    # do other stuff
except:
  # do different stuff

?

It's a zillion lines long and requires dropping out of expression context to statement context!

If I was writing some complicated boolean expression along the lines of

if a<b and c==d and (dict[key]==value or other_equally_good_case):

and realised that the dictionary might not contain the key at all, then I'd much rather correct the code to

if a<b and c==d and ((key,value) in dict.viewitems() or other_equally_good_case):

than have to turn it into something like this:

try:
    tmpvar = dict[key]==value
except:
    tmpvar = False
if a<b and c==d and (tmpvar or other_equally_good_case):

There just doesn't seem to me to be any contest there.

Actually, that's close to what I'd probably write (though not always; context is everything, and I'd do "except KeyError" not bare except, of course). Though the reason for that is that I wouldn't call the variable tmpvar; I'd give it a meaningful name, and think of it as an opportunity to write more self-documenting code.

I don't think I've ever come across this in practice, though; normally, I just work with dicts which either don't contain None, or where None would be expected to be treated the same as the entry being missing.

That should be "except KeyError:" :p

[sorry, "except:" is one of my pet python hates]