1. Summarise the issue (what were doing, what went wrong?)

I tried:

  darcs obilterate --dry-run, which resulted in:

"darcs failed:  Can't obliterate patch without reverting some unrecorded 
change."

2. What behaviour were you expecting instead?

I already knew there was a conflict with unrecorded changes. I was trying 
to use "--dry-run" so that I could parse out the file names from the 50 
patches involved to narrow down where the conflict with "whatsnew" was, 
since darcs wasn't telling me.

I see no read for darcs to "fail" in this case. It seems more useful to 
print the --dry-run report even if there are conflicts with unrecorded 
changes.

3. What darcs version are you using? (Try: darcs --exact-version)

darcs 2.9.6.

UI question:

1. do we want --dry-run to behave like obliterate would, or to be more 
adventurous; what's the design principle?

2. is this feasible, implementation wise (guess so, it would be 
basically getting it to show what would happen if user had done the 
revert)

3. FUTURE GARDENER: please re-read Mark's message.  He's trying to do 
something and obliterate --dry-run may not be best way.  Does darcs 
offer alterative solutions? Should it?