For a while now, I've been growing increasingly suspicious that the
interactivity logic is broken when dealing with the last patch in the
list of matches.  For example, here's the end of an obliterate
interaction which looks like "x" has been treated as "n".

In this interaction, I was repeatedly typing "x", then "y".

I guess the 64th patch was skipped because it adds the .bin directory.

  $ darcs obl --match 'not touch .bin/ru'
  [...]
  Shall I obliterate this patch? (62/64)  [ynWvpxdaqjk], or ? for help: x
  yMon Jan 16 20:20:51 EST 2006  Trent Buck <trentbuck@gmail.com>
    * Initial import.
  Shall I obliterate this patch? (63/64)  [ynWvpxdaqjk], or ? for help:
      Mon Jan 16 20:20:51 EST 2006  Trent Buck <trentbuck@gmail.com>
        * Initial import.

          A ./.Xdefaults
          [...]
          A ./Emacs/Config/windmove.el
  Shall I obliterate this patch? (63/64)  [ynWvpxdaqjk], or ? for help: x
  Skipped obliterate of 1 patch.
  [waiting...]

Sounds like a bug and something that could be boiled down into a test case. 
Marking need-volunteer for somebody to submit such a case to the bugs dir,
perhaps Trent?

On Sun, Dec 21, 2008 at 05:34:13PM -0000, Eric Kow wrote:
> Sounds like a bug and something that could be boiled down into a
> test case.  Marking need-volunteer for somebody to submit such a
> case to the bugs dir, perhaps Trent?

I won't take this ticket right now, as it is lower priority (for me)
than other open tickets.

Maintenance bump.
Test case still needed.

The interactivity logic has changed with Florent's SelectChanges 
refactor.  I suggest we mark this ticket given-up until Trent or somebody 
else can reproduce the issue.