Self-Junk Fix

[Eric Angelini]

Message published a few minutes ago on rec.puzzles by Ed Pegg
(a warm thank you to him):

>Here is a potential fix for 3) and 4):
> 
>3') When adding a new integer to the sequence, always choose the
>    smallest unused positive integer that does not contain any
>    impossible digits.  (0 is impossible from the start; 1 is
>    impossible once all of 1 through 9 have been used up, etc.).
> 
>4') Whenever we are blocked, go back to the first integer N that was
>    partly or completely unable to be re-used as a series of length
>    specifications, and replace it with the smallest unused integer M
>    of the same length as N that allows the sequence to be extended
>    further. If there is no valid M, then stop.

The question remains: last integer?
Best,
E.