Problem Statement

On a set of coasters we have, there's a picture of a teacup.   On the rim of the teacup the word   TEA   appears a number of times separated by bullet characters   (•). It occurred to me that if the bullet were removed and the words run together,   you could start at any letter and still end up with a meaningful three-letter word. So start at the   T   and read   TEA.   Start at the   E   and read   EAT,   or start at the   A   and read   ATE. That got me thinking that maybe there are other words that could be used rather that   TEA.   And that's just English.   What about Italian or Greek or ... um ... Telugu. For English, we will use the unixdict (now) located at:   unixdict.txt. (This will maintain continuity with other Rosetta Code tasks that also use it.)

Search for a set of words that could be printed around the edge of a teacup.   The words in each set are to be of the same length, that length being greater than two (thus precluding   AH   and   HA,   for example.) Having listed a set, for example   [ate tea eat],   refrain from displaying permutations of that set, e.g.:   [eat tea ate]   etc. The words should also be made of more than one letter   (thus precluding   III   and   OOO   etc.) The relationship between these words is (using ATE as an example) that the first letter of the first becomes the last letter of the second.   The first letter of the second becomes the last letter of the third.   So   ATE   becomes   TEA   and   TEA   becomes   EAT. All of the possible permutations, using this particular permutation technique, must be words in the list. The set you generate for   ATE   will never included the word   ETA   as that cannot be reached via the first-to-last movement method. Display one line for each set of teacup rim words.

Let's start with the solution: