Length-2 puzzles don't allow twisting with all grips.

[GoogleCodeExporter]

We had a problem where this actually caused puzzles to be unsolvable 
because the scrambling selected random grips from the full set, but that is 
fixed in  r127 and r128.

We would like to make these puzzles use a fuller set of twists at some 
point though, perhaps similar to how we are doing the 2^4 (special casing 
of clicking on 2x2x2 cells).  By "fuller set of twists", we mean a set of 
twists that allows the full permutation state-space of the puzzle to be 
reached.

When we do enhance this, it does change the nature of the puzzles since 
more permutations will be reachable, so we might not want to compare 
solutions of these "different" puzzles.

Original issue reported on code.google.com by [email protected] on 28 Oct 2009 at 6:22

[GoogleCodeExporter]

See related issue 73.

Original comment by [email protected] on 28 Oct 2009 at 6:24

[GoogleCodeExporter]

r156 improved this for most duoprisms, but there are still missing twist 
possibilities 
on some puzzles (e.g. the triangular duoprism and the dodecahedral prism).

Original comment by [email protected] on 19 Nov 2009 at 6:20

[Sonicpineapple]

An example log of a {3}x{3} 2 with an inaccesible move generated by the scrambler.

MagicCube4D 3 1 0 {3}x{3} 2 0.7274252398595265 -0.3816640585800157 0.5702500037733808 0.0 0.6861771309230953 0.40015126183016714 -0.607486553475774 0.0 0.0036695249637737344 0.8331935633716547 0.5529692763098009 0.0 0.0 0.0 0.0 1.0 * 112,-1,2 m|.