The problem in 'P' was introduced a few days ago in 1.0.9 when I tried to pretty up some formatting. I was mainly trying to curtail the verbose output of floating point numbers in structures

 perl -e 'my @a=(1/3, 1/4, 1/5, 1/6, 1/7);
print "[".join(", ",@a)."]\n"'
[0.333333333333333, 0.25, 0.2, 0.166666666666667, 0.142857142857143]
-----------
vs.
--------------
> perl -e 'use P;my @a=(1/3, 1/4, 1/5, 1/6, 1/7);
P "%s", \@a;'
[0.33, 0.25, 0.20, 0.17, 0.14]
[download]

Not as numerically accurate, but when doing dev/debug, easier on the eyes. I hoped it wasn't due to undef detection, but it had nothing to do w/that:

> perl -e 'use P;my @a=(1/3, 1/4, $x, 1/6, 1/7);     
P "%s", \@a;'   
[0.33, 0.25, (undef), 0.17, 0.14]
[download]

As usual, thanks pointing me in the right direction!