[PATCH 1/2] Get rid of superfluous variables in stem tracking
Some minor things I came across when working on a fix for [SV 54161],
which I'll send next.
Sure, the comments say that we should get rid of stem_str, but the
reality is that this variable is probably here to stay. Using it for
constructing the stem makes the kinship to another place where the same
thing happens (see next mail) clearer.