Comment #12 on issue 2514 by [email protected]: Patch: gcc 4.6.3 does not need -fno-optimize-sibling-calls any more
http://code.google.com/p/lilypond/issues/detail?id=2514
Sometimes being paranoid pays off. Faking the affected version numbers made me realize that [0-2] does not travel unharmed through m4 if it is not cushioned in enough layers of enclosing brackets.
Totally annoying. Since I now manually corroborated that this pattern works out, I'll just wait for the next round of Patchy to consider this reviewable for good.
