hmmm, I see, not a regression (same happened in 0.51-4), but definitely some worth fixing as long as we're at it. Now $0-x becoming #0-x is a regression. Well, I put it all into an issue: https://github.com/pure-data/pure-data/issues/1476