We should compare 140.0a1-15.0-1 with a branch from the 128 series to catch any problems we might have missed between subsequent rebases. I think this branch should be more or less on-par with 128.10.0esr-14.5-1. Probably, we can split this work, even though @brizental and me should probably take not too much part in it, since we've already seen the range-diffs many times, and we risk of not catching our errors. However, I could prepare a "commented" version of the HTML'ified range-diff, to make this task easier. /cc @morgan