This problem appeared (and was reported) when the regexp engine was replaced by ICU, but I'm quite sure that it was fixed some time ago on previous dev builds... problem is that it seems the problem was reintroduced: if you use search / replace with regexp and on the "replace" box insert a reference, you'll get $1 instead of the reference. For example, searching for duplicated words with
\b(\w+) *\1\b and writing $1 on replace, the duplicated word will be replaced by $1 instead of the single word. Another problem that was fixed but returned is the autoreplace that do not match case of the target word. For example, if you type "onomatopoeia" and then start typing ONOM... you'll end with ONOMatopeia, i.e., mixed case. That also worked on a previous dev build. Regards Ricardo
