Tim and all! >> diff --git a/lisp/org.el b/lisp/org.el >> index 1a1375461..fdeec0d67 100644 >> --- a/lisp/org.el >> +++ b/lisp/org.el >> @@ -19695,7 +19695,8 @@ non-nil." >> (save-excursion (forward-char -1) (delete-horizontal-space)) >> (delete-horizontal-space) >> (indent-to-left-margin) >> - (insert-before-markers-and-inherit fill-prefix)) >> + (when fill-prefix >> + (insert-before-markers-and-inherit fill-prefix))) >> >> I don't have anything better. I think this is a good patch. It makes >> M-j work again. >> >> Possible refinements and improvements can follow. >> >> +1 for applying of your patch. > > I was finally able to reproduce the error. It depends to some degree on > the text in the buffer and where the cursor is when you hit M-j. Adding > some additional text and moving the cursor to different locations > enabled me to reproduce the error and I now agree it is a bug in > org-comment-line-break-function. > > I don't know if your patch is the right fix or not because I don't > understand what the purpose of insert-before-marks-and-inherit is - in > fact, the doc string for that function doesn't even state what the @rest > args argument is for, so I don't understand why it is passing in > fill-prefix. For example, is it safe to assume > insert-before-merks-and-inherit does not need to be called if > fill-prefix is nil? Why is that function even called with the > fill-prefix as an argument?
Thanks for staying critical! I can't answer your questions. I agree that we should deepen the understanding. Your questions are a start. Pragmatically I still vote for applying the patch immediately since it is a step in the right direction AFAICS. Ciao, -- Marco