Alexander Kobel <n...@a-kobel.de> writes: > On 04/09/2013 08:26 PM, Janek Warchoł wrote: >> 2013/4/9 Werner LEMBERG<w...@gnu.org>: >>> >>>> It's a very hard problem and probably not solvable. >>> >>> Ah, I misread, sorry. Yes, it might be worth to test with line >>> strokes. >> >> I find it most interesting that apparently there is some way to solve >> this problem. If line strokes render perfectly, why rounded >> rectangles don't? > > A completely random guess from an almost, but not quite, entirely > clueless person: because lines have a unique direction, while > rectangles have two. Hence, for a rectangle one of the two choices is > bogus. > It might be clever for one type of documents (e.g, Lilypond scores > with lots of stems) to take the smaller extent as stroke width, but > you never know about the side effects for other documents. And that's > why they don't do it. > Besides, rectangles bound /regions,/ and I see a valid point in not > willfully changing regions or areas. For lines, it's slightly more > likely that the designer intended the object to be a line than for > rectangles. > > But it might just be that this SA thingie has to be set: > <http://lists.gnu.org/archive/html/lilypond-user/2006-06/msg00275.html>
strokeadjust is useful when drawing lines which should show a given thickness in preference to having best-fit left and right boundaries when viewed independently. However, rounded rectangles are not lines. -- David Kastrup _______________________________________________ lilypond-user mailing list lilypond-user@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-user
