On 2015/11/10 20:52:18, dak wrote:
https://codereview.appspot.com/270640043/diff/40001/scm/stencil.scm File scm/stencil.scm (right):
https://codereview.appspot.com/270640043/diff/40001/scm/stencil.scm#newcode73
scm/stencil.scm:73: (width (- (car stop) (car start))) I'd not use `width' (and `height') here but `dx' and `dy': the
expectation for
variable names `width' and `height' is that they will generally be
non-negative. https://codereview.appspot.com/270640043/diff/40001/scm/stencil.scm#newcode81
scm/stencil.scm:81: ;;;; beginning '(0 . 0), ending at (cons
length-to-print
0) Ah, here's the rub. Why end at (length-to-print . 0) instead of (1 .
0)?
Scaling by length at the start instead of after you are done makes
things
unnecessarily complex and indeed requires Pythagoras. If you instead
work with
a "unit" bow, scaling and rotating the resulting points to match some
actual
line does not require calculating the line length but just the
equivalent of a
complex multiplication in cartesian coordinates.
https://codereview.appspot.com/270640043/diff/40001/scm/stencil.scm#newcode123
scm/stencil.scm:123: length-to-print) See? You've effectively multiplied everything with length-to-print at
the
start, just to divide by it again here.
Total waste of effort.
Well, outer/inner-control rely on the user-settable variables bow-height and thickness. Intended is that the user may set/modify them depending on the final bow, not the unit-bow. Changing bow-height/thickness of the unit-bow only 0.01 will have a more heavy impact for the final bow than most users will expect and would be inconsistent with setting thickness for Ties, etc. I could hide this from the user by doing some division maybe with 10, although the appropriate value would be `length-to-print', which would make you not happy again... lol I'll think about it, though, any suggestion is highly welcome as well! https://codereview.appspot.com/270640043/ _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel