On 2010/07/05 05:48:16, MikeSol wrote:
I see 3.5 places where the patch may need improvement. 1 is that
lineto and
curveto seem unnecessary, as they can be automatically detected by the
number of
function arguments.
For purposes of human readability, I think we should keep lineto and curveto, even if they are automatically detectable. But I also think that even if curveto is detectable, lineto isn't because moveto has the same number of arguments as lineto.
Two is that I try to use predefined lilypond commands as much as possible when they exist - could the moveto's be calls to ly:stencil-translate?
I suspect that moveto's could be rewritten as calls to ly:stencil-translate, but then you'd need to change all the arguments. And I don't see the benefit of it. Both ps and svg have the moveto functionality available; we might as well use their functionality as the common calls, in my opinion.
Third is that I am wary of any loop and/or for-each construct in scheme: I think there is a way to do this with
tail-regression that
dispenses with the loop and is more Schemy.
I, too am wary of loop and for-each constructs in Scheme in general, but I think that in this case it works quite well.
Three.5, I think the extents are off for the curves in certain problematic cases,
Certainly the extents are off for the curves; the control points bound the curve but don't define the curve. So the extents may be a little bit larger than exact.
but that'll work hopefully work itself out via this proposition, to wit:
I think the best way to move forward on this patch would be to work on
merging
its functionality and nomenclature into the connected-shape stencil.
If path and connected-shape are to be merged, I'd like to see the resulting stencil named "path" and take the input that's defined in Patrick's patch. I think that path is a very user-friendly way to create complex stencils, and connected-shape somewhat less so.
I'd be more than happy to iron that out with you on the sidelines - just
shoot me an
email and we'll get that up and running.
I'm sure that you're good for this offer. Thanks for doing so. Carl http://codereview.appspot.com/1730044/show _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org http://lists.gnu.org/mailman/listinfo/lilypond-devel
