From: Alexandre F. <ale...@gm...> - 2012-12-06 00:23:54
|
On Thu, Dec 6, 2012 at 12:42 AM, Simon Geard <si...@wh...> wrote: > On 05/12/12 21:42, Alexandre Ferrieux wrote: >> >>> .c create arc {*}[mkArcByChord $x1 $y1 $x2 $y2 $h] -style arc >>> >>> where, obviously, mkArcByChord does the appropriate trigonometry (in >>> script or in C). > What you've highlighted here is the lengths a user > has to go to just to draw a circular arc segment - it really shouldn't > be that difficult. Seriously, is the above one-liner difficult ? To summarize: two options. (a) Keep computations and graphic primitives separate. Chord arcs just need one expand and a library function. (b) Conflate them. Sacrifice orthogonality. Introduce tricky interactions between options. Complicate canvas-to-postscript even further. Not really a dilemma, right ? While if you accept to isolate the helper function (in C), it'll be at least debatable to include it in the Tk core (justified by your main use case) instead of pushing it behind the Extension wall. You'd trade a NO for a MAYBE ;) -Alex |