[metapost] envelope proposal 4

Larry Siebenmann laurent at math.toronto.edu
Thu Mar 17 06:45:41 CET 2005



Hi Taco,

It seems I finally correctly understood the intent of proposal 4.

You write (Wed, 16 Mar 2005 08:52:08 +0100):

 > The [polygonal pen] restriction is easy to avoid if you need to
 > anyway, because you can always use
 > 
 >    envelope <path> withpen makepath(pencircle scaled 20);

Agreed. 

Quiery: Could one arrange to get a warning in the logfile 
rather then an error, in case inadvertantly one uses the envelope 
function with a pen that is not polygonal?  Suggestion: in 
that case return the trivial path (0,0) of length 0 rather than 
generating an error.  Then one can test for such mistakes by 
asking whether length of the result is 0. I am imitating the 
well-known convenience of having intersectiontime = -1 when 
there is no path intersection.

Cheers

Laurent S.



More information about the metapost mailing list