[metapost] MetaFont: Unexpected behavior of intersections times

[luigi scarso]

On Wed, Apr 6, 2011 at 10:54 AM, Boguslaw Jackowski
<B_Jackowski at gust.org.pl> wrote:
>
> Hekko, Everybody,
>
> Larry:
>>
>> One helpful sufficient condition for one or more genuine
>> intersections of the two paths p and q is this
>> criterion:
>> ($) existence of a quadrilateral Q [...]
>
> Is it possible to construct such a quadrilateral efficiently?
>
> Dan:
>>
>> I once had two semicircles that crossed
>> (theoretically) at the midpoint of each, and MP said
>> they did not cross.
>
> Larry:
>>
>> Loosely interpreted, this sounds to me like a bug, so I
>> would like to see a concrete example and its explanation.
>> It seems to contradict the claim I have just made.
>
> To me also sounds strange; it was several times mentioned on this
> list that: `p intersectiontimes q = (-1,-1)' implies that the curves don't
> touch/cross each other, although `p intersectiontimes q <> (-1, -1)'
> may sometimes occur for non-touching curves, but very close to each other.
>
> I'd be astonished if the former assertion was invalid. But, frankly
> speaking, I am being astonished repeatedly... ;-)
>
> Cheers -- Jacko
I will send you off-list tonight the example from which I've take the paths
(I hope to show more for bachotex).
In the context of MetaFont, I consider this case a "border case"
because it seems to happen
in situations where, in the end,  the pixels are anyway correctly put .
-- otherwise the font designer should change
something in his/her code if he/she uses directly these kind of paths.
At least this is my brief  experience .
-- 
luigi