Efi Fogel

Here are some comments after a quick review. First, I assume that using Handles, which in turn use reference counting for points and curves, is still inefficient, because we copy...

Now, a kernel is not constructed locally, but every curve stores an additional pointer. I was hoping that the use of the kernel could be eliminated. The Caching_polyline_2 is just...

ok, there is a better and preferred option and it's trivial. The member function line() of the polyline curve is called only from the subcurve traits, so the code that...

Right. I saw your message right after I pushed my changes.... I'll address them during the day (and make sure that the manual builds...) The tests still currently fail, but...

1. Indeed I made the Curve_2 and X_monotone_curve_2 type unspecified. 2. As you guessed, the documentation resides in fake headers for historical reasons, and yes, there is an internal doxygen...

I've fixed the test infrastructure. Now, we have some real failures. So I pass the ball to you, Simon. For example, the split test fails because the precondition in line...

Merge is used by the demo. (Naturally, the demo does not use the caching-polyline traits.) Consider the following scenario: Add a curve c1. Add another c2 the intersect c1 in...

right. I think you can suprees the "assertion" test as well. On Mon, May 17, 2021, 13:05 Simon Giraudot ***@***.***> wrote: > Alright, I tried to do a partial implementation...

Use an exact construction kernel. ____ _ ____ _ /_____/_) o /__________ __ // (____ ( ( ( (_/ (_/-(-'_(/ _/ On Wed, 16 Oct 2024 at 07:06, Garbage123King ***@***.***>...

Regarding the OutputIterator, if you ask me, and this should be consistent across the entire documentation, the text should be much more precise and say: \pre Dereferencing `` must yield...