Mesh_3 - add surface_only() named parameter
#8781
Conversation
to completely skip the "refine_cells" part of Mesh_3
|
no diff in the PMP header? |
|
We could also have a function named |
I think it's not needed here. It's a parameter function like |
OK I thought that you would need to pass the option in the call to |
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
|
/build:v0 |
|
The documentation is built. It will be available, after a few minutes, here: https://cgal.github.io/8781/v0/Manual/index.html |
janetournois
added
Not yet approved
janetournois
added
the
pre-approved
| * When this option is enabled, the ouput mesh has no complex cells, only complex facets, edges and vertices. | ||
| * Full-3D optimization steps such as mesh perturbation and mesh exudation are automatically disabled. | ||
| * |
There was a problem hiding this comment.
"complex cells/facets": sounds strange. The correct wording, used elsewhere is "cells in the 3D complex"
| facet_distance(0.001)); | ||
|
|
||
| // Mesh generation | ||
| C3t3 c3t3 = CGAL::make_mesh_3<C3t3>(domain, criteria, params::no_perturb().no_exude()); | ||
| C3t3 c3t3 = CGAL::make_mesh_3<C3t3>(domain, criteria, params::surface_only()); | ||
|
|
There was a problem hiding this comment.
Now that there is params::surface_only(), this example and its documentation in the user manual will change drastically: there is no need to use a trick with the constructor of Polyhedral_mesh_domain_with_features_3 from a sequence of polyhedra with only one polyhedron.
| if(surface_only) | ||
| { | ||
| for(auto cit = r_tr.finite_cells_begin(); cit != r_tr.finite_cells_end(); ++cit) | ||
| r_c3t3_.remove_from_complex(cit); | ||
| } | ||
| else if(!forced_stop()) | ||
| { |
There was a problem hiding this comment.
I wonder if there is a way to use that Boolean surface_only at compile time. Compiling Mesh_3 with parameters::surface_only() should be faster than without, and the resulting code should similar to surface_mesher. Maybe even the type of Vertex and Cell could be simpler with surface_only().
There was a problem hiding this comment.
You can make surface_only a bool template parameter to the class Mesh_3_options instead of a data member, but I think it is probably out of the scope of this PR and a dedicated issue should be opened.
Summary of Changes
Add the option
surface_only()tomake_mesh_3(), to completely skip the "refine_cells" part of Mesh_3, and cancel perturbation and exudation.@soesau and I noticed that scanning cells, even in the context of "Surface mesher", takes a lot of useless time.
@soesau you can use this branch for your benchmark
@lrineau do you think
surface_onlyshould be a member ofMesher_3or a parameter ofMesher_3::refine_mesh()(as done here)?Release Management