v2.6.4

Patch release. Notably fixes a build regression in igl::linprog (missing
std:: qualification on cerr/endl introduced in v2.6.3) plus assorted fixes
and additions since v2.6.3.

igl::triangle::remesh_at_points (#2556)

* working remesh

* working example

* tests for igl::triangle::remesh_at_points

Covers face-points, vertex-points (not inserted, only tracked in K), edge
points on boundary/manifold/non-manifold edges, all edge midpoints of a
cube (= one step of 4-1 subdivision), and many interior points per face on
decimated-knight.

A shared checker validates the invariants promised by the header: V is a
prefix of VV, indices are in range, VV.row(K(i)) is exactly the barycentric
interpolation and is referenced by FF, and each original face is exactly
tiled by consistently oriented, coplanar descendants.

All inputs are deterministic (dyadic barycentric coordinates that sum to
exactly 1) so a CI failure is reproducible.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* remove stray scratch notebooks

Untitled.ipynb and .ipynb_checkpoints/Untitled-checkpoint.ipynb were
committed by accident and are not part of this change.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

Fix dqs not checking if quaternions are on the same hemisphere (#2390)

Co-authored-by: Alec Jacobson <alecjacobson@gmail.com>

Fix AABB::find to not try to run 2D code on 3D input and vice versa (#…

…2294) [ci skip]

* pull out DIM from find

* AABB find test

* templates

use line_color alpha

Remove smileys.

Merge pull request #1468 from libigl/dev

Release 2.2

Merge pull request #1206 from libigl/dev

Version 2.1.0

Update .travis.yml

Former-commit-id: 941b403

Update .travis.yml

Former-commit-id: 941b403