save camera settings on selection to avoid different result when re-projecting the selection 2d screen points transformation (does not compile because of canvas.h cyclic dependency)

src/canvas.h

@@ -6,6 +6,7 @@
 #include "brush.h"
 #include "action.h"
 #include "canvas_modes.h"
+#include <stack>
 
 NS_START
 
@@ -100,6 +101,17 @@ struct PPIHeader
     }
 };
 
+struct CameraData
+{
+    glm::mat4 m_mv{ 1 };
+    glm::mat4 m_proj{ 1 };
+    glm::vec4 m_box{ 0 };
+    glm::vec4 m_vp{ 0 };
+    glm::vec2 m_pan{ 0 };
+    glm::mat4 m_plane_unproject[6] = SIXPLETTE(glm::mat4(1));
+    glm::vec3 m_plane_dir[6] = SIXPLETTE(glm::vec3(0));
+};
+
 class Canvas
 {
 public:
@@ -112,6 +124,7 @@ public:
     bool m_dirty = false;
     bool m_commit_delayed = false;
     bool m_dirty_stroke = false;
+    std::stack<CameraData> m_camera_stack;
 
     static Canvas* I;
     NodeCanvas* m_node = nullptr;
@@ -241,6 +254,11 @@ public:
     void project2Dpoints(std::vector<vertex_t>& vertices);
     glm::vec3 project2Dpoint(glm::vec2 pt);
     std::vector<glm::vec2> face_to_shape2D(int plane_index);
+    void push_camera();
+    void pop_camera();
+    CameraData get_camera();
+    void set_camera(const CameraData& c);
+    void reset_camera();
 };
 
 class ActionStroke : public Action