CPU-side scroll offset for a bounded content region. More...
#include <ScrollState.hpp>
Collaboration diagram for MayaFlux::Kinesis::ScrollState:Public Attributes | |
| AABB2D | content_bounds { .min = glm::vec2(-1.F), .max = glm::vec2(1.F) } |
| glm::vec2 | offset { 0.F, 0.F } |
| AABB2D | viewport_bounds { .min = glm::vec2(-1.F), .max = glm::vec2(1.F) } |
Detailed Description
CPU-side scroll offset for a bounded content region.
The interactable half of scrolling. offset is a content-space translation, clamped so a viewport of viewport_bounds's size never scrolls past content_bounds. offset increases as the view moves deeper into content: zero shows the top-left of content_bounds, the clamped maximum shows its bottom-right.
content_bounds is not clamped to NDC. It is whatever extent the content actually occupies (the accumulated height of every row a LayoutCursor has advanced through, typically), routinely far taller than the [-1, 1] a viewport can show at once - that excess is the whole point of scrolling.
A consumer applies offset however its own rendering model wants it. Portal::Forma::LayoutCursor folds it directly into placed bounds (CPU reflow, always hit-test-correct by construction, since the same value that moved the geometry is available to move the hit region too). A buffer wanting to shift many vertices at once without a per-element CPU reflow can instead derive a ViewTransform via scroll_view_transform() and skip that per-element work, at the cost of needing to remap hit-test coordinates by the same offset itself: nothing does that remapping automatically, so a GPU-only shift and CPU hit-testing will disagree the moment offset is nonzero unless something closes that loop.
Definition at line 60 of file ScrollState.hpp.
The documentation for this struct was generated from the following file:
