fix(imgui): keep the plot inside the child's clip rect and reject unusable input

The widget drew a region derived from the child's content rect alone, so a child
the hosting window cut off cleared and painted over the host's pixels; it also
scaled ImGui's (-FLT_MAX, -FLT_MAX) no-pointer sentinel and differenced it against
the last real position, turning a drag out of the window into a ~1e35 pan, and it
forgot the host's button levels across a re-bind, handing the next plot a press
that never happened.

The region is now the content rect intersected with the child's clip rect (a
no-op when nothing clips it), an invalid position is no position update, and the
button levels survive a re-bind. The controller sanitises non-finite coordinates
and deltas for any host, and the widget's input translation now has a committed
test on the headless EGL context.

tests/CMakeLists.txt

@@ -40,6 +40,17 @@ add_executable(plot_tests
   egl_headless.cpp
 )
 target_link_libraries(plot_tests PRIVATE plot::core ${plot_egl_library} Threads::Threads)
+
+# The Dear ImGui widget is the one adapter with a test here: its input translation
+# (edge synthesis, the hover gate, drag capture, wheel containment) is logic this
+# project owns, not a toolkit's. The suite is otherwise ImGui-free, so the source
+# and the link are added only when the adapter exists -- which the top-level
+# CMakeLists.txt guarantees by adding the adapters before this directory.
+if(TARGET plot::imgui)
+  target_sources(plot_tests PRIVATE test_imgui_widget.cpp)
+  target_link_libraries(plot_tests PRIVATE plot::imgui imgui)
+endif()
+
 # Same switches the library is built with: the cost tests read the counters
 # behind them (see include/plot/data/detail/{cover,merge}_count.hpp).
 target_compile_definitions(plot_tests PRIVATE PLOT_PYRAMID_COUNT_COVERS