Merge lp:~alan-griffiths/mir/better-AdorningDisplayBufferCompositor-composite into lp:mir

[Alan Griffiths]

examples: AdorningDisplayBufferCompositor::composite() no long ignores output boundaries and occlusions.

examples: AdorningDisplayBufferCompositor::composite() no long ignored output boundaries and occlusions.

Turns out this is significantly more efficient.

[Mir CI Bot (mir-ci-bot)]

PASSED: Continuous integration, rev:3297
https://mir-jenkins.ubuntu.com/job/mir-ci/256/
Executed test runs:
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build/44
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build-0-fetch/50
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build-1-sourcepkg/release=vivid+overlay/46
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build-1-sourcepkg/release=xenial/46
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=amd64,compiler=gcc,platform=mesa,release=xenial/46
        deb: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=amd64,compiler=gcc,platform=mesa,release=xenial/46/artifact/output/*zip*/output.zip
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=cross-armhf,compiler=gcc,platform=android,release=vivid+overlay/46
        deb: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=cross-armhf,compiler=gcc,platform=android,release=vivid+overlay/46/artifact/output/*zip*/output.zip
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=i386,compiler=gcc,platform=android,release=vivid+overlay/46
        deb: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=i386,compiler=gcc,platform=android,release=vivid+overlay/46/artifact/output/*zip*/output.zip
    SUCCESS: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=i386,compiler=gcc,platform=mesa,release=xenial/46
        deb: https://mir-jenkins.ubuntu.com/job/build-2-binpkg-advanced/arch=i386,compiler=gcc,platform=mesa,release=xenial/46/artifact/output/*zip*/output.zip

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/256/rebuild

review: Approve (continuous-integration)

[PS Jenkins bot (ps-jenkins)]

PASSED: Continuous integration, rev:3297
http://jenkins.qa.ubuntu.com/job/mir-ci/6234/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5834
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4741
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5790
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/406
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/558
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/558/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/558
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/558/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5787
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5787/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8191
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/27401
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/402
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/402/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/256
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/27406

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6234/rebuild

review: Approve (continuous-integration)

[Kevin DuBois (kdub)]

lgtm. Its merely in this specific case that the default compositor and the adorning one need the same logic; so its better to dup than to share the internal code and have to publicly support occlusion detection that each compositor should be coming up with itself.

review: Approve

=== modified file 'examples/server_example_adorning_compositor.cpp'
--- examples/server_example_adorning_compositor.cpp	2016-01-29 08:18:22 +0000
+++ examples/server_example_adorning_compositor.cpp	2016-02-08 18:08:56 +0000
@@ -118,10 +118,68 @@
     glBlendFunc(GL_ONE, GL_ONE_MINUS_SRC_ALPHA);
 }
 
+namespace
+{
+bool renderable_is_occluded(
+    mg::Renderable const& renderable,
+    mir::geometry::Rectangle const& area,
+    std::vector<mir::geometry::Rectangle>& coverage)
+{
+    static glm::mat4 const identity;
+    static mir::geometry::Rectangle const empty{};
+
+    if (renderable.transformation() != identity)
+        return false;  // Weirdly transformed. Assume never occluded.
+
+    auto const& window = renderable.screen_position();
+    auto const& clipped_window = window.intersection_with(area);
+
+    if (clipped_window == empty)
+        return true;  // Not in the area; definitely occluded.
+
+    bool occluded = false;
+    for (auto const& r : coverage)
+    {
+        if (r.contains(clipped_window))
+        {
+            occluded = true;
+            break;
+        }
+    }
+
+    if (!occluded && renderable.alpha() == 1.0f && !renderable.shaped())
+        coverage.push_back(clipped_window);
+
+    return occluded;
+}
+
+void remove_occlusions_from(mc::SceneElementSequence& elements, mir::geometry::Rectangle const& area)
+{
+    std::vector<mir::geometry::Rectangle> coverage;
+
+    auto it = elements.rbegin();
+    while (it != elements.rend())
+    {
+        auto const renderable = (*it)->renderable();
+        if (renderable_is_occluded(*renderable, area, coverage))
+        {
+            (*it)->occluded();
+            it = mc::SceneElementSequence::reverse_iterator(elements.erase(std::prev(it.base())));
+        }
+        else
+        {
+            it++;
+        }
+    }
+}
+}
+
 void me::AdorningDisplayBufferCompositor::composite(compositor::SceneElementSequence&& scene_sequence)
 {
     report->began_frame(this);
 
+    remove_occlusions_from(scene_sequence, db.view_area());
+
     //note: If what should be drawn is expressible as a SceneElementSequence,
     //      mg::DisplayBuffer::post_renderables_if_optimizable() should be used,
     //      to give the the display hardware a chance at an optimized render of
@@ -133,6 +191,7 @@
     auto display_height = db.view_area().size.height.as_float();
 
     glUseProgram(program.program);
+    glViewport(0, 0, display_width, display_height);
     glClear( GL_DEPTH_BUFFER_BIT | GL_COLOR_BUFFER_BIT );
 
     mg::RenderableList renderable_list;