Merge lp:~kdub/mir/fix-1485070 into lp:mir

[Kevin DuBois]

android: reject overlay optimization if DisplayBuffer is rotated.

fixes lp: #1485070

android: reject overlay optimization if DisplayBuffer is rotated.

fixes lp: #1485070

HWC is able to handle this, but a bit more work is needed (https://trello.com/c/G1dPmg12). For now just reject the optimization.

[PS Jenkins bot (ps-jenkins)]

FAILED: Continuous integration, rev:2847
http://jenkins.qa.ubuntu.com/job/mir-ci/4579/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/3543
    FAILURE: http://s-jenkins.ubuntu-ci:8080/job/mir-clang-ts-wily-amd64-build/404/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-wily-amd64-build/1076/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/3491
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-wily-amd64-ci/728/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/3491
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/3491/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-mako/6269
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/22596

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/4579/rebuild

review: Needs Fixing (continuous-integration)

[Daniel van Vugt (vanvugt)]

Seems to be consistent with the Mesa implementation now.

I always confuse myself with the function naming. Although the inline documentation explains it, the function name DisplayBuffer::orientation() does little to explain the intended semantics...

    /** Returns the orientation of the display buffer relative to how the
     * user should see it (the orientation of the output).
     * This tells us how much (if any) rotation the renderer needs to do.
     * If your DisplayBuffer can do the rotation itself then this will
     * always return mir_orientation_normal. If the DisplayBuffer does not
     * implement the rotation itself then this function will return the
     * amount of rotation the renderer must do to make things "look right".
     */
    virtual MirOrientation orientation() const = 0;

My fault.

review: Approve

=== modified file 'src/platforms/android/server/display_buffer.cpp'
--- src/platforms/android/server/display_buffer.cpp	2015-06-24 08:17:02 +0000
+++ src/platforms/android/server/display_buffer.cpp	2015-08-14 18:31:31 +0000
@@ -80,7 +80,9 @@
 
 bool mga::DisplayBuffer::post_renderables_if_optimizable(RenderableList const& renderlist)
 {
-    if (!overlay_enabled || !display_device->compatible_renderlist(renderlist))
+    if (!overlay_enabled ||
+        !display_device->compatible_renderlist(renderlist) ||
+        orientation_ != mir_orientation_normal)
         return false;
 
     layer_list->update_list(renderlist, offset_from_origin);

=== modified file 'tests/unit-tests/graphics/android/test_display_buffer.cpp'
--- tests/unit-tests/graphics/android/test_display_buffer.cpp	2015-07-21 04:30:03 +0000
+++ tests/unit-tests/graphics/android/test_display_buffer.cpp	2015-08-14 18:31:31 +0000
@@ -313,3 +313,18 @@
     db.configure(mir_power_mode_on, orientation, offset);
     EXPECT_THAT(db.view_area().top_left, Eq(geom::Point{origin + offset}));
 }
+
+//lp: #1485070. Could alternitvely rotate all the renderables, once rotation is supported
+TEST_F(DisplayBuffer, rejects_lists_if_db_is_rotated)
+{
+    ON_CALL(*mock_display_device, compatible_renderlist(testing::_))
+        .WillByDefault(testing::Return(true));
+    mg::RenderableList const renderlist{
+        std::make_shared<mtd::StubRenderable>(),
+        std::make_shared<mtd::StubRenderable>()};
+
+    db.configure(mir_power_mode_on, mir_orientation_inverted, geom::Displacement{0,0});
+    EXPECT_FALSE(db.post_renderables_if_optimizable(renderlist));
+    db.configure(mir_power_mode_on, mir_orientation_normal, geom::Displacement{0,0});
+    EXPECT_TRUE(db.post_renderables_if_optimizable(renderlist));
+}