Merge lp:~saviq/unity-mir/fix-inputarea into lp:unity-mir

[Michał Sawicz]

Map rect to scene in setMirInputArea and set the geometry onEnabled.

[PS Jenkins bot (ps-jenkins)]

PASSED: Continuous integration, rev:100
http://10.97.0.26:8080/job/unity-mir-ci/92/
Executed test runs:
    SUCCESS: http://10.97.0.26:8080/job/unity-mir-saucy-amd64-ci/34
    SUCCESS: http://10.97.0.26:8080/job/unity-mir-saucy-armhf-ci/92
        deb: http://10.97.0.26:8080/job/unity-mir-saucy-armhf-ci/92/artifact/work/output/*zip*/output.zip
    SUCCESS: http://10.97.0.26:8080/job/unity-mir-saucy-i386-ci/92

Click here to trigger a rebuild:
http://10.97.0.26:8080/job/unity-mir-ci/92/rebuild

review: Approve (continuous-integration)

[Daniel d'Andrada (dandrader)]

9 + setMirInputArea(m_geometry);

Speaking of setMirInputArea, should it call "m_surface->removeInputArea(this);" before "m_surface->installInputArea(this);", like InputArea::doSetSurface does?

Because MirSurface::installInputArea() appends the given area to an internal list of areas.

I know it's not the objective of your patch to fix it but still, we should let it pass if it's indeed wrong.

review: Needs Information

[Daniel d'Andrada (dandrader)]

> [...]
> I know it's not the objective of your patch to fix it but still, we should let
> it pass if it's indeed wrong.

s/should/shouldn't

[Daniel d'Andrada (dandrader)]

Looking at it better, this is not a problem since it's sending a pointer to the inputarea class instead of sending the geometry directly

review: Approve

[Michał Sawicz (saviq)]

We're looking whether this caused input to fall through the OSK when in shell.

[Michał Sawicz (saviq)]

Confirmed, need to see how this can affect maliit.

[lp:~saviq/unity-mir/fix-inputarea]

101. By Michał Sawicz on 2013-10-07

Fix *for real*.

102. By Michał Sawicz on 2013-10-07

No need to store the parent.

[Michał Sawicz (saviq)]

Fixed.

[PS Jenkins bot (ps-jenkins)]

PASSED: Continuous integration, rev:102
http://10.97.0.26:8080/job/unity-mir-ci/93/
Executed test runs:
    SUCCESS: http://10.97.0.26:8080/job/unity-mir-saucy-amd64-ci/35
    SUCCESS: http://10.97.0.26:8080/job/unity-mir-saucy-armhf-ci/93
        deb: http://10.97.0.26:8080/job/unity-mir-saucy-armhf-ci/93/artifact/work/output/*zip*/output.zip
    SUCCESS: http://10.97.0.26:8080/job/unity-mir-saucy-i386-ci/93

Click here to trigger a rebuild:
http://10.97.0.26:8080/job/unity-mir-ci/93/rebuild

review: Approve (continuous-integration)

[Gerry Boland (gerboland)]

Works, thanks!

review: Approve

=== modified file 'src/modules/Unity/Application/inputarea.cpp'
--- src/modules/Unity/Application/inputarea.cpp	2013-08-29 14:46:43 +0000
+++ src/modules/Unity/Application/inputarea.cpp	2013-10-07 21:14:57 +0000
@@ -58,6 +58,9 @@
 
     if (enabled != m_enabled) {
         m_enabled = enabled;
+        if (m_enabled) {
+            setMirInputArea(m_geometry);
+        }
         if (m_surface) {
             m_surface->enableInputArea(this, enabled);
         }
@@ -161,8 +164,7 @@
 {
     DLOG("InputArea::setMirInputArea (this=%p, x=%lf, y=%lf, width=%lf, height=%lf)", this, relativeGeometry.x(), relativeGeometry.y(), relativeGeometry.width(), relativeGeometry.height());
     using namespace mir::geometry;
-    //FIXME(greyback) mapping geometry to scene giving wrong output, figure out why
-    const QRect rect = relativeGeometry.toRect();
+    const QRect rect = parentItem()->mapRectToScene(relativeGeometry).toRect();
 
     m_mirInputArea.top_left.x = X{rect.x()};
     m_mirInputArea.top_left.y = Y{rect.y()};