1253812 - Return types possibly swapped through AndroidBridge (GetScreenOrientation <---> GetScreenAngle)

Status: RESOLVED WORKSFORME

(Firefox for Android Graveyard :: General, defect)

Description

[Mark Capella [:capella]]

Can someone reality check me here? This looks inconsistent / wrong ...

If so, it would be way tempting to blame this on my previously reported:
Bug 1242571 - Hard crashes / Android UI Crash / Android reboots on device rotation

http://mxr.mozilla.org/mozilla-central/source/widget/android/AndroidBridge.h?mark=289-290#284
http://mxr.mozilla.org/mozilla-central/source/widget/android/GeneratedJNIWrappers.h?mark=833-833,835-836,846-846,863-863,865-866,876-876#832
http://mxr.mozilla.org/mozilla-central/source/mobile/android/base/java/org/mozilla/gecko/GeckoAppShell.java?mark=2516-2516,2521-2521#2514

Comment 1

[Mark Capella [:capella]]

?ni to be sure we're noticed

Comment 2

[(inactive) Jim Chen [:jchen] [:darchons]]

Looks like screen orientation [1] and screen angle [2] both have Java short type (C++ int16_t type). In GeckoAppShell, getScreenOrientation stays as short and getScreenAngle is expanded to int. Finally, in AndroidBridge, getScreenOrientation is expanded to int (uint32_t) and getScreenAngle is reduced back to short (uint16_t).

So the types are definitely not consistent, but because the values are both 16-bit values originally, the inconsistency should be harmless, and shouldn't cause crashes.

[1] https://mxr.mozilla.org/mozilla-central/source/mobile/android/base/java/org/mozilla/gecko/GeckoScreenOrientation.java?rev=be593a64d7c6#37
[2] https://mxr.mozilla.org/mozilla-central/source/mobile/android/base/java/org/mozilla/gecko/GeckoScreenOrientation.java?rev=be593a64d7c6#280

Comment 3

[Mark Capella [:capella]]

Thanks for the explanation. I'd meant to close this earlier :)