I fixed that already.

Indeed, Qt.Labs is not a great option, also as they state "Note: Types in Qt.labs modules are not guaranteed to remain compatible in future versions."

I just added a little thing to improve the look of the ok button (I could do it myself afterward, but thought it was good for you to know about this subtlety).

Looks good for me.

It looks mostly good for me now.

Thanks. Little comment though...

I added inline comments for all the needed changes after the changes I made on last commit, and all options editing will work properly.

Good, I could fix the issue with Speed slider value not restored correctly after pressing Cancel in Gletters.

I just made a commit on the branch to fix the issue with settings change not applied directly, and applied the fix accordingly to all activities using the DialogChooseLevel.
Note: now inside the Activities DialogChooseLevel, in onStartActivity, we need both background.stop() and background.start() , and in onSaveActivity we don't need anymore the background.stop() and background.start() .

Also, I think it would be good to hide the instruction box at the top when clicking inside the activity/on weights, right now it only hides when clicking on it directly, which is a bit bothering/not the best user experience.
And it will allow to have more easily the OK button in the top right corner and not be hidden by the instruction box.

The position of the OK button is not very good:
-It should be more on the top of the activity to be more visible
-right now in vertical window/portrait mode, it gets hidden under the score

After testing, I noticed an issue with Options:
-When changing some options and then click "save" or "save and start", the changes are not applied directly (it keeps the settings that were there while opening the Settings page).
Wether the Activity Settings is opened from the menu or from inside the the activity doesn't make a difference.

Tested again, now it looks ok for me, all work as expected.

Closing as Resolved as multi-dataset should deal with this problem directly.

Closing as Resolved as multi-dataset should deal with this problem directly.

Tested, indeed last think to fix is the reset of the position of the helicopter.
Otherwise it works good.

PS: see this task with the general TODO for multi-dataset to see how it fits the whole design idea: https://phabricator.kde.org/T12632

I would really prefer to stick to "cumulative dataset" principle rather than "exclusive dataset".
The reason is to be able to have all datasets selected by default, to not force users to select datasets.

I think this task is obsolete now with the multi-dataset plan; can we close it?

I think this task is kind of obsolete now with the multi-dataset plan; can we close it?

Thanks! :)

Yes please, adding the link to the play store entry would be appreciated.
Link to the full (paid) version:
https://play.google.com/store/apps/details?id=net.gcompris.full

upload finished.

Note: actually in our case, the issue was only on android 5.1, as the 0.97 apks published are only compatible with api 22+ (so starting from android 5.1).
Android 5.0 users only have access to previous release (0.96) apk (which requires api 16+ for the 32bit apks, and 21+ for the 64bit apks), and this release was using older Qt and OpenSSL versions which apparently didn't cause conflict (we didn't get report of downloading issues on those).

It is indeed a problem that happens with latest released apk (0.97) on android 5.0 and 5.1 (as far as I understand).

The Call for Proposals is finally officially opened! :)