After checking again the discussion here and on the tasks, and testing again the patch, I think it is good to go.
Adding the optional OK button can be done later, it is not really part of the dataset changes.

Closing as Resolved.

Great, yes a review can always be useful :)

I've almost finished the new version of the activity, it will be ready to commit to master probably tomorrow.

My sources are in WIP state now so I can not really test this patch now.
But if it follows what I proposed on the dataset description, which as far as I can see it does, it should be good for me :)

Thanks, closing it now.

Good, thanks!

Actually making good progress, now I think I can get it working good soon :)

OK, I'll persever a bit more to try to make my proposal work for now, it should be possible... Only if I fail after a few days I'll go back to plan B.

Hmm.. I started to implement my idea, but when I reached the point to add the gravity/collision part, I realized there was some technical issues I didn't think about.
The current code to simulate gravity is highly depending on the current design, it will be hard (at least for me) to adapt it to work the way I described.
Then I was thinking we could probably use QmlBox2D, but it will make the activity a lot more heavy than it currently is.

The patch itself looks fine, however I would prefer to fix this issue the other way for those activities.
That is: use the score as an actual score counter, so make it start at 0 instead of 1.

Additional note:
To increase difficulty of levels, I suggest to increase the frequency of planets and maybe also the length of levels.
Example: level 1 could have 5 planets to avoid, on a duration of 1minute (with a constant vertical speed), and level 10 would have 50 planets to avoid on a duration of 5minutes...
These are random values to illustrate the idea, the actual numbers need to be adapted of course.

I think we need to completly rework the intro_gravity to be more educative/less weird...

Right, personal informations would not be so much the issue.

Considering the implications of the GPDR nowadays, I would prefer not to gather datas like this...
Maybe we can close this task as "wont fix"?

No need to remove the hint, but please add context for translators.

Yes you can, no one is working on it currently :)
Please assign the task to you.

@dekumar I just opened a task for it T12809

Still needed is to move the images names as part of the datasets, as suggested by Johnny.

Tested, the layout can still be improved a lot...
Though as it is not really part of the dataset changes, please keep it as it is, and I'll work on it after your changes are merged.

indeed, it would make sense to specify the image names in the dataset as you suggest.

More specifically, layout issues are (as said on irc, but noting those here for reference):

As discussed on IRC, the layout has some issues, and can be improved in several cases...
I suggest to use some states to handle all possible situations, making the best use of the screen area available in any case.

After discussion on irc, we came to conclusion that it is anyway not a good idea to keep both model and rebuild-area on same line, as it doesn't work good past a certain number of items, and even worse in vertical mode.
I made some comments about it on the task, but basically it would be simpler to just have the model and rebuild-area on top of each other in any case for one-line levels.

After all the work I did last year to complete this activity, it works quite good on desktop (Linux at least, not tested on windows yet), but there is weird bug on Android that prevents using it properly.
(some kind of issues when loading canvas... to investigate further).

Stamp tool is working quite good now, but still has the issue of minimum size that should not be 0, so I keep that task opened for now as reminder.

Bucket fill tool could still be faster, but at least it works good now, even on empty area.
Closing as fixed.

Closing as fixed.

I think it's better to keep selected BG color when creating a new drawing.

This is fixed already