June 5, 2012

Android drag and drop; it's broken...


(Disclaimer: All opinions expressed herein are my own and do not necessarily reflect that of my employer)

Drag and drop seems like a great thing to have in a platform. Since Honeycomb, Android has had one... Unfortunately, it's severely lacking, and in some places just flat out broken.

For the past several months I worked on a customized Launcher app for a rather large phone manufacturer. We used the official google Launcher from the recently released ICS as a starting point.

Now, Android is a great platform; they've done a lot right. Some things though, like their Launcher code, isn't as quality as I would have expected.

We spent the first month refactoring several parts of the code. Pretty early on I had the bright idea to get rid of the monolithic Launcher.java and separate out the home and menu into two activities. I ended up just having a HomeFragment and a MenuFragment in the same activity, which worked much better, but the process resulted in me refactoring the drag and and drop to use what's provided by the platform.

Before this point, the Launcher used a transparent view called DragLayer that covered the window, intercepted events, drew the drag preview etc.

This method doesn't work when you want a drag to be started in one activity and end in another because the drag is tied to a view. (Long pressing an icon in the menu and dropping on home.)

The platform's drag and drop is much more robust in that it's handled completely separate from your activity. Your activity more or less hands off a surface to the window manager which takes it from there.

This has the great benefit of not having to manage dragging yourself. Unfortunately we also lost a lot of flexibility.

The first thing I ran into was you can't change the preview of what you're dragging after the drag has started. We needed to turn a dragged app icon red when hovering over a trash icon. It's simply not possible, at least not with the public API. Using a lot of reflection and manually exposing various functionality I was able to get around this limitation; but most people don't have the luxury of knowing the platform isn't going to change from underneath them.

Next up, views receive every event EXCEPT for start and end if they have a child that returns true on a start event. You can see more details on that bug here: http://code.google.com/p/android/issues/detail?id=25073
Or maybe I'm just using it wrong?

The next problem I ran up against was a lot trickier to figure out. Basically, it's possible to receive two drop events in a single drag. And before android 4.0.3 it would cause a crash that looks like this:

    java.lang.IllegalStateException: reportDropResult() by non-recipient
        at com.android.server.wm.Session.reportDropResult(Session.java:311)
        at android.view.IWindowSession$Stub.onTransact(IWindowSession.java:375)
        at com.android.server.wm.Session.onTransact(Session.java:111)
        at android.os.Binder.execTransact(Binder.java:338)
        at dalvik.system.NativeStart.run(Native Method)

More details on that can be found here: http://code.google.com/p/android/issues/detail?id=25061

There are several more bugs with androids platform drag and drop that other members of the team found. And then there are problems inherit with it operating in a separate process. (Such as if you want to animate your drag shadow once it's been dropped, but that's a rant for another day.)

Long story short; life is fun.

May 9, 2012

Voronoi diagrams in C++

Why? I can't tell you... or wont? Anyways, I was looking for an implementation that was C++ friendly, didn't leak memory and let me easily build triangles from the resulting computation (for use in OpenGL).

I found an implementation that did none of those. But it was fast, and close-ish to C++ friendly... ish.

It was open source so I encapsilated it all, converted all those pesky malloc and frees to... not  malloc and frees (which indirectly fixed the memory issues) and changed it to output the result to something a little more useful.

And now I'm sharing the result with you. https://bitbucket.org/jlm/voronoi-c/

May 6, 2012

Why minecraft's terrain loading is slow

Fuzzycraft - not minecraft!
Oh I don't know, it's in Java? (Sorry, just getting the flame-bait out of the way.)

It's an interesting problem; sure it's just a bunch of cubes - but who wants to stop at a few thousand of them? You want a massive world where you can potentially see tens of thousands of theses blocks at a time. And due to its nature you're left without many of the techniques often used to optimize terrain rendering.

Sounds like a good opportunity to get a little more comfortable with OpengGL. Enter fuzzycraft-webgl; poorly textured terrain in all it's blocky glory! This ended up being horribly slow. More on that in a bit.

Trying out minecraft; the terrain seems to load very slowly. With my fairly high-end GPU I'm pretty sure OpenGL isn't the bottleneck here. Generating what blocks go where isn't slow either. (Test it by toggling the graphics options between fancy and fast; you can still see the chunks appearing one by one, oh so slowly) The only thing left is the process of building the geometry to send to the GPU.

I wanted to see if I could do better. Coffeescript was something I was wanting to try out and I always like to see how far WebGL can go. Turns out, getting good FPS was next to impossible with larger worlds. Generating the blocks of geometry to send to the GPU wasn't too bad; although to be fair, I had 8 cores chewing through it via web-workers.

Really, using OpenGL just works best in C. I was confident that it was possible to get it fast; both generating and rendering it. So I re-wrote it all in C++. There's really not a whole lot to say other than yes, it is very possible. There are still a few optimization I could do to make it run even faster. I may do them someday; but for now I've satisfied my curiosity.

So why does minecraft take so long to get those chunks of vertices to the GPU? I don't really know. Part of me wants to decompile minecraft's bytecode and find out. But I work with way to much Java as it is at my day job; and honestly, it's not a language I adore.