Damn Microsoft #defines small as char in some stupid COM header that I don't care about that gets pulled in when you #include <windows.h>, like their documentations says to do.

Why the hell didn't they use a typedef?  Now I need some ugly hack to work around all my code that uses small as an identifier.

If you #define WIN32_LEAN_AND_MEAN before including windows.h then that prevents some junk from being added. I don't know which junk though: it doesn't seem to stop the min and max definitions so it might not help you here.

Inside the mine.

The Day The Bomb Went Off!

I really like this one.

Strange, as I'm using SWFObject 2 to embed the swf, which is supposed to be "oh so compatible" and all that. I've tried changing to a direct embedding method... could you let me know if it loads successfully for you now?

No, it's still not working. Really strange, I'm not sure why this might happen.

Statically linked version works: doing the same for the past challenges too would be useful.

Can't immediately spot anything in your code that would make it use the thing: my gut instinct would be maybe the preloader or something else Flex-related that's built-in.

You might try using System.setClipboard instead of Clipboard.generalClipboard.setData to try fixing the Mac clipboard issue.

Haha, now I'm getting error #2046, which I remember happening intermittently before, and I'm sure other people have mentioned it in this thread too.

If you post the code I'd be happy to take a look, or you could try adding the compiler parameter -static-link-runtime-shared-libraries=true which I think might help.

I don't have a clue what's going on here.  As far as I can see it's a silly runtime library that I don't even use..  Do the previous excercises work (I can't see that I've changed anything relevant since)?  Do you have flash player 10?

Previous exercises are broken too, I'm on a new laptop so that's what's changed. I do have Flash 10.

Are you using any Adobe text classes? This page says:

That is because TLF text requires a specific Runtime Shared Library to be available to the Flash Player at runtime. If the library is not already present on the user's computer, it will be downloaded from the Adobe site at runtime. The download will be necessary only once for a given computer. It is a good idea to place textLayout_1.0.0.595.swz on your server as a backup in the same directory as your swf file that uses TLF. Should the Adobe site be unavailable, the swf will look for the RSL in its directory.

I can't see any evidence that it's looking on the Adobe site though, the only requests it's making seem to be to your server.

I'm getting error #2032 which is apparently because it's looking for the file http://jonathanwhiting.com/coding/ldw/textLayout_1.0.0.595.swz which doesn't exist.

Apparently it's some silly runtime library?

Whenever I press any key with the game focused, I get a message saying "The Adobe Flash plugin has crashed".

Firefox 3.6, Ubuntu.

It doesn't happen if I load the SWF directly though, so I think it's something to do with how you're embedding it in the page. I don't have the same issue with Freedom Bridge.

Started a simulation thing with a friend.

My rule of thumb is:

If game objects can rotate, use a physics engine.
Otherwise, it's probably easier to code it yourself.

The tool used in a lot of Linux projects is autoconf.

I believe CMake also includes dependency analysis.

Personally, I don't like either for my projects. I have a hand-written makefile (so that I know what's going on) that makes every .cpp file depend on every .h file.

This does mean some unnecessary compilation when you edit a .h file (the whole project is recompiled rather than just the parts which directly use it) but it has the advantage of being really simple.

I would advise to create both an installer version and a zipfile version. There are always people who are annoyed by installers, and similarly there are people who are annoyed by something not having an installer.

I was surprised recently by my (fairly technically minded) friend, who said that he was annoyed by zipfiles, not because that meant he needed to find dependencies, but because he lost track of these tiny games dotted around his computer. He wanted installers so that he could keep track of them and easily uninstall them.

That's a personal organisation problem as much as anything, but he was someone I would have expected to be firmly in the "just give me a zipfile" camp.

Another thought is that 2d  games have the immediate advantage of the sense of unknown created by a scrolling view, there is less to show, and when you're trying to nail a certain atmosphere, it always helps to only have to show the things necessary (to do so)... also working in 2d with a plane of space allows usage of fundamental design concepts that can effectively lead the Player's eye and ultimately their thought process as they take in information.

I actually think the opposite: in 2D you can always see all of your immediate surroundings, whereas in 3D you never know what's directly behind you.

You have less control over the direction the player is facing (and thus what they're seeing) in 3D though (that said, most of the time I think you'll have a pretty good idea which way they're likely to be going).

"contents" is an array of ints

Whilst this is correct, I would recommend avoiding describing things in this way because it perpetuates the confusion which caused this bug in the first place.

Surely this is not correct, or the bug would not have occurred?

If "contents" had been an array containing only ints, then there would be no need to cast the array element to an int. Therefore, "contents" must have been an array containing at least one non-int element.

I checked this myself just to make sure I wasn't missing something obscure:

Code:

var a: Array = [5];
var n: int = 1;
trace(a[0] + n); // 6

If you're loading files from a website, crossdomain.xml must be in the top-level directory for that website, e.g. http://www.dropbox.com/crossdomain.xml

You can't do that on Dropbox, so if you want to load files from a website you'll need a different host (possibly you'll need to get your own hosting).

Are you sure the game isn't just running at 60 frames per second and only accepting one click per frame?

I much prefer FlashPunk, I think it's easier to use and the API is better.

FlashPunk is still young, so it isn't quite as stable as Flixel yet, but it's a great library.

I made another level subverting the normal difficulty curve: this time if you fail a challenge you may fall down to an area filled with checkpoints. The only way to continue from where you were is to treat these checkpoints as spikes and make your way to the nearest spike (for the purposes of this level, the reset button does not exist).

Damn checkpoints!

Not massively happy with it. Perhaps next I will try making a level which actually follows the theme.

P.S. Inverse difficulty with checkpoints: now actually feasible, though that first fall is still ridiculously difficult.

Huh, that's weird. fcsh-wrap should behave exactly the same as mxmlc.

Unless it's not running properly and not telling you. What output are you getting?

Also try running `fcsh-wrap killall` and then trying again. In theory that should only do anything if it was working to start with, but I'd like to make sure.

Ah right, in that case you can do this:

Code:

FCSH_COMMAND = "'/path/to/flex sdk 4.1/bin'"

With two sets of quotes.