Blog Archives

Minesweeper sample

So, I didn’t have anything better to do, so I got again down and tried to make something of my own. I always liked to check my ability to code prototypes. I try to participate in game jams to test my skills, and I tend to watch a lot of tutorials for any type of game out there. Minesweeper was one of the games that people suggested that good coders should be able to make in less that one day, and I’ve never made a minesweeper before. Also, lately I got inspired by a student of mine, whom I teach ActionScript so I got motivated to create a minesweeper game to teach classes (among other things). The result was quite well and took me less than 1 hour, although I tried to optimize it later and make it as easy to learn and as efficient as I can within the limits of knowledge of my student.

I’ve uploaded the code to GitHub and you can find it here. The game was inspired by Windows Minesweeper. It’s plain ActionScript. 3 classes, and 1 more event class. Cloc reports 367 lines of code which is pretty small number, considering the auto completion I got from FlashDevelop and all. To sum this up here is the hierarchy:

  • In the Main class I just initialize the app. I create the Minefield class and a restart textfield which serves as my button and upon click it calls a Minefield function to reset its grid.
  • In Tile class I set a rectangular tile which serves as a tile in classic minefield. I create the color with a simple shape, and I just store information of the position of the tile (without any good reason. Just housekeeping). then I have a tile id which indicates if the tile is empty or contains a mine. A status variable which store the state of the mine if its shown or hidden (that’s an integer though it could become a boolean). Then I have one more integer which stores the amount of mines that exist in neighboring tiles. That one is print in a Textfield above the shape, when the tile is shown. Functions are the usual, and pretty much self described. Show, hide, reset/update, and a checkTile function which shows it for the player. That one dispatches an event in case it’s empty without any mines arround, and creates a flood effect to show up all other tiles until there are mines around. More on that on the Minefield class.
  • A TileEvent class which serves to dispatch events from Tiles. Most of it is not used but I liked to show up the abilities it could have. Mainly it stores the event type, and the information of the tile (position, and if it contains a mine or not[Tile id]).
  • Finally the Minefield class. First of all I set the minefield’s dimensions and mines. Interesting values are 8×8 with 10 mines, 16×16 with 40 mines, and 30×16 with 100 mines. Then I create the grid, which is a 2D array. I prefer to use Vectors for array but that one could be done the same with the Array class. Grid is created upon player click in order to avoid placing a mine where player made his 1st click. When we have the information of player click and which tile it hit, we then randomly place the mines avoiding the same tile for 2 mines and the tile which player clicked. With mine placement I increase by one the surrounding tiles info about neighboring mines to update their respective text. Minefield is ready. Then I respond to player clicks by doing a check at the respective clicked tiles. If tile was totally empty with no surrounding mines I dispatch an event to unfold surrounding tiles until a mine is around. This is called a flood-fill algorithm with a really simple implementation. In function showUntilMineAround, I check each surrounding tile if it has any mines around or not. If it doesn’t this means that we can unfold more neighboring tiles, so I set the current tile as open and then continue my search calling the same function recursively. Although calling the same function recursively could lead to an infinite loop causing the app to hang I avoid this case by setting in advance the tile to open and then if the tile is already open I return to the parent caller (that might seem a little confusing at first so if you didn’t understand it I urge you to search about recursively calling a function). When I spot a tile that has a mine around I also stop the search at this tile by returning to the parent caller.

That’s just about it. This is a simple sample and tutorial on how to create a minesweeper. I hope you liked this guide, the same as I’ve enjoyed creating it, and as my student appreciated it.

FlashDevelop with ASC2 compiler

In case you haven’t heard Adobe has released a new compiler for Flash and ActionScript. The new compiler does exactly what the legacy does, and mostly without any changes, but faster and better. Or at least that’s what it is supposed to do, but it’s still in Beta. So far you could only try it with Flash Builder 4.7 but thankfully there is a workaround for flashdevelop.

This post is based from this one by Alama, so in case you know French go on and prefer his guide.

Read the rest of this entry

How an embedded swf messes up your code and how to avoid that

When you embed another swf into your code, maybe this operation will break everything appart. That depends on the classes you use. If these classes have same names but different functions here’s your reason.

The problem is that these swfs share the same application domain. You have to set a LoaderContext with the parameter allowCodeImport to true, but if you want to make them completely seperate, then in the same LoaderContext you’ll have to set a new ApplicationDomain. That sets the two swfs seperate, but that also means that both of them can’t communicate with each other, and neither share the same stage property. So you won’t be able to dispatch any events, except set the initial LoaderContext parameters.

That thing took me some hours to find out…

One last, maybe useful thing. You can also embed a swf, but it won’t work as is. The embed must set the mimeType=”application/octet-stream” variable. That’ll make the swf a ByteArray. After that, you can load it with a Loader, as you would load an external class by instantiating with the new keyword.

Center pivot point on a sprite

There’s been a long time since I’ve updated this place. Well, real life has gotten me. My time has been divided between work, and Diablo 3 (yeah I know….). These days at work have been very anxious (and still are). There are a tons of things that I want to write, tons of game tutorials I want to add up, but I just keep delaying it. My bad I suppose, but it’s never too late (or so they say).

This time I had an old school problem in as3. I wanted to rotate a sprite, with its pivot point centered. But as you probably know in as3 the pivot point is always at the top left of the sprite, and it can’t be moved except if you are using Adobe CS Flash. Actionscript junkies are left ouf of this. Now there is an algorithm based on mathematics that you can use, but there is an even easier way for this. You probably know it but as I am self-taught I didn’t and spent a lot of time to find it, I guess I could help others who doesn’t know it either. Read the rest of this entry

Should as3 constructor be lightweight?

When I work in flashdevelop I usually use the code quality tool (PMD) provided by Adobe, to check on the welfare of my code. It can be found under Tools>Flash Tools>Analyze Project Source Code. Generally I try to stick on its warnings and keep my code as clean as I can by its proposals, although sometimes I ignore it on purpose. But there is a warning that always caught my sight.

Constructor must be as lightweight as possible. No control statement allowed, whereas a cyclomatic complexe of 2 has been detected. The Just-In-Time compiler does not compile constructors. Make them as lightweight as possible, or move the complexity of the code to a method called by the constructor. Then the complexity will be compiled by the JIT

I always wondered how much difference can that make and if I should follow it or not. So today I put that on a test. Read the rest of this entry