Glad to see that it works out well for you chris!

Just a little heads up though: there has still been a little bug in the engine, which allowed the player to click a block while there were still operations pending, if he happened to click at the right tick.

Most of the time this shouldn't have occured, since you basically only realize it if you're going about rapidly clicking blocks while others are moving. Even then it's not that easy to spot.

Anyway, this is fixed now. The engine should be rock solid. <img src="smileys/smiley20.gif" border="0" align="middle" />

So please redownload Part 3.1 if you downloaded before this post was made. Thanks and sorry for the inconvenience.

But as many other say, i would love to see the function, where they dont match, and that they will be swapped back.

I had this functionality for a while in an unfinished match-3 game of mine, which uses a slightly modified version of the engine of this tutorial. Didn't bother to add it here so far, as there didn't seem to be a demand.

But I here you guys and I just added it in, you can download the updated part 3.1 from the first post and enjoy swapping back. <img src="smileys/smiley4.gif" border="0" align="middle" />

It's a bit of a hack job, but I'm fairly sure it works correctly. If not feel free to reply with any issues you might run into.

In Your method can't make blocks with animations?

Let me just quickly come back to that: I think I just realized what the problem was. Of course you can't use frames only for the different blocks like in this example if you want to animate them. You have to use animations instead. To randomize the blocks in a similar fashion you have to use naming conventions for your animations (1_moveright, 3_moveup,...) and a PV to keep track of the actual type (color) of your block. Hope that helps.

That's a very cool little gamemaker already! I would like to see options to mess with the particle effect and score display, then you would have covered almost all the visual aspects of the game.

Also what's obviously missing: switching back blocks that don't result in a match. I will upload an updated cap of my tutorial this week which includes this functionality. <img src="smileys/smiley2.gif" border="0" align="middle" />

Well done so far chrisbrobs, looking forward to new versions!

Nice thread! Here are some of my alltime faves that come to mind:

Super Mario Bros 3

That's Mario in perfection. SMW comes close, but can't beat this NES gem imo.

2. Double Dragon II NES

By far my favorite NES game of all time. Real fun, and has one of the best endings I've ever seen in a game.

Played this game to death with my brother back in the day. Double Dragon 2 is one of the finest scrolling beat'em up games there is, period!

Castlevania 3

Say what you want, to me Castlevania was still best before it became all metroid-ish with Symphony Of The Night. Once again a true highlight on the NES!

X-COM: Terror From The Deep

That's where I truly fell in love with TBS for the first time. Had a cheesy poster of the game hanging in my room for the longest time. :D

TES IV: Oblivion

The total freedom of what you do and how is unbeaten.

Are you sure you aren't confusing this with The Elder Scrolls III: Morrowind? <img src="smileys/smiley17.gif" border="0" align="middle" />

Cause that's where the real freedom was.

Oblivion feels almost restrictive in comparison, also it suffers from having been designed for consoles. That's not a proper

inventory, damn it! Morrowind is still my alltime favorite RPG. Bring on SKYRIM!

No One Lives Forever 2

I love the humor! And the gameplay was great too. The charme of this game can hardly be matched by any number of mindless COD-style shooters. Also great fun in multiplayer (banana slip, explosive robot cat).

When i run tutorial(3) in debug mode, the following occurs:

<img src="http://dl.dropbox.com/u/22173473/match3%20debug.png" border="0" />

The number above the red arrow is constantly increasing.

Any way to stop this ?

The debugger recognizes the count expression, if there isn't any instance of that object. This isn't a bug, but a quirk of the debugger and shouldn't affect performance.

Hi!!!

AweSome tutorial!!!

but...ofcourse...few questions)))

In Your method can't make blocks with animations?

and one more important: If block don't math than they should swapped back.

I try to make it for 2 days and... don't know what to do)

Please help!!!

Thanks alot for tutorial!)

Of course animations are very possible with this method. Why would you think otherwise? What problem do you face trying to animate the blocks?

Also I mentioned earlier in this thread that I considered adding unmatched blocks being swapped back. I have a cap with that functionality floating around my hdd, I just need to dig it up. Can't tell about the needed events on top of my head, as I haven't looked into this example for quite a while. Just give me a little time.

Thanks for your interest in this tutorial! <img src="smileys/smiley10.gif" border="0" align="middle" />

Very true, I posted this already: http://www.scirra.com/forum/forum_posts.asp?TID=43267&title=empty-about-me

<img src="smileys/smiley14.gif" border="0" align="middle">

Did you test Firefox 5? FF5 added decent hardware acceleration, FF4 didn't have that.

The initial test was Ffox4 against Chrome 11, but it appears to be the same picture with Ffox5 against Chrome 12. Ffox5 may have come a little closer.

If you test my example Chrome beats out Ffox still, while with the Ghost Shooter example of this thread, Chrome loses pretty harshly. Weird. And yes, for my testing the 2D Canvas Acceleration is always active in Chrome.

IE9 seems to be dominant all the time btw. Microsoft did a great job there with the hardware acceleration and their javascript engine I imagine.

Right now I think performance is mostly whether or not the browser has hardware acceleration, rather than your hardware. I think right now Chrome is still using a software renderer, so it will always be slow - they're working on hardware acceleration AFAIK.

Not sure if that's true, since Chrome usually performs quite well with HTML5 stuff. As I mentioned in my earlier post it did beat Ffox in a little test I made, which didn't involve fullscreen or scrolling though, it was also a completely passive benchmark. It creates a lot of objects while constantly rotating them and changing their opacity.

Also when you go to the address about:flags in Chrome you can see an option 'GPU Accelerated Canvas 2D' which should be activated by default. So if I'm not completely mistaken, Chrome does indeed do hardware acceleration already.

Having an empty 'About me' block in your profile will actually display the following info: "Tom is a game developer registered on the Scirra website."

I'm sure everyone would love being Tom, but unfortunately we can't... <img src="smileys/smiley17.gif" border="0" align="middle" />

Another little issue: I can't seem to find a way to edit the website address in my profile.

This is shaping up nicely lucid! Even I did once an example that had to deal with line intersections. Now I don't have to use the math formula anymore, very handy! :)

Very interesting. It's running around 60fps in both FFOX 5 and IE 9. Opera plummets to 12fps.

The performance results seem similar to an older test I did comparing HTML5 speed of all major browsers. IE9 still appears to be ahead of the bunch, while FFOX5 apparently can keep up with certain examples.

What's interesting is that Chrome only manages ~22fps with this fullscreen example, while it was clearly the second best after IE9 in my earlier test. Maybe it can't handle fullscreen and/or scrolling so well.

Oh, my specs:

X4 640 3.0Ghz

4GB DDR2

1024mb Radeon 6850 OC

Windows 7 Home Premium 64bit