K1kk0z90

I'm back!

BUG: Using the "Set source string" action of the String Parser object causes an HTML5 app to freeze. I've tested it on Firefox 14, IE9, and Chrome 20, and all browsers report this error to Console:

ReferenceError: index is not defined @ Please login to see this link.

Line 1592 of parser.js is the one underlined in this function:

function CRunparserElement(t, i) {
this.text=t;
this.index=i;
this.endIndex = index + this.text.length;
}

Maybe "index" should be "this.index" or "i".

Attached the mfa showing this bug.

BUG: If an Active object, whose sprite angle is rotated using the "Set angle" action of the Active object, collides with another Active object, and there is an event for the collision of these two object and it contains an action, the HTML5 app freezes and I get this error in the Console:

SCRIPT5009: 'ArrayBuffer' is not defined
Sprites.js, Row 341 Character 9

Row 341 of Sprites.js is inside the createRotatedMask function.

I've tested it with Firefox 14, Chrome 20 and IE9. This bug happens only with IE9.

Attached the mfa showing this bug.

BUG: I've tested the mfa of the bug on qualifiers I reported for last Beta, but unluckily it still happens in Beta 5.
I've cloned the old project, and built it again with Beta 5, and the behaviour is the same. In few words, have an Active object (in the example is the Yellow one), that does not belong to any Qualifier. Create another Active (in the example Red), and add the Enemies qualifier to it. Put more than a copy of it in the frame. Create another Active object (in the example Blue) and add the Good qualifier to it. Put more than a copy of it in the frame.
In the Events, put the following 2 events:

Collision between Yellow and Group.Enemies
-Destroy Group.Enemies

Collision between Yellow and Group.Good
-Destroy Group.Good

If you run the app in HTML5 runtime, when Yellow collides with Blue, nothing happens. When Yellow collides with Red, the Red active which collided with Yellow gets destroyed, and ALL Blue instances get destroyed as well.

Attached the mfa showing this bug.

Now sounds on Chrome 21 work perfectly!

BUG: On Firefox 14 there is still the bug with sounds I reported for the previous version. Attached the updated mfa (in which I removed the bug description for sounds in Chrome, that now you solved). In the mfa the bug is described in details.
In few words, in Firefox 14, when a sound is played while another was already playing, when the frame changes (using the Storyboard object actions in the events), that sound is played at the beginning of that new frame.

This bug happens only with Firefox 14.

BUG: The "Change frame while sound is playing" bug is still present in Beta 5.
Have a MMF2 application with "Play sounds over frames" option ON.
If an MP3 is playing, if the application goes to another frame and this contains a Backdrop object or an Active object, a little part of the MP3 is skipped.
This bug happens on Chrome 21 (and also on IE9, but it is negligible on IE9 because only few milliseconds are skipped).
More details in the attached mfa.

BUG: Open an mfa and run it as an HTML5 app using the "Build and run" option of MMF2. The app will start correctly. Now, in MMF2, close this project and load another mfa. Run this as an HTML5 app using the "Build and run" option. When the browser opens, it will load a completely empty page. Copying the URL and pasting it to another browser will give the same result. If you close MMF2 and launch it again, if you load this last mfa and run it as an HTML5 app using the "Build and run" option, the app will start correctly.

BUG: In an HTML5 app, if a sound is played with the "Play sample" action, and the same sample is played with the "Play sample" action while the previous sound is already playing, this causes the previous sound to interrupt.
The correct behaviour (like in C++ runtime) would be that the two sound instances are played together. Instead, the HTML5 runtime behaviour is like if the "Play sample on a specific channel" action was used, using the same channel.

You can see this bug very well using Chrome. Firefox has a big delay playing a sound after a Play Sample action is fired, so it's not the ideal browser to test this kind of bug.

Attached the mfa showing this bug.

BUG: The current number of lives at start of frame, becomes the max number of displayable lives by the Lives object, until the frame is restarted.

I've tested it with Firefox 15, IE9 and Chrome 21, and the behaviour is the same in all these browsers.
The mfa is attached to this post. You can also view the HTML5 app Please login to see this link..

Steps to reproduce it with this mfa run as HTML5:
1. At the start of the frame, you can see the number of lives is 3.
2. Click on the button "Add 1 life". The "lives" variable is incremented, as you can see in the String object on the frame. But the Lives object displays only 3 lives.
3. Click on the button "Restart frame". The frame is restarted and the Lives object now correctly displays 4 lives.
4. Click on the button "Subtract 1 life". As you can see, the "lives" variable is decremented and it is also correctly displayed in the Lives object.
5. Click on the button "Add 1 life". The "lives" variable is incremented, and the Lives object correctly displays the number of lives (4). Infact, at the start of the frame, the number of lives was 4.
6. Now click on the button "Add 1 life" again. You can see, in the String object, that the "lives" variable is incremented to 5, but the Lives object only displays 4 lives again.
7. Click on the button "Restart frame". The Lives object correctly displays 5 lives.

BUG: A MMF2 generated HTML5 app freezes with the action "Frame: Set Background Color". I've tested it in Firefox 15, IE9, Chrome 21, and all browsers report this error in the Console:

ReferenceError: ACT_SETFRAMEBDKCOLOR is not defined @ Please login to see this link.

Probably it hasn't been implemented yet.

Attached an mfa showing this bug.

BUG: When Internet Explorer 9 loads a MMF2 generated HTML5 app, the app works correctly, but there is an error in the Console:

SCRIPT5009: 'Float32Array' is not defined
webgl-2d.js, Row 477 Character 3

I've tested it in Firefox 15, IE9, Chrome 21, but it happens only in IE9.

Attached an mfa showing this bug.

BUG: I've encountered a difference between C++ runtime and HTML5 runtime, that makes not work a simple custom platform movement I've made with fast loops. In C++ runtime, when the "Player" active object jumps on a "Spring", it bounces up and then falls down again. In HTML5 runtime (and also Flash runtime), it starts going up forever without never fall down.
I've attached the mfa showing the bug, but unluckily this time I haven't managed to completely isolate the problem and understand exactly why this happens. I've managed to identify a condition in an event that, if removed, makes it work in the HTML5 runtime.

BUG: In the computeNewDisplay function of Objects.js, at line 1900 there is:

[COLOR=#AA0D91][FONT=Consolas]this[/FONT][/COLOR][COLOR=#000000][FONT=Consolas].[/FONT][/COLOR][COLOR=#000000][FONT=Consolas]hoImgWidth[/FONT][/COLOR][COLOR=#000000][FONT=Consolas] = [/FONT][/COLOR][COLOR=#000000][FONT=Consolas]rsBoxCx[/FONT][/COLOR][COLOR=#000000][FONT=Consolas];[/FONT][/COLOR]

but I think it should be:

[COLOR=#AA0D91][FONT=Consolas]this[/FONT][/COLOR][COLOR=#000000][FONT=Consolas].[/FONT][/COLOR][COLOR=#000000][FONT=Consolas]hoImgWidth[/FONT][/COLOR][COLOR=#000000][FONT=Consolas] = [/FONT][/COLOR][COLOR=#AA0D91][FONT=Consolas]this[/FONT][/COLOR][COLOR=#000000][FONT=Consolas].[/FONT][/COLOR][COLOR=#000000][FONT=Consolas]rsBoxCx[/FONT][/COLOR][COLOR=#000000][FONT=Consolas];[/FONT][/COLOR]

because while playing a MMF2 game I'm making exported as HTML5, the game suddenly freezed, and in Chrome's Console there was the error:
Uncaught ReferenceError: rsBoxCx is not definedPlease login to see this link.

BUG: If an Active object is next to a Backdrop object of type Obstacle with the property "Collision with box" enabled, in the HTML5 (and also Flash) runtime, a collision between the Active object and a backdrop is detected.

Attached an mfa showing this bug. If you run this mfa with the C++ runtime, the collision is not detected, and it is the correct behaviour. But if you run it with the HTML5 (or Flash) runtime, you can see that the collision is detected (and it's wrong).

If you disable the "Collision with box" property of the Backdrop object, everything works correctly also in the HTML5 and Flash runtime.

This week I'm busy, so if a new beta will be released these days I won't be able to look at it until next week. Sorry for that.
Next week I'll be able to continue beta testing!

BUG: In Movements.js, at line 2937 there is:

x=this.hoPtr.hoAdRunHeader.rhLevelSx-this.hoPtr.hoImgWidth+hoPtr.hoImgXSpot;

but probably it should be:

x=this.hoPtr.hoAdRunHeader.rhLevelSx-this.hoPtr.hoImgWidth+[U][B]this.[/B][/U]hoPtr.hoImgXSpot;

Infact DJFuego's game in Please login to see this link. crashes giving the error "ReferenceError: hoPtr is not defined. Movements.js:2937".

BUG: In HTML5 runtime, the action "Set RGB coefficient" has no effect.
Attached an mfa showing this bug.

PS: Tested on Firefox 16, IE9, Chrome 21, and the result is the same on all browsers.

BUG: The Layer object's "Set X of layer <layer name>" action is bugged in HTML5 and Flash runtime. Using it makes the layer's X Position change of a different amount than the parameter entered in the action.
Attached are 2 different examples showing the bug.

Note: the action "Set X of layer <layer number>" works correctly.

Started testing the new Beta!

Quote from K1kk0z90

BUG: Open an mfa and run it as an HTML5 app using the "Build and run" option of MMF2. The app will start correctly. Now, in MMF2, close this project and load another mfa. Run this as an HTML5 app using the "Build and run" option. When the browser opens, it will load a completely empty page. Copying the URL and pasting it to another browser will give the same result. If you close MMF2 and launch it again, if you load this last mfa and run it as an HTML5 app using the "Build and run" option, the app will start correctly.

This bug is still present in the new beta.