Windows XP SP3 32-bit

Centrino Mobile

Construct r78

Attempting to edit existing Family instance variables by right-clicking on the Family, and trying to change values (not the name or type) just attempting to change initial values results in:

"Name used by object type in this family"

The object 'XXXX' in this family is using this name 'FireInterval_Min', so your variable needs a different name.

<img src="http://dl.dropbox.com/u/12667027/Construct%202/Family%20Instance%20Variables%20Bug.png" border="0" />

<img src="smileys/smiley19.gif" border="0" align="middle" />

EDIT: Forgot to mention, Archer_Ally is a sprite with no instance variables of its own.

Windows XP SP3 32-bit

Centrino Mobile

Construct 2 r78

Getting this error constantly. Every time I try to create a new sprite, I get this same error. (cries) I can't even move forward with my project now. However, if I use Ctrl to copy a family sprite already on the layout it doesn't throw an error.

Freesound is a pretty nice place for sound effects, and all of their sounds are licensed under Creative Commons, you can also preview the sounds before downloading them, although you will need to register for an account before doing so.

Hmm, I just did some more checking, sometimes it seems like with this code when the CPU usage is maxed out and the frame rate drops, the jump release key code sometimes won't detect, so even if I just barely tap the key, the player sprite will jump to its furthest height.

Pile of barf bags. I forgot with my solution you'd have to worry about other overlapping sprites and pictures and backgrounds. HERP DURP DURP.

EDIT:

If you were willing to create solid color "cut-outs" (like masks) of all the animation frames, you could do the pin, making sure to update the overlay frame to correspond to the currently displayed sprite frame. The only benefit to this approach versus just making a copy of all the sprites frames in the colors you want would be that you can combine multiple overlays with additive properties at once allowing for blends or combinations, or change the opacity of the overlay without affecting the main sprite's opacity.

Scirra, regarding other languages and the difficulty inherent in moderating them, I thought I'd mention that I can read and write Chinese, so if you guys ever need help with moderating a Chinese forum (like how the XNA App Hub forums have specific language boards) or anything else related to that, I'd be more than happy to assist in anyway I can. Let's promote Construct in any way, shape, form or language we can! ??????????????????????????????Construct??????

-- cacotigon

EDIT: lol, that was amusing, the Chinese I typed just displays as a bunch of question marks. Yay unicode!

You have two possibilities:

You could create alternative colored images for every picture you have. Kind of a pain though.

What I would do is:

1. Create a bunch of solid colored blocks in paint.

2. Make the block be wider and longer than any sprite you intend to use it with.

3. Load all of these into a single sprite called OVERLAY on different frames, set animation speed to 0, and give this sprite a pinnable behavior.

4. Write a bunch of global variables COLOR_RED, COLOR_BLUE, COLOR_GREEN which numerical values correspond to the OVERLAY's respective frame.

Whenever you want to change a sprite's color, simply do the following:

1. Have that sprite spawn an OVERLAY

2. Set OVERLAY frame to the appropriate color

3. set its effect to ADDITIVE

4. Pin it to the sprite (assuming origin is the middle point of that sprite).

You can destroy it later, or change its visibility, etc.

Hope this helps,

-- cacotigon

EDIT: I whipped up a quick capx file to demonstrate it:

Color overlay capx project

I have a design/structure question related to player platform behavior movement. Custom controls is set to true. The main question is related to ON_KEY_PRESS vs ON_KEY_DOWN.

Here is the keyboard code:

<img src="http://dl.dropbox.com/u/12667027/Construct%202/Keyboard%20Movement%20Handling/Player%201%20Keyboard%20Events.png" border="0" />

Here is the actual movement code itself.

<img src="http://dl.dropbox.com/u/12667027/Construct%202/Keyboard%20Movement%20Handling/Player%20Movement%20Events.png" border="0" />

For example, the jump is implemented as ON_KEY_PRESS, but I could have just as easily implemented it as ON_KEY_DOWN.

Which do you think is better? Sometimes it just feels like my character isn't too responsive when jumping, especially when attempting to do lots of little hops.

Keyboard logic is separated from movement logic for purposes of handling multiple players.

Looking at this code, what do you think? Is this a poor way of implementing keyboard controls, player movement and animation?

Thanks!!!

-- cacotigon

Woot! Say goodbye to really long nested pick / reset clusters!

By the way, speaking of feature lists, have you guys ever considered implementing something along the lines of a voting system? For example, Scirra would show a long list of potential features on the website, and every paid user is allocated X points (for example three), to which they can spend on any of those features. At the end, the features that are more important to the users would float up to the top. At any time, a user is free to deallocate one of his points, and put it in something else, or even put all three points in a single feature that they really want. Whenever a feature is implemented, all of the points that users had allocated are given back to them.

I would devour an entire planet full of unicorns for a UID indexer, something like how object() parentheses takes IID, using brackets would refer to the UID, so you'd have object[UID] and object(IID).

This would save me so many times where I've saved off a certain UID during runtime, and in another expression (which has already done a pick operation on the same object type but a different instance) I suddenly need to get one of the saved_UID object's inner values like speed, velocity vectors, etc. Right now I have to setup lots of { reset pick, pick object uid, save object's vector to local variable } subevents.

-- cacotigon

Hmm, I messed around with saving IID values instead of UIDs, since the arrays are going to be around for the duration of the program. It appears that accessing an object via its IID, aka object(IID), is not affected by any prior picking; I don't have to do any resetting or worry about selections. And unless I'm mistaken, the IID values should remain constant for a given object type *provided* that nothing is destroyed, right? Is there anything I should be concerned about when using IIDs?

KYATRIC. Thanks that clears it up, I didn't realize that the ( ) index was a reference to a IID. I'm going to have to re-read the manual about picking/indexing/IID/UID so that I avoid making this kind of mistake again.