Hi everybody, aridale & Ubivis

gotta revive an old thread/question but theres a new thing nagging me with my project.

I managed to layer all parts correctly but as I'm using rex_zSorter from rexrainbow I'm getting problems with all the small sprite parts. I have 2 ideas for a workaround so far:

1) increase size of all sprites to match and let rex sort them out.

• I have the feeling that this won't work and it would be a lot of work to change all sprite images for this attempt

2) only sort the base/collision sprites of all mobs and get the z-index from those to sort all sprite parts afterwards

• every tick
• get z-index of base sprites
• start loop for all mobs
• - start loop for all parts of each mob
• - sort parts based on internal z-index for sprite parts with a function
• - end loop for parts
• end loop for mobs

Does that sound realistic? I have about 12 - 15 sprite parts for each mob and on a busy screen there might be about 10 - 20 mobs which leaves us with something between 120 and 300 sprites to be sorted every tick...

Before I start coding that I'd like to know what you think. Should this solution (2) work and might this result in performance issues?

tap

I just wanted to ask about the status of android and support for directCanvas as well as multitouch. I browsed the last 4 pages but couldn't find anything on that topic.

As far as I can remember, export with directCanvas for optimal app performance was only available for iPhone/iPad.

Could you point me to a place where I can see current features as well as a roadmap with comming features (such as full android support)? That would be a great help as I'm only developing for desktop and android (handheld/tablet) right now due to not owning any i-devices. I couldn't find any definite info on the mobius website... not sure if I looked at the right place.

Thanks a lot, your work so far and any help is appreciated!

MC

Thinkin about it now you can see why most 2d games dont have a visual gear effect right?

But people love eye candy (in addition to good gameplay and stuff) ^^

Customization is a big plus for many people. Gotta find a way to get it to work if I want people to actually like their characters.

I'll try and find a good way to "bake my layer cake" without making a mess in my development kitchen. I'll probably write a "change animation" funktion so I only have to change animations in one place and I can call that animation whenever an event calls for a new animation.

Ubivis

I've read a bit about Spriter during the kickstarter campaign. Could you explain the planned Importer Plugin in 2 - 3 sentences? I'm not aware of any details.

Hi

I managed to implement multiple characters and other mobs with lots of animations in my 8-way isometric game and everything is fine so far. I worked out how to trigger the right animations, sort by y, custom shadows, jumping and basic movement. As I'm planning to make the player character and enemies customizable (custom cloth/armor, weapons, hair styles) and don't want to include every possible combination (my game size would explode) I'm looking for a way to build my mobs as a combination of modular parts.

Is there a way to achieve this in Construct2?

I have humanoid mobs with

• lots of animations (walk, run, jump, fall, attack 0 - x, cheer etc.)
• customizable weapons (sword, shield, dagger, mace etc.)
• customizable armor (head, back, torso, arms, legs, shoes, hands)
• customizable body (currently only different skins)

I would maybe implement different character sizes and proportions but that depends on the possibilities of Construct2 (or custom plugins/behaviors) to assemble a character from all those parts.

My first idea was to create all parts with all animations, create a layer cake character and trigger the animations for all parts. As this would take quite a lot of time, I wanted to ask the community if there is a better way do go about this that would either save me lots of time or provide me with a scalable solution.

One Problem I see is the sorting of all parts. To keep it more simple I wouldn't sort all parts but the collision objects of all items and characters in the game (which are on a lower layer). This way there would be only one collision item per character but about 9 (or more) animated sprite parts and I could sort all parts as one entity.

I'm not developing for the arcade as I allready have lots of plugins and behaviors from rexrainbow 00Rez and Pode so any solution involving behaviors and plugins is welcome as well!

Ok, I'll have to see myself then when I'm done prototyping.

Thanks a lot!

cmr333

thanks for the quick reply.

Is there a way to determine or even influence, how much of the game is then installed to SD? I had mixed experiences with games from the android market. Some shrunk from 8mb to 780kb and others only went from 6mb to 3mb.

Hi,

as I need MultiTouch (or keyboard) for my game to work, I'll go with appMobi. Therefore I visited their site and community portal to search for info on "Save to SD". I couldn't find anything. A quick forum search for appMobi + "Save to SD" didn't yield anything here either...

I think there should be a few people out there by now that have already finished a C2 appMobi game. Could you tell me/us what manual steps you had to make to implement Save to SD on android.

[pre posting edit]

I realized that directCanvas isn't yet released for android.

1) Does the regular appMobi wrapper that uses the stock browser (without multitouch) have "Save to SD" support?

If not, I'll have patience until directCanvas is released for android. I have good faith that it won't be long. ^^

2) Does the iPhone have (or even need) a "save to SD" option?

If the pin effect is constant, no need to overload body code and "on every tick events" more than needed !.

In my case, the shadow is pinned to a character that can move 8 ways and jump (like "Double Dragon", "Streets of Rage", "TNMT" etc.)

When the player (or any other mob) jumps, I used to unpin the shadow, move the mob up a bit, shrink the shadow and repin it. Worked fine until I tried jumping and changing the direction mid-jump (I'm not sure whether I'll deactivate this) and the shadow got out of sync.

This is why I created a shadow function that moves the shadow of each mob according to the mobs X/Y (collission sprite) as well as Z value (animated sprite). By seperating shadow calculation and jump function I can implement in-jump movements easier (double jump, interceptions and collisions, in-jump attacks) but bloat my code a bit more.

I'm running out of events soon and don't have the allowance (of my better half) to spend money on this pet project of mine. She said I should build something like Bejewled or a BubbleShooter...

Maybe I'll start an IndiGoGo Project or something to gather the money for some better artworks, sprites and a license to keep going.

I'm not sure I get what you mean.

Useless as in "shouldn't be used"? Allways do x, then y, then z is more or less what I thought it would do. That's why I had concerns about the performance when doing all those actions for "5*number of mobs" sprites + lots of other things "every allways" <img src="smileys/smiley2.gif" border="0" align="middle" /> compared to a pin behavior.

I'm interested in this as well.

I'm working on a "pseudo" iso 8way platformer with "jumpable" things and a z attribute for mobs that jump or fly.

All my mobs (player and enemy) consist of:

• an invisible collision/movement sprite
• a shadow sprite
• a player sprite for animations
• an effect sprite (emoticons for sleep, anger, talking etc.)
• (optional) a health bar sprite

When the player jumps, I increase his z-value, unpin him, move the animation sprite up to his z-value and repin him. Meanwhile I shrunk the shadow sprite while he's in the air.

Worked perfect until I started smashing the buttons/keys and had him move around while jumping. Then there where moments, when his animation sprite and his shadow and collision sprite lost their relative positions because while unpinned, other movement events changed the position of the collision sprite but didn't update the position of the animation sprite.

This could not happen if I maintain all sprite positions based on "every tick", the position of the collision sprite and sprite variables like z. Is it better to replace the existing PIN behavior with events that update the position every tick? I thought that with more mobs on screen PIN might be better performance wise...

As mentioned in another thread I'm thinking about building my own behavior. I'm working on game prototypes that use grid movement (8way) and currently all my mobs (pc and npc) are using the same logic for movement (plus some basic pseudo path finding for npcs).

Thanks to the awesome help of the community I managed to achieve a good grid movement, that snaps to the grid, checks for walls and even mob collision. Would it be possible to create a behavior that does the same:

• check if mob is moving
• set movement distance (set to grid size via action)
• set movement speed
• support state/condition/variable to set mob as immobile
• support temp. deactivation of control input
• check for collision / overlap with solid objects (for walls, objects and other mobs) as 8way does
• allow custom controls (should be easy to recreate as

  rexrainbow already did it in his 8wayMP Behavior)

Bonus (+100pts)

• allow to set a "move to" grid position (based on grid size) and simply set movement direction to that angle (snapping to either 4way or 8way grid angles as mentioned in this thread)

Bonus (+10.000pts)

• check if next block is empty and if not, go either left or right to avoid simple obstacles (And enable/disable this feature via setting or action. It should be only required for AI movement)

Not sure, I got all requirements but that should be about it.

1) Is it possible to build this as a behavior?

2) Is there maybe a behavior that does this already? (not looking at path finding behaviors because they tend to be to slow for larger numbers of mobs)

3) Any tips, where to start or how to tackle this biggie in js?