Animmaniac's Forum Posts

  • R78 32 bit win7.

    Description: expression editor doesn't allow to set the second parameter (b) of max(a,b) and min(a,b) to a variable. It drops an error: "Invalid Parameter - Parameter 2 of 'max': must be a number"

    <img src="http://dl.dropbox.com/u/7871870/construct/bug-06.jpg" border="0" />

    This happens in any expression, not just in the "Set text" like the example.

  • Try Construct 3

    Develop games in your browser. Powerful, performant & highly capable.

    Try Now Construct 3 users don't see these ads

  • I must agree with what VampyricalCurse said about performance earlier. Html5 only runs OK in powerful machines. On old computers (sometimes not too old) they perform badly. Don't need to say anything about mobile. Currently, if you want your game to run smooth you need an audience with powerful hardware and restrict your creations to simple games heading the casual market. You can only have the liberty to create more elaborated games if your audience has top of line machines, and even then you don't get an ultra performance.

    Then there's this paradox: the users who have powerful machines are mostly hardcore players for whom the casual games don't appeal that much, in the other hand the majority of casual audience have slow machines. It's a complicated situation for Html5.

    I'm not an expert on the subject, but from what I can get it seems that besides fully supporting WebGL, the only thing that could make Html5 faster would be to drop javascript in favor of a faster language. But unfortunately I don't see that coming soon. :/

  • Hmm, pop-ups are annoying. I would rather put a comment in the parameters description explaining the performance hit.

    I must say that I also miss the ability to use multiple frames and animations. Being able to spawn particles with random frames helps add variation as well as animated particles.

    In 'Can't Turn It Off' I used a single object with multiple animations to make all the particles. It would be handy if the Particle plugin allowed the same.

  • Strange behavior. When I changed it before, everything was working fine. Now that I re-opened the file it doesn't work every time. Almost seems like if I stay messing with it and previewing for a while it works. Then if I close and come back later it doesn't.

    By the way it works fine in Chrome, and I'm running in win7 32bits desktop.

    Edit: I just noticed that for some reason it doesn't work in Firefox, only in Chrome.

    Edit2: Forget what I said, it works randomly and buggy. Let's hope someone can shine a light on it.

  • Here's a fix:

    ninjaTurn Fix

    The problem is caused when you flip frame 3 instead of the frame after frame 3 during the turn. So it gets flipped one frame before it should.

    I fixed by flipping frame 3 in the image editor to keep the events simple the way they are, so when you flip it in events it will be in the correct orientation.

    This solution is kind of hacky though. It would be better if there was a way to check if the ping-pong animation is forward or backward.

  • I don't know why it's not working sometimes, maybe it's a bug.

    I modified the events a little an now it works:

    Corrected example

  • Well, by definition code is "a system of symbols, letters, or words given certain arbitrary meanings". In Construct you use visual symbols with arbitrary meanings to create a computer program. So you are also coding. If you study semiotics you will probably agree with it.

    For me the best is to call it programming since it's a broad term and everybody understands.

    Some people like to use the terms to promote some languages and depreciate others, but they are all the same in essence. The only thing that changes is the language, that is nothing but a different system of symbols and rules.

    Making analogy with language: some people program with phonetic languages, other use ideograms, some can't see and use braille, others can't hear and use gestures, some are masochists and prefer Morse Code. In the end they are all doing the same, programming.

  • Here it is:

    Physics drag and drop example

    Since there's no way to destroy a joint after it is created you need to destroy the object.

  • For me is because of precision issues that varies depending on the rotation of the object and aggravates through scaling.

    I started using the physics behavior and it's circular collision, but since it lacks the feature to disable collisions in runtime I decided to do my own movement through events. Now the problem is not the inability to disable collisions anymore, but the fact that I can't check for circular collisions.

    In another project is a different situation. I used the physics behavior with circular collision and test for collisions with other objects to apply forces. The problem is that when you use Is Overlapping or On Collision it checks for collision using the sprite's collision polygon, but in the same time the physics behavior is using circular collision to interact with other objects. This difference between the collision models, where the space the object occupies is different than the check space, generate errors and unexpected movement.

    Anyway, just wanted to make you aware of the problems I'm having.

  • I support your petition and must say that I'm also in need of such feature.

    A drop-down in object properties like the physics behavior has, with the options for Collision Polygon, Bounding Box and Circle would be enough. A None option to ignore collision detection would be welcome too.

    In the future I think would be better to add collision primitives through the image editor, but for now I would be happy if I could use circular collision in any way.

  • OctaveEight

    It uses a frame with "no collision polygon" to ignore collisions.

    Although it doesn't work perfectly since C2 doesn't allow a frame to have no collision polygon. If you jump over the top-left spot of an "deactivated" platform the player collides with the tinny collision polygon, when you are expecting not.

  • Yeah, I imagine it would need a big deal of code rewriting, but you can't deny it would be a great feature for C2.

    But I understand your situation, it's a difficult task for a two man team.

    Well, at least I can dream about it. And it also serves as a good exercise in problem-solving for me!

  • Ashley

    I would expect the family object to collect all the ACEs of all the object types that belongs to the family. It could filter the repeated ACEs and display them just once. The most common, or the ACEs that belongs to all of the types could be grouped in a Common category.

    Then if you add a condition that only belongs to a specific type of object, during execution it would first filter all the family objects that have that same condition, and find the matching objects to be picked and added to the SOL.

    For actions would be the same - first it would filter all the objects that have that same action, then for each object type of this list it would check if there's specific instances picked. If yes apply the actions to those instances. If not apply the actions to all of the objects of that type that belongs to that family.

    This would mean that if there's an action that belongs to two kind of objects (sprite and text), and only one type of object is picked in the conditions (sprite: instances 2,3 and 4), that action would be applied to the picked instances of one type (sprite: instances 2,3 and 4) and also to all instances of the type not picked (text).

    Then by adding a Pick object type condition to families, would be possible to disambiguate which object you want to apply the actions in case you want to affect just one type of object.

    For expressions it would also work the same, with the only difference that when there's no specific instance picked that match that expression, it would return the value of the first object in the list.

    I really wish that families could support multiple object types, specially to allow instance variable comparisons between different types of objects that could be used similarly to classes or tags.

    Take a family of the respective objects and it's variables as an example:

    Family "Scene":

    Sprite: varClass = "rock", varShow = "False"

    Sprite2: varClass = "grass", varShow = "False"

    Sprite3: varClass = "rock", varShow = "False"

    Text: varClass = "rock", varShow = "False"

    Text2: varClass = "grass", varShow = "False"

    TiledBG: varClass = "grass", varShow = "False"

    TiledBG2: varClass = "rock", varShow = "False"

    You could do this:

    Family "Scene": variable varClass == "rock"

    ->Family "Scene": Set variable varShow to "True"

    That would result in:

    Sprite: varClass = "rock", varShow = "True"

    Sprite2: varClass = "grass", varShow = "False"

    Sprite3: varClass = "rock", varShow = "True"

    Text: varClass = "rock", varShow = "True"

    Text2: varClass = "grass", varShow = "False"

    TiledBG: varClass = "grass", varShow = "False"

    TiledBG2: varClass = "rock", varShow = "True"

    This could open a whole new can of possibilities.

  • Are there any plans to include more than one object type in a single family?

  • I would love if there was a way to reference objects in actions like HotGod described. The ability to associate the precise instances picked in the conditions with the actions is something I desire since I started using CC.

    The work-arounds to achieve the same picking precision are far, far less intuitive than if such feature existed.

    Lets hope it can be added to C2 someday.