Greetings Constructors,

First, I didn't post this subject in the How do I section because according to me, it's more a general question about the overall functioning of Construct 2 and its physic behaviors than a simple request for help. I don't really need help anyway (although your replies on this topic could be of use, of course). If I'm wrong and this subject belongs to the How do I section, sorry, I'll respost it there and delete this one. That's being said, let's get back to the real topic : PPC (Physic, Polygons, Collisions).

Under precise conditions, the little armless dummy that I use as main character/sprite can go through the platforms around him. The player's sprite overlaps the platform's sprites while the collisions are disabled. It works efficiencly without any problem when doing any vertical movements, however, I've noticed that it may causes some odd stuff when doing any horizontal movements. I may precise that I use Chipmunk physic behavior, and not the basic physic behavior.

To be clear, when jumping through a platform above or beneath him, everything is fine.

When running through a platform however, I have seen three main cases :

1) The character is slowing down (like he was suddenly going through some kind of big water wall or something) but manage to pass

2) He remains stuck in the middle of the platform (pretty rare tho, but fact is, it happens over time)

3) Collisions are still enabled when running directly toward a platform, but they are not if I jump through it

It's pretty odd, as in the event sheet I specify that under the concerning circunstances, the collisions don't apply anymore...so what's going on exactly ? I was sure the character will just run like nothing was there anymore, the sprite being drawn but physically inexistant. I was wrong, but why ? Also, if I suppose we can't change anything in the behavior to fix this mess, I'm sure there is a tricky way to avoid that thanks to the events, but I have no idea how, honestly. Any ideas ?

ErekT Thanks for the names, really useful ! Yep, I'll ask them !

No problem mate !

Interesting. Looks like some kind of subtle hybrid between the Shadow Caster Behavior and Dynamic Lighting. You can combine the two, after all.

Take a look at these :

https://www.scirra.com/tutorials/9401/e ... ing/page-1

https://www.scirra.com/manual/178/shadow-light

It's either full Shadow Caster Behavior with clever use of sprites color, either the said combination.

NetOne I see. In this case, the best answer I could give would be "No", I suppose, for my current project at least. Clever answer sir, thanks.

newt I think he was quoting Aurelien Regard regarding the said contracts. Not his own experience.

glerikud Ok, alrighty then, I'll do it someday !

As for my case, I'll finish my prototype as far as I can thanks to C2, then build everything in Unity, Java or C3 if grass is greener there. It's...safer. My current sandbox only has more than 150 sprites (characters or environment), dynamic lighting, 8 layers and uses a physic engine. Well. It's fine right now. But a whole game with several levels ? Correct me if I'm wrong but based on what I saw in the forum recently, I'll end up with some serious problems at the end or even before if I keep working on this in C2. This game project is simply not suitable for these tools. They are not designed to support medium or big stuff...so I'll keep them to continue my experiments and to work on smaller stuff.

This actually looks freakin great. Nice job !

glerikud Not a problem at all, however if I recall correctly, you already know Unity, so if I decide to use this tool instead of C2 while I'm waiting for C3, I don't think my opinion will be really useful for you. Nobody uses Flash anymore, so once again, my point of view about it would be pretty useless. If I use Haxe Flixel or Godot or Fusion or anything which is not Unity, and if you wonder how useful and powerful the LibGdx library for Java truly is, I'd be glad to shate my experience on the matter, yep.

ErekT I'm gunning for mobile (and C3 seems to be a good way for it). And yes, nothing is perfect and I'll always end up frustrated because of something, that's for sure. But when it comes to export or to check performances in anything which is not little games, grass is known to be greener elsewhere. And my doubts are about performances and export...I'm not worried by anything else in C2. But yeah, as I said, I think I will start to remake everything I have in Unity (or Java, or Godot, or Haxe flixel) while waiting to test C3.

This way, I should be able to decide what to do. Sticking with C2, sticking with Unity/Java/Godot/Whatever, or using C3. I need to gather as much informations as I can about whatever alternative to C2 or C3 I might consider, to perfectly acknowledge its pros and cons. Once again, it's all about this specific project, Scirra's tools remain the best when it comes to create small but advanced and sympathic contents. Everything is still possible for my case. I may be end up working on this in C2 or C3.

Thank you guys for your answers, I'm gonna answer you as well !

glerikud Hum, I don't know. I've been through the whole post-publication entries written by Aurel on The Next Penelope for example, and it's pretty much what we all fear and what we all know already : The guy respects Construct 2 and claims that it's a fine tool (which is true), but also says that the said tool is a real pain in the ass on several aspects to the point that he's now working on the Unity version of his game. It's a grey area. No black nor white here. It's great, and it's not at the same time. In the end, this post-publication is not really reassuring for someone in my current position.

newt Although your answer is entirely true, it's not related to my doubts. I perfectly know where I want to publish my work afterward. It's my first commercial game, I'm nobody, and I'm probably working on nothing. I target Desktop and mobiles. For a first product made by another anonymous indy, it'd be entirely enough I think. If the game is a successful one, there, I could consider an export on consoles or something. But for now, such an export is nothing but a vague dream. Not even a dream, a thought. So, in my case, using C2 and C3 is not an issue regarding this point. Consoles doesn't have my consideration for now.

Prominent It makes sense. And I'm a huge adept of the Experiment and suffer to grow stronger and better concept in life anyway. However, as game makers (hobbyist or not), we also need to stay pragmatic and logical. Today, I have a sandbox. A great test area, to play with physic, animations, gameplay elements, lighting. I experiment. I make experiments. And I learn. While learning, I still wonder if this is the right engine to use, because I could be screwed at the very end or even before. And I'm still able to step back. To recreate everything elsewhere, using what I learned and will learn thanks to C2 or C3. So...what's the most logical way available ?

— and tgeorgemihai I see. You're right. I think I'll try Unity, and see how much work it is to remake what I already have on C2. By that time, C3 Beta should be available, so I'll be able to see what it can offers as well. To learn, to see, and to choose between all the possibilities. Maybe even with little old friend Java again.

I'll take a look at Haxe Flixel as well.

NetOne I don't really understand your point. I mean, your answer is basically :

You fear to work on something which will then become a tiring hell on Construct 2 if it's bigger than expected. Hey, take a look at this link which talks about one of the most known C2 user who confirms your fears about the export and the performance aspects.

The story of Aurelien Regard is one of the sources of my worries on the matter. So why do you say I should continue to work on my first big project with C2 ?

STARTECHSTUDIOS I've heard that there are several problems when it comes to export to mobiles. C3 will apparently fix that. On Desktop as well ?

Greetings everyone, the following text may be long.

First of all, no, this is not a Oh Shit Construct 3 Oh shit related topic. The title implies some doubts regarding a project, not the tools themselves. I freakin love Construct 2, I love this community and I'll take Construct 3. That's being said, after reading a bunch of stuff in Construct 3 General discussion, I've come to realize that the project I'm currently working on (for almost two years now) might be a bad move on Scirra's tools. The engine is great, but maybe not the one I should be using for this.

Context :

Some years ago, I was disregarding the softwares and tools like Construct Classic, Unity and co because from my point of view, they were nothing but poor attempts to help lost users to create uninteresting and basic games because coding was soooo complicated for the plebe. You know, this kind of guy so proud and arrogant because he knows how to make a freakin stickman move on some freakin platforms thanks to LibGdx/Java or AS3/Flash. Useless to say, I was completely wrong, but the fact is :

I've made some little games using Flash and Java for a time.

Then someday, some dude told me that I could do the exact same stuff - even better and faster - using Construct 2. I was kinda condescendant at the time, but this same dude being someone I respected, I decided to take a look and actually learn some facts about this kind of tools instead of sticking with my dumb ideas. One year later, I was buying a licence. And I don't regret anything to this day. Yup, that's my story and the context, now let's go back to the problem.

But now :

I've made some stuff. Silly, serious, polished, bullshit or not. Various kind of stuff. Using Java, Flash or Construct 2. But in 2015, I have finally started to work on a serious project, a little 2D world which really matters for me, the first game which I could eventually sold someday (I'll probably not earn anything but still, to achieve a long time dream and for the sake of the process, you know, the usual speech). I've started it with Construct 2. As you know, it's awesome to quickly build something which works, so it's easy to create prototypes. And that's basically what I have for now : A prototype. A sandbox. An environment wich exists only to test the AI, the animations, the gameplay, the fluidity, the rendering, the lighting and so on. A little something to make several experiments and adjust everything before starting the creation of the real levels.

I can still go back. I'm not too advanced yet. I have the core. Not the world. But I don't know if I should. After reading the Official Information Thread of Construct 3, I've learned that Construct 2 seems to be a huge hell when it comes to create and publish big or medium length projects. I don't really know how big will be mine, I'd say medium for the moment, but still. Now that I'm working on something serious, what is better ? To polish my prototype using Construct 2/Construct 3 then build it with another engine, or to continue with Construct 2/Construct 3, despite the risk of ending up with something unadapted to this engine ? I feel kinda lost these days.

This is not a tiled game, there are several layers and several assets...and it's just a sandbox right now, not even a real level. I don't know. The tool is great, the event system is amazing, but I fear to find myself stuck with limitations or problems from third parties or god know what while working on this. I'd like to hear your opinions, guys. Thank you

morton Hansel Yeah, that's a Tutorial, that's the concept !

sivricmarijan You're welcome, I'm glad if it helps !

NanoCake I didn't note any performance impact. I use it in my own platformer, and It's quite light. Calculations to generate your shadows are far to be heavy (as explained in the tutorial. It's nothing but depth simulation after all, your sprites are still flat as hell, no matter the resolution). Any problem at this point would be from your own design and not from the feature itself I think. For instance, if you set a lot of light sources and use a lot of normal maps for some reason (knowing that you normally don't need more than one or two normal maps and don't use more than 3 or 4 light sources for the entire scenery/level...) you will have some problems. But in regular case, which means one single normal map to cover your whole level and some light sources (or maybe even one single light source like a sun or something), I really doubt anything could go wrong.