Colludium's Forum Posts

  • Multiply X by a fraction that represents each object variable - the sum of the fractions should add up to 1.

  • To add to my earlier bug report - when I Save As my working backup and replace the .scml and .scon file, a bunch of information is omitted from the new save:

    Here's an example from the working version:

    <entity id="0" name="player">
        <obj_info name="player_torso" realname="torso" type="sprite">
            <frames>
                <i folder="0" file="0"/>
            </frames>[/code:lit83fyl]

Here's an example of the same section of the .scml from the newly saved version:

[code:lit83fyl]<entity id="0" name="player">[/code:lit83fyl]

ie The obj_info information appears to have been omitted from the new save...

  • MadSpy - my thoughts as well - an automatic rename to avoid replacement, perhaps?

  • A bit of a random question to throw out there - is anyone able to shed light on why my html5 export has created an index.html and an index2.html file? It's on Dropbox (my umbra demo) and nothing links to the second file (dated a day later than the first one, both appear to contain exactly the same information). Is this a c2 bug and will my demo explode if I delete the index2 file? Thanks.

  • A cloud has appeared over Umbra game-dev land. A dark and oppressively-angry looking cloud...

    I am experiencing a problem with Spriter that has stopped my game's progress.... Like any good paranoid developer I have been collecting zip archives of all of the important source files for my game every few days. Just recently I found that if I imported the player spriter .scml into c2 it caused c2 to delete all of the player's image objects. After some investigaing I've learned that simply importing a newly saved version of the project into c2 does not work. Even importing the Spriter file into a blank c2 project shows the same bug, so it's almost certainly a problem in Spriter/plugin compatibility (the image files are not being dragged into the project with the .scml - a file association is broken somewhere). This happens on more than one computer after editing a backup or current player Spriter project. I have noticed a Spine plugin is being developed - I don't want to re-animate everything in Spine (all 80 player animations!) just yet.

    So, fingers crossed that this can get fixed. Meanwhile, I think I need to do some investigating of alternative engines. For example, Corona SDK now builds for desktop, don't you know....

  • Hi lucid,

    Do you mind if I pm you a zip of a spriter project I'm having difficulty with? I seem to have reached some sort of limit with spriter; if I duplicate and edit a couple of animations and then import the updated scml into my game, it causes all of the spriter object's images to delete from the project... I have 80 animations and I'd rather not make it available publicly if possible. I have reverted to a backup copy and everything imports normally into C2 again, but it seems that I am unable to make any more changes to the animations without the import into C2 failing.... I'm using C2 r211 (although this happens in each version down to r206 minimum), Spriter r4.1 and the latest Spriter plugin, tested on both W10 and W7...

    Thanks.

  • It depends on what sort of game or app you are looking to create. The best thing to do IMO is create a test layout with the behaviours and object numbers on screen that you would expect - then run it over wifi in preview and see whether it meets your satisfaction.

  • Aha! I had my suspicions - the asm.js physics is still rather beta and, in my opinion, may never reach full functionality because its use here depends on 3rd party updates.

  • Out of interest, is that with box2d web or asm.js?

  • I thought this was going to be about the not rather hacky steam plugin :p

  • Try Construct 3

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

    Try Now Construct 3 users don't see these ads

  • Do any of your sprites have behaviors that are not disabled? It's almost impossible to guess what could be going on from a gif....

  • Same here.

  • I expect that the problem you highlight will only ever get resolved if it is addressed by the emscripten developers - which I find unlikely unless we all start telling them of bugs in their code. I understand that they have an algorithm that to convert c to javascript but it clearly causes subtle errors in function. Although it's a great concept, without the backing of a major developer behind it, I have little confidence that this problem will get rectified (because they are an unpaid 3rd party to scirra). There are also joint problems associated with asm.js that have been documented for many months now. I personally consider asm.js to still be in beta and my advice would be to use it accordingly - I think we'll have to wait for the next (annual?) release of asm.js box2d to see if that will fix these problems while not introducing any more.

    I personally think scirra would be better off focusing on fully implementing one proven and versatile physics engine as a full-up plugin (ie full box2d?) as I think that would be a better selling point than indefinitely waiting to see if emscripten will one day become reliable.

  • I would be interested in learning about whether this could ever be possible. I recall the attempted 0.5*vsync change to the engine to allow us to set 30 fps - which, for anyone trying to create a game for mobile, was very attractive: 0.5 * the CPU and GPU work per second.... And that was a failure, unfortunately.

    When I first thought about this having read the OP i thought that there could be a problem that stops this being viable - Tweak Guides on tearing. However, I'm not so sure now - that article is discussing GPU vsync rather than browser sync with GPU sync (vsync) of the monitor, so tearing may not may not be a problem. I don't know if it is possible to make a browser run effectively at half vsync using c2 because the 30 fps attempt went so badly.

    But I agree - at 120 Hz refresh rate, the single thread of the browser that's running your game is working to do the same calculations in half of the time when at 60 Hz. And that means that some monitors could, in theory, cause frame drops just because they have such a high refresh rate.

    Often GPUs allow the user to set the refresh rate - perhaps the solution is to state that, in the game hardware requirements, the user should ideally set 60 Hz.

  • Colludium - THAT is actually a very good idea. Running faster is running faster!

    Yes! I totally agree. I use that technique mostly for traps and puzzles in each section of the game - if the player is somewhere else then the groups of associated code are disabled. To make all of that work, like others have posted, I have event sheets for: E_player, E_effects, E_enemies, E_traps, E_common, E_LayoutX - and most of the group enabling/disabling takes place in E_traps, E_enemies and the event sheet specific to the current layout.

    It's the same philosophy for turning off work for other behaviors - but doing that might earn you bigger dividends. For example, I turn off particle objects if they are far away from the player. That lets me add a LOT of particle objects without loading up the cpu by making it have to track thousands of particles that cannot be seen. When I introduced that little trick to umbra during dev of level 2 the cpu utilization went down by 10%! So, depending on what your events do, enabling/disabling groups may not make any discernible difference to a game (but could cause you a bit of pain setting it up) whereas managing off-screen behaviors might be worth giving more attention to (I just do it out of caution so I don't later find I have to go back and try it).

    Regarding the code groups - the way I imagine it is that it enables you to block off great swathes of code by setting a single boolean. That boolean will be checked each tick - so each disabled block of code will add a small but relatively insignificant burden to the work the cpu has to do each tick.