• Content Count

  • Joined

  • Last visited

  • Days Won


stupot last won the day on July 28 2018

stupot had the most liked content!

About stupot

  • Rank
    Advanced Member

Recent Profile Visitors

2573 profile views
  1. I modified the plugin too, for this very same reason
  2. you need to look into 'Web Workers'
  3. I've used the closure compiler in a phaser project with maximum optimisations, the output was smaller and faster than uglify2. But it took a lot of work to manually create and maintain all the required externs, I don't do this anymore becasue of this, mobiles have become much more powerful to compensate. Phaser has a lot functions/properties created programatically that cannot automatically be scanned for, so debugging the optimised code back to the source to see what else whould be extern'd was a nightmare. Luckily I only used a minimal portion of Phaser so didn't have to do a full job on it.
  4. Found an apparent mac download here:
  5. there's a project plan here: according to this it's currently (possibly) in development
  6. Hi akwa6971, The image hasn't loaded by the sounds of it, check your console for error messages. Is the image file name/path correct, is the upper/lower case correct? Also. this post should be in the main Phaser thread, rather than this Phaser3-Beta thread - you'll get more help there.
  7. Thanks @rich sounds like a good plan, that's given me some food for thought.
  8. Looks like I'm going to have scratch a little deeper than the surface, conceptualy Phaser3 looks to be quite different to Phaser2. I've a few different ideas about how parent transforms could be implemented and applied only to required objects but I need to get my head around the new P3 internals first. Any pointers or suggestions you might would be useful now.
  9. Removing always-there pre-post updating/rendering/etc is great for performance but will slightly increase the difficulty of using Phaser whether an alternative is native or pure user implementation. I'd like to see a native solution to this otherwise the forums will be filled with people asking how to hook in an update function and getting several differing answers. It's common enough functionality to expect something built in to standardise the way people do it. Looking at UpdateList.js now though reminds me of several problems I've had in the past where the triggering criterea was hardcoded into it, ie here it depends on How about a more general callback processor - like a simple event/subscriber model? The system would have several built-in reserved events ("preUpdate", "postUpdate", "preRender" etc) which any object could subscribe to, along with any custom user events. Possibly the system could have the ability for subscriptions to also register the triggering criterea, maybe by a list of properties with boolean values. Eitherway, always calling the callback function would provide the ultimate user control.
  10. I was just running through the and it says you can do that, though I do realise this was cloned from the PhaserV2 version which says exactly the same. I did fix it up myself because I wanted to compare the output and play with the uglify compress options but my fix (bodge) was for my purposes only rather than a generic one so didn't try to submit it.
  11. ok, got it. The biggest speed optimisations I made to a version of Phaser I used was to bypass a lot of repeated scenegraph traversal for the functionality I didn't require, so I can see why this has been done, but it'll leave a functionality hole for many. Sounds like the skelly-bone classes are what I'm looking for, not a native/global performance hog, but something available when needed, any idea when dev on these begins?
  12. I don't think this has been working for a while. There used to be a rocket button for 'run' but this was also problematic. The examples heve never 'all' worked and never will, probably a little frustrating for newcomers, it should have a warning at the top really. I think the sandbox was created partly in reason to pickup where the examples were failing. It's usually quite easy to port an example to the sandbox, eg In preload(), just has to add the baseURL and crossOrigin then modify the asset paths accordingly (remove the leading assets/), so you'll have the same example running that you can modify and run and save for later. Not sure if there's anyway to target a specific version of Phaser though.
  13. Hi, Just had my 1sy proper quick play with with Phaser3. I notice that the traditional scenegraph approach for transform inheritance was dropped earlier in the year and now runs with single flat display list. Is that right or have I got the wrong end of the stick? If this is right, then how do we now have a child object whose transform is offset from a parent object? Cheers, Stu
  14. you'll need to fix up the webpack.production.config.js to be able to run 'npm run deploy'
  15. Hi Tangaroa, If an image coordinates aren't whole integer values then it can't be drawn to the screen with exact pixel mapping, the renderer may have to sample outside the source pixels which is where you sometimes see a little bit of the adjacent image ("edge pixels from the catcher's net"). If you want to keep your movement at += 5.5 then add some blank space to separate each cell of your spritesheet (look into texture atlases and the texturepacker app, this makes doing this easy). This will also help with this problem showing during scaling. Or change your movement +- to whole integer values so there are no 0.5's being added to the coords.