Ashley's Forum Posts

    • Post link icon

    If you run in to a bug or issue in Construct, please post it to the GitHub issue tracker here:

    https://github.com/Scirra/Construct-bugs

    You must follow the bug report guidelines or your issue will be closed without investigation.

    If you have suggestions for new features or improvements in Construct, we also have a feature suggestion tracker here:

    https://github.com/Scirra/Construct-feature-requests

    Your must follow the suggestion guidelines or your suggestion may be declined.

  • Third-party cordova plugins can be supported in the build service, but we maintain a whitelist of allowed plugins for security reasons. If you want a third-party cordova plugin to be considered for whitelisting, please file an issue.

  • Why not just save directly to the cloud?

  • You do not have permission to view this post

  • Closing, please report C3 bugs to the C3 tracker. This forum is for C2.

  • Honestly, I struggle to see anything wrong - I've had tests like this one which I've been using for ages. I just ran it on my HTC 10 and it consistently measures it hitting v-sync within 1ms. On my desktop it measures <= 0.1ms. In Firefox it's so accurate, the v-sync jank is locked on 0ms. So I can't see anything wrong.

    I would guess it's somehow device or OS specific. But so long as everything looks fine on my end, what do you expect me to do? It looks like our engine is v-syncing just fine. As ever, a minimal reproduction with the right hardware and OS combination is the best way to investigate any issue. You can also report issues to browser vendors yourself directly, which I would encourage you to do so - if you report them to me all I would do is forward the report to them, but I won't if I can't reproduce it myself. You can easily cut me out of that process.

    If you have general performance concerns, the C2 engine is very mature by now, having been used widely for ~6 years. The performance characteristics are generally excellent, there are no significant known bottlenecks or weaknesses, and it's robust and proven for a wide range of games. I have tested and profiled heavyweight games like The Next Penelope and reviewed our engine's performance, and everything was working fine. In the past, it's also been very, very common that people blame us or browsers for their own mistakes, like extreme fillrate abuse. I am very confident our engine is actually working great. That's not to say there aren't any bugs or issues, but by now finding any kind of serious deficiency in our engine is actually a rare and pretty surprising case. I think you should weigh up the likelihood of that accordingly when running in to issues.

  • It should save to the same location you last saved to. If it doesn't, please file a bug.

  • Cloud save is as fast as the cloud service. Maybe you should try a different service?

    You don't have to keep downloading copies either, you can always just save to the browser local storage and cloud save at the end.

  • > Last I checked, most browsers could v-sync reliably to within 0.1ms. I'd be interested in any .capx examples I can try out. It's also important to say which browser and OS you're testing on.

    >

    https://www.dropbox.com/s/tyfj5687kmvf8 ... .capx?dl=0

    On my Windows 10 PC it shows consistently one color. It drops a drame about once every ten seconds. That's about the same as our old native engine in Construct Classic managed. So we have reached parity with native.

    (Why would a native engine drop frames? The OS has dozens of threads to schedule, and sometimes it might just schedule some work somewhere else and miss a 16ms frame deadline. Often if you go fullscreen the OS boosts the process priority and this happens less.)

  • ...and "Letterbox integer scale" fullscreen mode, to avoid fractional scaling.

  • Try Construct 3

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

    Try Now Construct 3 users don't see these ads
  • "Letterbox scale" fullscreen mode can distort pixel art because it allows fractional zoom levels, e.g. 2.7x scale.

    That's what "Letterbox integer scale" is for - use that for pixel art.

  • The C3 manual is complete, and it lists system requirements.

  • Well, I posted that about eight years ago, but I think I meant just a simple subtraction between the current mouse position and the last mouse position gives you the change.

  • Why not just implement the collision flag yourself in a boolean instance variable? Ignore the condition if the boolean is false, and you have a per-instance, per-event way to handle collisions.

  • Overlap is a collision check, so if you turn off collision checks, all collision checks are ignored and return false. That is by design.

    Construct has a really well optimised collision engine, generally it's not necessary to disable collisions to improve performance.