Ashley's Forum Posts

  • I would be skeptical of OS-level memory measurements. At the OS level, memory management is surprisingly complex, and for various reasons Task Manager can report memory shared with other processes as part of the memory usage of a single process, resulting in inflated numbers or double- or triple-counting the same memory allocations. Also if you have loads of RAM, from the OS's point of view it's like a big really fast cache space, and not using that is a waste. So as long as you are not actually running out of memory, then the OS will use as much RAM as possible, to make best use of that. Often OSs like Windows are accused of "wasting memory", but thinking the OS should keep RAM as empty as possible is similar to thinking that your CPU should keep its L1 cache as empty as possible, just in case you suddenly need to put lots of stuff there, whereas it actually performs far better to fill it up completely and evict unused things if you run out. So if Windows sees you have gigabytes of space free it's going to start filling it up with file caches and stuff to maximise performance. There is a whole cascade of other effects too - like Chrome might decide, why waste CPU time running a GC now if it has room to make several gigabytes of further allocations? There's no pressure at all there.

    So overall the task manager reported numbers are pretty misleading. If you see it using 1 GB RAM there, that does not at all mean it is actually using, or requires, 1 GB of RAM.

    Music is not held decompressed in memory, but sound is. You can easily test the memory impact by turning off 'Preload sounds' and comparing the startup memory use. If you don't preload sounds though, you have to have a whole system of determining when to preload and when to unload the sound effect, and mistakes mean high latency on the sound playback, so I don't really think it's worth the bother.

  • The VM is probably using a special network configuration that doesn't work the same as having the machine on the Wifi/LAN. I think it works something like by simulating a network connection between the VM and host machine. Unfortunately I'm not sure how you'd configure this so the VM can access the same Wifi network in the same way. Maybe there are some networking options in Parallels.

  • — - it's entirely sensible to guide optimisations based on measurements. However I would be amazed if anyone could measure any significant performance difference for the question asked in this thread. Changes on that level are really not important.

  • Try Construct 3

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

    Try Now Construct 3 users don't see these ads

  • It's by design - I can't think of many reasons you'd need to update in the background and couldn't do that upon waking up, especially given C2 is mainly a game engine. Also if you do too much work in the background you'll drain the device battery and annoy users.

  • That CPU doesn't look very powerful. I'd go for something 2 GHz+, but on the other hand if your games run well on that system it will probably run well everywhere!

  • Lots of users are successfully publishing to iOS and Android. If notnsane has any specific criticisms I'd be interested to hear them, because AFAIK things are working well at the moment.

  • Brute mode could save anywhere from 0-5% on the file size, and takes several times longer to process, so generally it's not worth it.

  • Post to the Bugs forum following all the guidelines. There is nothing anyone can do with a screenshot of an error message.

  • I can reproduce, but there are no errors reported by Opera so it appears to be a browser bug in Opera itself. It works fine in Chrome. Opera Developer is also pre-release software so may be subject to issues that don't occur in the main release channel. I'd recommend reporting this issue to Opera instead.

  • Closing as by design: like the other behaviors, the Physics behavior handles collisions and moves objects so that they are no longer overlapping, before the "Is overlapping" condition is tested.

    However "On collision" still triggers. You can test this by changing the text object when "On collision" triggers instead.

  • Closing, please follow the bug report guidelines. I cannot reproduce the problem here, and presumably nobody else apart from you can either, since zooming is a very common operation and nobody else is reporting this.

  • Closing as not a bug, it's just someone trolling the room by leaving themselves as host while having a poor quality connection. The movement is lag compensation due to the poor quality connection and your movements not correctly syncing with the host. Try changing the room name and hosting your own room.

  • Does this still happen after exporting? We've had a couple of bug reports in the past that are related to the C2 preview server, and the problems go away when hosting on a different server.

  • The C2 debugger runs the game in an iframe, and if I debug from C2 preview with IE11 it works correctly. Can you confirm it works correctly for you there as well?