Scoremonger's Forum Posts

  • WackyToaster Interesting, thanks for the idea! That is generally what I do already - create a family-specific function to spawn arbitrary members of the family by name, so this test could slot right in. Much-appreciated.

  • winstreak Thanks for the suggestion! That would work in a test where I know exactly when things are created, but I'm talking more about confirming object spawns using event sheet logic so I can log errors and otherwise respond to an object that failed to spawn. I'm often spawning at unpredictable (and very fast) rates, so it's not always possible to manually monitor which spawns succeeded and which failed.

  • Noah1 that looks very similar to what I'd tried, but I'll copy what you have verbatim and see how it goes, thanks!

  • In the Example Browser you can find a template project called "Gamepad Twin Stick Shooter" that illustrates how this works. It's using the right analog stick for aiming, not the left, but otherwise it sounds like what you're looking for.

  • Ah, thanks R0J0hound, that's a great idea. It seems so obvious now that you handed it to me on a silver platter, haha. XD

    And yeah, I am using create by name as a half-baked way of composing more complicated objects in a way that the pieces can be swapped in and out dynamically (a very poor man's component system, I guess.) So, I'll add some name variables or an array of names to an object, and at runtime I'll spawn objects with those names and add them to a hierarchy or manually store their UIDs. If I ever want to change them to another type, I can delete the instances, change the names, and spawn new objects. Since we don't have object or family type variables in this engine, I used this workaround, but it's also fraught with peril. It's so error-prone.

    Anyway, thanks again, much-appreciated.

  • OK, thanks, I'll use that one!

  • Noah1 Thanks, great tip! I tried it, and it does work in a vacuum, in terms of confirming that the function created an object when that function was just called once.

    Unfortunately, waiting a frame to do that test breaks downstream logic that depends on immediately getting a valid UID back from the object's creation. But worse, if I'm calling the creation function multiple times in a frame, passing in different object names, I can't say which function calls succeeded and which didn't. At least not if I'm using logic that simply checks how many objects exist after the function call(s) and comparing that to how many existed before them. Maybe there's a different method I don't know about, or maybe there's just no simple way to do this in C3 yet.

  • In most engines there is a way to create an object and then immediately know if it was created or not. I.e., if (createObject("spaceship") != null) etc. In Construct you can reliably create a family or object instance if you set it explicitly, but if you create an object by name there seems to be no way to know for sure that the thing you tried to create actually got created (or even exists, for that matter.)

    I tried making my own function to mostly do this, but it doesn't work. The logic for it is like so. I did it in event sheet logic, btw; this is just pseudo code. 

    Local number startingObjectCount = 0
Local number finalObjectCount = 0

startingObjectCount = customobjectfamily.count;
System.CreateObjectByName("MyObjectName");
finalObjectCount = customobjectfamily.count;
if (finalObjectCount > startingObjectCount) {
	PickLastCreated(customobjectfamily);
	return customobjectfamily.UID;
} else {
	return 0;
}

    Unfortunately this always returns 0, I'm guessing because the object isn't actually created until the next frame? Does anyone know if there's a simple way to test at runtime whether the object type you're trying to create got created. I love creating objects by name because it's so powerful, but man it's annoying when there's a typo or something gets renamed and not all the calls that create that object type are manually updated too. Thanks.

  • Does anyone know if there's a new suggestions platform in the works (or already existing?) I was about to add a suggestion, but there's a note on what I think is the current platform that it will be replaced in early 2022. Did that happen? I did a quick search but couldn't find anything newer than that.

  • Try Construct 3

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

    Try Now Construct 3 users don't see these ads

  • "If a replica is modified individually, the affected property will stop taking the value from the template and instead will take the value from the replica itself. To indicate this, the property will no longer be highlighted."

    From a user perspective i would suggest having modified fields also have their color modified.

    I understand not wanting to have a rainbow of colors, but it this case having the color distinction instead of the regular grey should help users.

    Agreed, I was just about to suggest this myself. Highlighting overridden values in some way is standard in most engines and quite useful.

    The other feature I would love to see in this system is support for creating objects by name, which is something I use all the time. Perhaps there can be a "create object by template" (which takes an arbitrary template name string as an argument) action or something similar?

    All that said, templates should be a huge workflow improvement for Construct!

  • When you mouse over thumbnails in the store bundles the images tilt and glitch. I assume this is a bug?

    It would be nice to be able see the full thumbnail of the asset in the bundle my mouse cursor is over. Otherwise it seems better to just disable that tilting effect because it's pretty irritating at the moment.

  • The old ideas page did get archived, and there is a new one.

    https://www.construct.net/en/forum/construct-3/general-discussion-7/announcing-new-construct-161682

    *facepalm*

    Thank you oosyrag, I totally missed that.

  • ~6 months ago there was a discussion about wiping the ideas platform on a cyclical basis to try to get some value out of it. Any news on this, Ashley?

    Tagged:

  • We have to work with how the existing feature suggestions platform works. It can't stop unlimited submissions. So as I said before, part of limiting voting is also to stop unlimited submissions, with a policy that ideas with zero votes will not be considered.

    Yep, that was just a lament. ;)

  • Maybe I don't understand something, but if Ashley is going to pick 5 most upvoted ideas, why does it matter how many people voted on them? Limited or unlimited votes - in both cases some ideas will get more votes than others, so just pick 5 from the top!

    .

    It's a moot point because it's not happening, but... ^this^.

    I can't think of any reason why unlimited votes would result in a less accurate view of the consensus of the entire user base on which ideas are best. More votes will still accrue on the issues that are the most appealing to the most users. It will just make engagement with the platform easier. It's a shame there's no way to limit the number of ideas each user can create so we could at least try unlimited voting. Ah well, such is life I guess.

    My hope is that the six month reset will result in more engagement by Scirra on the top ideas, which should still be a nice improvement over the current situation. So I'm excited to see how this goes, despite my grumbles.