Problem Description

Mouse & Touch plugins not detecting mouse - up event when a browser alert is issued.

Attach a Capx

https://www.dropbox.com/s/78xzr94z2e33w ... .capx?dl=0

Description of Capx

Demonstrates that the Mouse up event does not fire after a browser alert is issued, causing in game problems when trying to detect if the mouse button (or touch) is being used currently.

Steps to Reproduce Bug

• Run game, note the text boxes display current status of mouse
• Click the left mouse button and hold it down, and then press any key on the keyboard, and it'll issue a browser alert (after the alert shows up, let go of the mouse button)
• With the mouse, or keyboard, press the OK button on the browser alert.

Observed Result

It still thinks the mouse button (or touch) is down even though it isn't.

Expected Result

It should either send a mouse up event, or simply detect the mouse is no longer pressed, even if not sending the mouse up event.

Affected Browsers

• Chrome: (YES)
• FireFox: (YES)
• Internet Explorer: (YES)

Operating System and Service Pack

Windows 7 SP1

Construct 2 Version ID

r209

Problem Description

A one shot particle eventually takes 100% CPU

Attach a Capx

https://www.dropbox.com/s/z95oc6jjqdsiq ... .capx?dl=0

Description of Capx

I, on purpose, blasted the screen with a bunch of particles. Remove some of these lines if it brings your computer to a complete halt. The purpose is to show that eventually, the FPS at first goes right down... then eventually speeds up as the particles exit the screen. Then after waiting approximately 15-20 minutes (900+ seconds), the FPS drops down to near frozen state again... eventually it sometimes crashes the browser. This works BEST on a slower computer... and must be done without WebGL.. but will work on any speed computer eventually.

Steps to Reproduce Bug

• Load game and observe FPS
• Wait till most particles leave the screen and observe FPS
• Wait 15-20 minutes (900+ seconds), and observe FPS drop again

Observed Result

FPS Drops down, and CPU consumption rises.

Expected Result

CPU consumption should not rise at all.

Affected Browsers

• Chrome: (YES)
• FireFox: (YES)
• Internet Explorer: (YES)

Operating System and Service Pack

Windows 7 64BIT

Construct 2 Version ID

204.2... but has happened for many months... only recently found out what the issue was after many hours of testing.

- i have developed apps like , you can even make an app using Construct 2

Are you saying you have something like this?

If so, I'd be interested in using it.

Thanks!

I'm not talking large dlc... I'm talking new levels. And json is only an option if you have less than a couple dozen objects... but lets assume you have hundreds of objects.... good luck with that. C2 is all about doing things quick and easy.... pre-determining with accuracy, hundreds of objects in a layout, through json is not exactly doable.

Possible feature request for Construct3?

It really doesn't seem like it would be that difficult to incorporate, if you have the same graphics, etc etc... all you're doing is loading the same .xml file... just externally, instead of internally.....

If your plans are to remove Box2D web, as your post suggests, then you'll need to fix your Box2D ASM so that it works on a Power PC first.

Interesting... thanks.... I'll keep that in mind.

What kind of bug if you don't mind me asking?

Is it possible that a Boolean takes longer to set than an integer, so that it needs that little bit of extra time to retrieve the information? I'll test that theory tonight... but it honestly makes no sense.

Ok, so I made this plugin... and today ran into a weird situation that hasn't happened on any other of my projects with this plugin. So I have a trigger that is called, when you perform a certain action. When this trigger from within the action is called, it does, but funny things happen. Here's an example:

Lets assume a Sprite has an instance bool "hair"

A Specific Sprite.Hair = True
MyPlugin.CallFunction       (at this point, my function calls back with "MyPlugin.Trigger"
[/code:3jfna26x]

[code:3jfna26x]Conditions:
MyPlugin.Trigger    // it triggers....but.....
    >   Sprite.Hair = True  (grab the only Sprite that has a Bool that = True)
             > Action:  Perform whatever.[/code:3jfna26x]
DOESN'T WORK... It doesn't SEE any Sprite.Hair as = true... even though it's there!!! 

Now.... if I leave it as this, it doesn't work.... but only with this one function, no matter how many times I try renaming it, moving it around in edittime, or whatever.   It's very odd... HOWEVER... if I change my conditions to this:

[code:3jfna26x]Conditions:
MyPlugin.Trigger     ->  Wait [b]0[/b] serconds,  then Call Function "Wierd"

On function "Wierd"
 >    Sprite.Hair = True  (grab the only Sprite that has a Bool that = True)

Action:  Perform whatever.[/code:3jfna26x]
THAT WORKS????!?!!!!!  WHY?!?!?!  NOW it see's it? Why not earlier?  I don't understand.  Shed some light perhaps?

Nitro187 - the only difference I can see from those logs is when it's moving around the backupN files it gets "SHARING VIOLATION" for one of them. But then it goes ahead and carries on anyway, which is why I don't understand why it hangs, it appears to show it continued after that and it must be hanging somewhere else.

Does this only happen if you have your actual .capx file in your Dropbox (or Google Drive) folder? You don't need to do that to back it up, just set one of your backup folders to the Dropbox and as you work on your local .capx it will copy it to Dropbox from time to time anyway. I'm pretty sure that backup process only copies a single file in one go, so it should work with these file syncing apps.

I honestly never use the backup folder options in the preferences.... those are just set to default. It never really occurred to me before that I could just turn off backups all together - which would remove this issue... but also leave me vulnerable too. This happens when I have everything saved in the same directory, like it would normally. I never save a .capx in the same directory, the youtube video shows that just to show you the bug occurs with caproj's only. I guess it would just be awesome if there was a cancel button to close that window, so you can 'try again' or something. If that's not possible, I guess I'll just have to disable backups completely unfortunately.

Awesome.