If the image name is the same it will not load.

add a suffix to the end of the name.

example

yoururl.com/image.png?1

yoururl.com/image.png?2

yoururl.com/image.png?3

Such a simple solution, I've been struggling for quite a while now.

Thank you my friend!

Added a global variable called suffix that increases with every load.

I take it the cache problem goes away with every restart and there will be no problem reusing suffix between runs. (?)

I have a sprite.

I load image from url, "1.png".

...Some time passes... (and the image on the server has been replaced by another.)

Same event runs again, load from url, "1.png".

But the sprites image does not update. (restart of layout does not solve, but stop/start new preview does.)

I have tried to destroy all sprites,wait a few ticks, and unload all images from memory.

The sprite however is a template from a (unloaded)repository layout.

The .htaccess on the server providing the png has no cache settings.

any ideas?

Any ideas on how to fix or work around?

([outline=#eeeeee] Mitt meddelande är ganska långt [/outline])

I use a global layer included on all layouts containing a transparent debug "console box" (sprite) with dragdrop behavious that has various "godmode" buttons and text about what is going on behind the curtains (ex. "looping through inventory").

Keep it in "is in preview" and set text to "right(textobj.text,100)" to prevent memory leak, be aware of any debug related code that may run even if its not in preview that could hog memory after hours of play.

I've just taken the project apart piece by piece until the bug revealed itself (Ockham's razor)

In a double tap event, under a very uncommon condition, one frame in a 5fps animation caused an infinite loop where a function kept calling itself (collision not triggering).

That caused the system not to recognize that the user ended the touch, and that touch index got frozen in place.

Ashley the bad programming was mine to own, but perhaps this is something that can be avoided? Some sort of failsafe? (on the other hand... a failsafe would have kept me with bad code)

I'm working on a game (tower defense type) using touch controls, it has plenty of touch tap and double tap (no multitouch required).

After a while playing on an android device (samsung A5 2017 "SM-A520F" and Js Galaxy tab A "SM-T510") the touch index gets "stuck" as if the user suddenly kept its finger in exact same spot while continuing using an additional finger.

This messed upp all "is touching" and set to position touch.X.

Solution (work around):

Create object called TouchPoint and variable TouchToUse.

All touch X/Y replaced with XAt(TouchToUse)/YAt(TouchToUse)

Any event previously asking: "is touching" replaced with "overlapping touchpoint"

and so on.

I've read lots of old threads regarding a similar problem to a device own built in 2+1 gesture, this bug behaves in the same way but the two mentioned devices in my test have no such feature.

I've just thrown out my work-around here in case anyone else faces the same problem, perhaps someone else has a better solution and maybe construct dev team wants to look into the bug.

It would depend on what else it should be able to do, and if others are supposed to work in a similar way.

A suggestion would be to add the object to a family called something like PointFollowers and add the behaviour moveto and a variable MovingTo to that family.

Add another object called MovePoint and make it invisible, give it a variable called PointID and another one called MoveTo.

Pointfollowers on collision with MovePoint -> set Pointfollowers.MovingTo = MovePoint.Moveto

-subevent pick all MovePoint

pick MovePoint where PointID = Pointfollowers.MovingTo

action Pointfollowers.MoveTo MovePoint.X MovePoint.Y

You could even experiment with using the same PointID and add pick random in multiple Movepoints to have the AI take random paths.

Any object in that family will then follow movepoints as you have set them up.

I think the only way for iPhones is to publish an app. Apple don't allow Safari on iPhones to go fullscreen, and because Apple ban other browsers (no idea how they get away with that), other browsers are just skins over Safari and so have the same limitations. So if you try Chrome, Firefox, Edge or anything else on iPhone, you're still using Safari but just with different branding.

Wow! that "skinned safari" was a bust, amazing they haven't been sued (yet).

I'll just wait and see what the future brings then, and users will have to settle for a slightly smaller play-area.

It's on iphone, not sure which one that is then, guess ios.

Anything that can increase the play area would be helpful, hiding tabs, using some html element or browser js command perhaps?

I've reached the limits in my knowledge in this area now.