Build 293.10 - Release version

Quote:

---

Originally Posted by NaitorStudios

Please don't tell me you're creating objects through events then deleting those from the frame...
Even tho it usually works, it causes MFA corruption and BAD objects... It's a unintended behavior, something that people exploit, not a real feature...

Honestly a warning should be added, or people will keep doing this although there are threads talking about how this causes corruption...
Same for setting global mode to "Identical", it's more of a legacy feature and it's known to cause corruption as well.
Load on Call for objects can also cause memory corruption sometimes (graphical artifacts), specially when having many objects with it.

---

What does this mean? i never really understood how you can even do this. Create objects through events then delete them from frame. Im guessing its not like unchecking create at start for things like bullets then creating them when needed to fire them and destroy/deleting them on collision.

How can you corrupt an MFA if you create in event and delete at runtime. The frame has to be running to delete destroy the way im thinking. I must have the wrong idea on what this even means. Please can some one explain it in more detail.

Quote:

---

Originally Posted by fredetmumu

"S’il vous plaît ne me dites pas que vous créez des objets par le biais d’événements puis que vous les supprimez du cadre..."

yes, i didn't know it could be a problem,

@ VOLNAISKRA:
i have 30 differents actives with qualifier "good"
15 are displayed on the frame editor and ,15 are not because i create it in the program then i remove it in the frame editor

i want to rename the alterable value A of all the "good" qualifier
i click on "select all actives with the same qualifier" and i rename alterable value A, but only the actives on the frame editor are concerned.
i must put one instance of every actives on the screen. Now i kno than it's preferable to let every actives on the frame editor and unchek "creat at begining"

---

If you enable "show created items", it will show all 30 items in the frame editor, including the ones that don't usually exist there. Then once you've finished changing Alterable Value A, you can disable "show created items" and it'll again show only 15.

Quote:

---

Originally Posted by ASD

"Animation is over" condition are triggered with a Vector movement. But, the Bouncing ball and Path movement (and more other movements) aren't triggerd condition (they are triggerd after change animation sequence).

---

I've compared Vector and Simple Ellipse and I'm puzzled, I don't see any difference that could explain this difference. What's happening is that both movements use the Walking animation, which is automatically replaced by another animation by Fusion when it doesn't exist. In one case it triggers an anim is over condition for the walking animation and in the other case it seems to trigger this condition for the real animation that is played.

If you create a Walking animation and check if this animation is over, then apparently it works in the Ellipse movement. I'm a bit busy with the current build currently, so maybe I'll retry later. Not sure if it's the same reason and fix with the other movements you mention.

@c4t "Create at Start" was added so people stop relying on this exploit, on the surface it behaves the same, but there's more to that...
When you create a object through a event and remove it from the frame, Fusion might have trouble identifying the object.
So the more you work on it after removing it from the frame, the more corrupt it gets, and it can start affect other objects as well.
Sure, events will be kept, but at what cost?

Internally don't store the events the way we see it (for obvious reasons), it stores it in binary, the same event copied from different places might have completely different internal code even tho the events look the same.
Once this internal code is messed either by file/drive corruption or event corruption, usually you can still open it, work on it, but things start acting up...
It even has a warning when corruption gets too far, and after a certain point it will just refuse to work all together.

Think like this: We can't delete objects created from events for the same reason we can't create objects that aren't on the current frame with Global event editor.

Quote:

---

Originally Posted by NaitorStudios

@c4t "Create at Start" was added so people stop relying on this exploit, on the surface it behaves the same, but there's more to that...
When you create a object through a event and remove it from the frame, Fusion might have trouble identifying the object.
So the more you work on it after removing it from the frame, the more corrupt it gets, and it can start affect other objects as well.
Sure, events will be kept, but at what cost?

Internally don't store the events the way we see it (for obvious reasons), it stores it in binary, the same event copied from different places might have completely different internal code even tho the events look the same.
Once this internal code is messed either by file/drive corruption or event corruption, usually you can still open it, work on it, but things start acting up...
It even has a warning when corruption gets too far, and after a certain point it will just refuse to work all together.

---

I think this is a bit outdated, MFA files don't get corrupt easily nowadays, I don't remember having seen a bad MFA for a long time. Though I agree anyway, having at least one instance created in the frame editor with "Create at start" unchecked is better, at least for practical reasons like the issue reported by fredmumu.

[bug] spotted: StringParser on [ios], expression "replace$"

simple example:
source text: "ttttest example"

if you replace "ttt" with "tt", the app will crash with an "out of bounds exception". (reason: cause it will find in the output the same replace-string again)

thanks StingRay i will check

Feature Suggestion: Support for Android display cutouts (https://developer.android.com/guide/...display-cutout)

My device currently has a big gap on the Top(Portrait) or Left/Right-side(Landscape) while in immersion mode.

Would be really useful if Windows runtime got a image compression feature just like Android (None, Color Compression or JPEG).
In addition to this, a new JPEG + Alpha Channel would be awesome, the best of both worlds.

Basically on building it would separate RGB and Alpha into 2 textures, convert those to JPEG, and then during runtime it could have the benefit of JPEG without losing transparency/alpha channel.
Would significantly decrease the file size, memory consumption (if it can handle compressed images on memory) and improve performance.

Got into this topic cause people asked if choosing a lower the number of colors would use less resources...
But got me thinking that if would probably be converted into uncompressed images on memory, meaning it's only useful for making sure it would look accurate on low color displays... (basically a legacy feature)
I could be wrong about it tho.

Not sure if this would be really interesting. With 2.5+ images are already compressed in RAM, a 2.5+ app uses much less RAM than normal 2.5 apps. JPEG is better but decompression is slower.

I see, does this apply for mobile as well?
Cause I think JPEG + Alpha could be nice on mobile at least

Image compression on mobile is defined by the Image Compression option in application and object properties. There is JPEG compression in 2.5+ apps, not sure if JPEG + alpha is easy to add.

I've just discovered the meaning on multiplying by In.Tint on DX11, it's actually what carries the RGB/Blend Coef.
Now I was wondering: Is there any way of getting this on DX9 or isn't it implemented? Cause I would like to apply it to a texture parameter...

Why not just pass it in yourself as a parameter?

The available PS versions are pretty limited on DX9, and I already have a lot of parameters...
I got used to DX11 lacking DX9 features, now suddenly DX11 has a feature that is lacking on DX9... lol

If I got no choice, sure, I'll try pass as a parameter, but seems like something pretty basic that should be built-in or it makes no sense that they have included it on DX11.

why not webp?

Would it be possible to add alignment options to the string objects? I find myself always creating 3 version of the same string object just to cover left, center and right alingments, would be nice to be able to set this during runtime for each instance.

Quote:

---

Originally Posted by Tyrexchip

why not webp?

---

Maybe CT can release image filter SDK and HWA SDK, then we can add new format support (e.g., webp/hevc) by ourselves, as CT is already too busy.

Quote:

---

Originally Posted by defisym

Maybe CT can release image filter SDK and HWA SDK, then we can add new format support (e.g., webp/hevc) by ourselves, as CT is already too busy.

---

I have the image filter SDK, but it has some limitations - I looked into porting SVG vector graphics, but Fusion expects the image filter to give the width to runtime, and rescales on the runtime side, not vice versa.
At any rate, the filter SDK is for Windows only.

Would be cool if we could get the total amount of backdrops, included on frame editor and pasted (if these are counted separately, it should be 2 expressions)
Might be useful for level editors and to debug in case backdrops stop showing.

BUG:

The alpha channel of sprites get broken(removed?) when resizing the Window Frame while in Direct3D 11. This bug does not occur in other Display Modes.

Attachment 31098

I solve with text rect object and a little bit of math, but i think It can be really usefull change the alingments in runtime!
If you wanna i send you my exemple tonight cause now im outside

@SirEatAlot

I make sometingh wrong, but the last message is for you :)

Would be nice to be able to use middle mouse button to pan on image editor, been missing this, I remember this was added to a few other places not long ago

Quote:

---

Originally Posted by NaitorStudios

Would be cool if we could get the total amount of backdrops, included on frame editor and pasted (if these are counted separately, it should be 2 expressions)
Might be useful for level editors and to debug in case backdrops stop showing.

---

question regarding this, because i found this was occuring; add backdrop - destroy object; i would have thought this was not contributing to any overhead - is it? i dont really know why theres a limit on it, unless it is having some sort of overhead.

Add backdrop and destorying the active does help with performance, cause Actives have a lot of stuff going on (although they are very optimized nowadays)
But, you have to keep in mind that there are two object limits, both share the same threshold (set on frame properties)
Backdrops will not increase the "general" object limit, but it will take from the "backdrop" object limit.

Creating backdrops and destroying actives can be a bit intensive if done multiple times at once tho.

Been having the oddest issue while porting DX11 shaders...
I'm trying to port ColorWheel, part of the Color Picker pack,for some reason alpha isn't working on the premultiplied version.
So I tried debugging it by displaying the alpha channel as RGB and it shows a perfectly good alpha channel texture.
I can't Demultiply and Multiply at the end cause there's no texture at all, it's all generated by math.

I THINK premultiplied DX11 causes an issue when the alpha value is outside the range of 0.0 to 1.0...
I don't have any idea on how to solve it.
Maybe this is a bug on the premultiplied DX11 routine?

Quote:

---

Originally Posted by NaitorStudios

I THINK premultiplied DX11 causes an issue when the alpha value is outside the range of 0.0 to 1.0...
I don't have any idea on how to solve it.

---

If that is the problem, the "clamp(x, min, max)" function might be what you need (or perhaps even "clip(x)").
https://docs.microsoft.com/en-us/win...ics-hlsl-clamp

I've tried clamp, min & max, clip, saturate, normalize... Nothing works.
If I return the alpha channel as RGB after any of these, it still looks the same (which looks completely fine), but checking with if conditions I can see that it's now in the proper range.
But still, it refuses to work, it kinda of affects the alpha of the final result a bit, but a lot less than it should.

I don't know if anyone else would care about this but I would like to be able to manually set the hotspot wherever I want when importing images, not only the options "Top-left, Bottom-right" etc. :D

i also think it would be nice if you could choose to keep the hotspots & action points where it currently is when importing new image/animations, as sometimes you have to tediously type them all in again X)

Quote:

---

Originally Posted by SHINBAXTER

i also think it would be nice if you could choose to keep the hotspots & action points where it currently is when importing new image/animations, as sometimes you have to tediously type them all in again X)

---

This seems like a good idea to me. For me, it's exactly this situation (having to redo hotspots when updating an existing image) that is the biggest problem. Being able to specify any hotspot at all upon import would be nice, but it's a similar effort to specify it after import. But having to repeatedly type in the hotspot for an image that you frequently update is the real annoyance, and a "keep original hotspots" checkbox would solve that.

You guys should use box-mode when importing graphics - it's just better.

"This option is very powerful: it allows you to capture several images contained in boxes in a single image file. In order for this option to work, you must respect some rules when saving the image file. First, every graphic you want to import must be surrounded by a rectangle of a color different from the background color of the image. The same color must be used to define all the rectangles surrounding the graphics contained in the image. Secondly (optional), you can define the hot-spot and action points of each image with a pixel of a different color in the horizontal and vertical lines of the rectangles. To define the hot-spot, indicate its position by a pixel of a different color in the top line (X coordinate) and left line (Y coordinate). To define the action point, put a different pixel in the bottom line (X coordinate) and right line (Y coordinate)."

Quote:

---

Originally Posted by Volnaiskra

This seems like a good idea to me. For me, it's exactly this situation (having to redo hotspots when updating an existing image) that is the biggest problem. Being able to specify any hotspot at all upon import would be nice, but it's a similar effort to specify it after import. But having to repeatedly type in the hotspot for an image that you frequently update is the real annoyance, and a "keep original hotspots" checkbox would solve that.

---

I confirm, just had problems with updating graphics and losing hotspots, it's terrible

Quote:

---

Originally Posted by MuddyMole

You guys should use box-mode when importing graphics - it's just better.

"This option is very powerful: it allows you to capture several images contained in boxes in a single image file. In order for this option to work, you must respect some rules when saving the image file. First, every graphic you want to import must be surrounded by a rectangle of a color different from the background color of the image. The same color must be used to define all the rectangles surrounding the graphics contained in the image. Secondly (optional), you can define the hot-spot and action points of each image with a pixel of a different color in the horizontal and vertical lines of the rectangles. To define the hot-spot, indicate its position by a pixel of a different color in the top line (X coordinate) and left line (Y coordinate). To define the action point, put a different pixel in the bottom line (X coordinate) and right line (Y coordinate)."

---

That's probably good advice for a lot of people, but it's not relevant to all workflows. I've got some good streamlined workflows set up to quickly get my graphics from Photoshop/ToonBoom to Fusion. If I had to consolidate everything into spritesheets and add special rectangles to everything, I would spend more time than I'd save. I can see how spritesheets would be very helpful to some workflows. I imagine they'd be particularly good for pixelart workflows where you work in raster from start to finish and manually draw each frame individually - especially if your source program is designed to export to spritesheets. But I can't see it being of much use to me.

Besides, something like spritesheets/box-mode is probably best suited to well-established assets that you plan out in advance because you know you're going to be modifying them heavily. But there are many times when you're working more on the fly (eg. testing out a quick idea) or just attaching a quick graphic to something (eg. an icon graphic for a collision detector or altVal storage object). In these cases, you may still need to reimport the graphics occasionally, but probably not to the extent that it's worth setting up a separate spritesheet for them.

Quote:

---

Originally Posted by Volnaiskra

"keep original hotspots"

---

perfect name for it, i like :)

Is there any chance of getting the custom icons for counters/strings (the icons you can set yourself in the properties panel) be persistent? Currently, they revert to their default icon whenever you make certain changes to the object (eg. editing the graphics, scaling it in the frame editor).

On a related note, could their custom icon be used in the frame editor? Currently, strings that have "visible on start" turned off all look like identical in the Frame Editor, which obviously makes it harder to work with them, since you keep having to step and check each tooltip to tell them apart:

https://bit.ly/33sI6kd

Perhaps there could be a "generate automatic icons" checkbox (enabled by default). When enabled, a string/counter would behave exactly as it does now. When disabled, it would use only the icon from the properties panel, and would use it everywhere (event editors, frame editor).

Quote:

---

Originally Posted by Volnaiskra

Is there any chance of getting the custom icons for counters/strings (the icons you can set yourself in the properties panel) be persistent? Currently, they revert to their default icon whenever you make certain changes to the object (eg. editing the graphics, scaling it in the frame editor).

---

Yes, it's in the list for the next build too, this will be done if no major issue.

Some people have been getting an glitch where Fusion stops loading most of extensions, even built-in ones, no idea what triggers it, but it seems to corrupt files that Steam isn't able to restore.
So it's necessary to uninstall and manually install all extensions back...
Unfortunately it happened to me yesterday, and I have a massively amount of extensions, so this is really frustrating...

Quote:

---

Originally Posted by NaitorStudios

Some people have been getting an glitch where Fusion stops loading most of extensions, even built-in ones, no idea what triggers it, but it seems to corrupt files that Steam isn't able to restore.
So it's necessary to uninstall and manually install all extensions back...

---

Hmm, never heard of this and I don't see at all why removing and reinstalling extensions would change anything. Have you tried restarting your machine before?