Will the next version of the 'haxelib swf' library 'optimize performance'?

.
Will the next version of the ‘haxelib swf’ library ‘optimize performance’?
.
I heard that the next version of the “haxelib swf” library will be released next month. I would like to know if there is any “performance optimization”, such as “MovieClip” supporting batch rendering to provide higher performance?
.
After all, the performance of the “haxelib swf” library is really poor now, so I use “starling” instead
.

.
‘haxelib swf’ has been released for many years,
At present, it is still a “beta version” with poor performance ..
.
I feel that the ‘OpenFL’ team is not focusing on performance optimization, but rather on compatibility
.

You can check for any improvements in the network graphs.
Network Graph · openfl/swf · GitHub
and also check forks and their commit history.
Forks · openfl/swf · GitHub
Also worth looking at g2md ( but it is only the ‘g2md/swf’ really that relates to swf ). It looks like the last change was on font last year.
Network Graph · nmehost/gm2d · GitHub

You really need to start looking through github code yourself as most stuff that is likely to be released in a new build will be in there already. Generally improved support for Animate swf’s are going to be a bit limited towards the starling users, and most haxe users have likely abandoned Adobe long ago!
Realise that libraries like ‘swf’ were largely created in the early days..

When Haxe’s main target was often FlashPlayer for stuff like facebook games, android apps, web microsites, but when adobe abandoned it’s as3 developers and flash designers, and also charged a lot for Animate without really improving it - the small haxe community found tools that were more dependable long term. Animate for instance can not be run on linux. While using timeline swf’s in OpenFL can be really powerful the swf ecosystem with openfl is likely very small these days. So I would not expect much.

Yes, even though the “swf” library has been released for many years,
However, the ‘swf’ library is still a poorly performing beta version,
From here, it can be seen that the author is not enthusiastic about the “swf” library

I think the original swf library was created mostly by Hugh of NME, his last demo was of chess and he used svg for assets in NME so rather than swf assets!
I believe that Josh of Feathers is improving the swf library as a small aspect of openfl with feathers, but the usecase is not focused on gaming specifically and it may not even be that practical to improve performance with swf assets ( once worked on a project where they took frames and converted them to a sprite sheet and similar with svg’s this stuff can be slow). But a more general point you can’t tell opensource projects what you want them to work on as they are individuals that have specific needs that interest them and spare time is valuable.. But you can work on aspect yourself and make a pull request and hope it is accepted. By contributing you may find that forgotten areas become more active but just complaining that others do not spend their free time on your project interest is unlikely to help. This is not an official view this is just reality of opensource as I understand it.

1 Like

When I say hugh created it, when nme was forked to openfl I believe swf was rebuilt as a standalone lib and some work on flashlite format, but the basics were from before the fork I suspect.

It is likely that the performance bottlenecks you encounter when using the SWF library are actually within OpenFL, and not caused by the SWF library’s code. Someday, we will hopefully be able to to render all vector graphics entirely on the GPU, including batching more things into less draw calls. However, that’s a significantly large amount of work to implement, and it may not happen for years. In other words, big improvements are unlikely in the near term. You will need to be very patient. This is not an improvement that can be made quickly with our limited resources.

It’s not because we don’t want to improve performance. Certainly, we want OpenFL to be faster, just like you do.

.
Thank you for your answer. I rarely use “vector graphics” unless I have to use it. Even for the “moveclip” animation created in “Adobe Animate”, I use “sequence frame images”. I use the “swf” library to use the animation I created in “Adobe Animate” (I use sequence frame images)
.
If openfl can merge and render multiple sequence frame images, I think its performance should be improved. This is just my speculation ..
.

In addition to performance, I hope the “swf” library can add the following features: 1. It can customize the converted format. Currently, it is converted to a “. zip” file because some platforms do not support uploading “. zip” files. I now have to manually modify it to “. png” every time. Can you provide support, such as converting to a “. zip” file, but the code can modify “. zip” to a “. png. xx” file? It’s just changed to ‘. png’ here, but it’s actually a ‘. zip’ file

The second feature we hope to add is the ability to set the “fps” for each “moveclip”, just like “starling” and “creatjs”, where each “moveclip” can have its own independent “fps” set